At Amazon, some coders say their jobs have begun to resemble warehouse work

https://www.nytimes.com/2025/05/25/business/amazon-ai-coders.html

by milkshakes • 1 year ago

580 856 1 year ago

nytimes.com

jrs235• 1 year ago

https://archive.ph/HVZRL

jt2190• 1 year ago

This is interesting:

“It’s more fun to write code than to read code,” said Simon Willison, an A.I. fan who is a longtime programmer and blogger, channeling the objections of other programmers. “If you’re told you have to do a code review, it’s never a fun part of the job. When you’re working with these tools, it’s most of the job.”
This shift from writing to reading code can make engineers feel as if they are bystanders in their own jobs. The Amazon engineers said that managers have encouraged them to use A.I. to help write one-page memos proposing a solution to a software problem and that the artificial intelligence can now generate a rough draft from scattered thoughts.
They also use A.I. to test the software features they build, a tedious job that nonetheless has forced them to think deeply about their coding.

timr2• 1 year ago

I was just thinking about this the other day (after spending an extended session talking to an LLM about bugs in its code), and I realized that when I was just starting out, I enjoyed writing code, but now the fun part is actually fixing bugs.

Maybe I'm weird, but chasing down bugs is like solving a puzzle. Writing green-field code is maybe a little bit enjoyable, but especially in areas I know well, it's mostly boring now. I'd rather do just about anything than write another iteration of a web form or connect some javascript widget to some other javascript widget in the framework flavor of the week. To some extent, then, working with LLMs has restored some of the fun of coding because it takes care of the tedious part, and I get to solve the interesting problems.

danielbln2• 1 year ago

I'm with you. I love solving puzzles to make something go. In the past that involved writing code, but it's not the code writing that I love, it's the problem solving and building. And I get plenty of that now, in a post-LLM world.

ape4• 1 year ago

Fixing bugs in my own code is fine. In other's code its less fun.

seanmcdirmid• 1 year ago

Bugs are awesome. I picked my current job at Google because it involved lots of bug fixing and crash investigation.

archagon2• 1 year ago

Code reviews are fun when you can trust the competence of the coder.

Vinnl• 1 year ago

I think for me, the fun comes from preventing bugs; being able to draw on my experience to foresee common classes of bugs, and being able to prevent them through smart code architecture, making it easier for future contributors/readers to avoid walking into those traps. I'm hoping I'll keep being able to do that.

overgard6• 1 year ago

I spend all day fixing bugs and that's why they pay me -- because for most people it's not an enjoyable task. I'm not denying your experience but I will tell you I think you're an outlier. For most people, fixing bugs they didn't create is the worst part of the job.

jerpint• 1 year ago

Part of the fun is also figuring out the “best” way to achieve a thing. LLMs don’t often propose the best way and will happily propose convoluted ways. Clean approaches are still hard to come up with, but LLMs certainly help implementing them once thought up

protocolture2• 1 year ago

100% agreed.

I solve a problem, let the AI mull on the next bit, solve another problem etc.

zeroonetwothree7• 1 year ago

I find both fun. Writing a web form is still kinda fun even after 20 years. I guess it’s like how some people still play the same video games after years and years while others want something new.

owebmaster• 1 year ago

I hate fixing bugs but I enjoy big refactoring, AI-assisted or not - sometimes you just gotta do it all

twobitshifter• 1 year ago

As far as vibes go, these guys sound much more negative.

Ericson2314• 1 year ago

Illiterates. Reading code was always very important. Hell may be mandatory vibe coding, but I'm glad this guy is now forced to read.

bpicolo• 1 year ago

Same, I love fixing bugs

FirmwareBurner2• 1 year ago

Welcome to the club.

Have coders really psyopped themselves into thinking their job is somehow that much more special than the rest simply because it paid better due to temporarily market conditions?

I thought that was a joke where everyone was in on it, not that they were serious. I assumed it was clear we're all replaceable cogs in a machine, baring a few exceptions of brilliant and innovation people.

rvz2• 1 year ago

Have coders really psyopped themselves into thinking their job is somehow that much more special than the rest simply because it paid better due to temporarily market conditions?

Yes. We don't need to pay $$$ for simply changing elements on a page or adopting the next web framework to replace another. The hype in many web technologies that lots of developers that have fell for also contributed to the low quality of the software that you use right now.

All of this work to pay developers to over-engineer inefficient solutions and to give a false sense of meaningful work contributed to the "psyop" of how highly inflated their salaries were to do their jobs in the ZIRP era.

And AI has shown which developer jobs it is really good at, and it is consistently good at web developer roles.

So I'd expect those roles to be significantly less valuable.

FirmwareBurner2• 1 year ago

Software development is a lot more than web development.

layer84• 1 year ago

AI is good at web developer roles because that’s what has been most prevalent in the training material.

danielbln2• 1 year ago

This begins the golden age of creative generalists (until that's also in the chopping block).

bix6• 1 year ago

Is it? We’re more specialized than ever imo.

NBJack3• 1 year ago

Joke's on us; this is going to rapidly drain what little creativity there is in places like Amazon as they rely increasingly more on a tool that at best intelligently regurgitates what it learned/gleaned/stole from the internet. As the AI models are further trained on their own slop, the signal to noise ratio will only get worse; this has already been noted in studies.

rxtexit• 1 year ago

Yes, this is the nature of being spoiled rotten.

I am not a software engineer and have never felt stable in my 30 year career.

It always feels like the rug could get pulled out from under me at any time.

So what? It is still better than working in coal mine. It is still more interesting than working at a gas station.

Hard to feel sorry for people basically complaining the work ping pong table doesn't have quite the quality ping pong balls they were expecting.

It is an interesting mix of being both super elitist and completely infantile at the same time.

chongli• 1 year ago

What even are most programmers at Amazon doing? It seems like all the interesting bits (the MVP of Amazon's major products) was developed long ago. All the low-hanging fruit is gone so the task is now focused on squeezing the few remaining drops of efficiency out of the stone.

John23832• 1 year ago

That’s most major tech companies at this point. The fun was building closer to the mvp of the core product. Once that’s done, it’s just corporate druggery with computers.

Bombthecat• 1 year ago

Yeah, most of the things in IT are just in maintenance mode now.

You need way way less people for that.

iLoveOncall5• 1 year ago

Until you join a FAANG you simply cannot imagine the complexity of the systems and how much has not been automated and how much there is to build.

I've been in Amazon for close to a decade, and I constantly think "I can't believe X hasn't been automated in the 30 years that Amazon has existed and is still done on Excel".

Most engineers will work on new features for at least half of the year, and I personally work on brand new projects or at least features constantly.

hellzbellz123• 1 year ago

working in a warehouse, just recently built over the last 4 years, and the pallet management is done via excel. im pretty sure the scanner app is just a frontend.

horrified to think the totes are done the same way.

saratogacx• 1 year ago

Doing work that helps another team will rarely result in getting any credit and is a great way to get stuck in the bottom of the stack rank (top grading or whatever they call it now). So automating anything that integrates across the org chart won't happen because the incentive structure is against it.

kjkjadksj2• 1 year ago

Last few drops of efficiency out of the aws stone scales to millions of dollars of savings potentially.

andrewstuart• 1 year ago

“OK Andrew, we’re looking for someone for this senior developer and they’ve got to be really really great at copying and pasting between Claude and an IDE.

Now we’re going to set up a whiteboard test here and you can demonstrate to us your best copying and pasting.”

“errrr, do I do any actual coding in the job?”

“Well, yes, inasmuch as anyone does these days. It’s mostly copying and pasting though, but hey that’s what coding IS now, right?”

“OK are you ready for your coding test, here it is: what key is COPY? And what key is PASTE?”

eitally• 1 year ago

But for about ten years, at least, the majority of Google swe rules have been talked about as "just moving protobufs", which is effectively the same. If your job is just plumbing other people's designs for existing products, how fun can it be?

biorach• 1 year ago

I think you do not understand what A Google SWE actually does

ath3nd14• 1 year ago

The place known for:

- High degree of warehouse injuries (https://www.nytimes.com/2024/12/16/business/economy/amazon-w...)

- Making its delivery drivers piss in bottles (https://www.forbes.com/sites/katherinehamilton/2023/05/24/de...)

- Illegally busting unions (https://www.thenation.com/article/activism/amazon-union-bust...)

- Forcing people back into their offices on five day RTO (https://www.businessinsider.com/amazon-5-day-rto-mandate-pol...)

Is now making its white-collar employees' life resemble ...warehouse work? Unthinkable!

NBJack3• 1 year ago

Ruby/Dawson in SLU even looks like a warehouse on the inside, complete with chain link fence decor, exposed concrete floors with a splash of paint, exposed ceiling infrastructure, and spartan decorations. And it was built over a decade ago.

pbiggar• 1 year ago

About 8 or 9 years ago, I talked to a friend who started their first software engineering job. They talked about their job as taking tickets from the JIRA board, completing the change, and putting the tickets back. They were expected to complete 2-3 tickets per day. They didn't enjoy the job, needless to say.

I found this ultra-depressing, and far from what coding was for me - a creative role with great creativity and autonomy. Coding was always solving problems, and never felt like some sort of assembly line. But in a lot of companies, this is how it was constructed, with PMs setting up sprints and points, etc.

Similarly, I spoke to a doctor about how much they loved being able to work remotely at their role - with 2-3 days a week where they just responded to email and saw patients over telehealth. It felt very "ticket" focused and not at all the high status doctor role I imagined.

I suspect that both those roles will be lost to AI. If your role is taking a ticket that says "the box should be green, not red", and nothing more, that's the sort of thing that AI is very capable of.

randallsquared• 1 year ago

They were expected to complete 2-3 tickets per day.

Based on my experience with sprint teams, breaking things down into just a couple hours of work per ticket implies that someone else is doing an enormous amount of prep work to create a dozen tickets per feature. I agree that your friend is performing the work of a development system. I've heard this called "programming in Developer" as opposed to whatever language the developer is using.

leoedin• 1 year ago

I've had plenty of colleagues who expect the job to be that. Just working on a small ticket and moving it to "to test" when done.

It's incredibly frustrating to try and get anything done in a team like that. The reality of most software jobs I've had is that problem discovery is part of it. The only people who know the code well enough to know the problens it has are the developers.

wiseowise6• 1 year ago

They talked about their job as taking tickets from the JIRA board, completing the change, and putting the tickets back.

Where do I sign up?

usrnm• 1 year ago

I said it before and I'll say it again: it's high time we got the taste of our own medicine. Getting people out of jobs has been the main selling point of our industry since its first days, this is what we've collectively been doing for decades. Do I want my job to be automated right in front of my eyes? Not really. Do I see some poetic justice in the whole thing? Absolutely.

Braxton19803• 1 year ago

How is it poetic justice? Were we advocating for automation?

Mistletoe• 1 year ago

You were writing all the code for automation.

perching_aix• 1 year ago

How is it poetic justice? Were we advocating for automation?

Yes? I know I did, still do, and will continue to at least.

Pet_Ant3• 1 year ago

Email made the corporate mailroom obsolete and the letter carrier.

A1kmm• 1 year ago

Software development is the most automated career in the history of all time.

Assemblers, Linkers, Compilers, Copying, Macros, Makefiles, Code gen, Templates, CI & CD, Editors, Auto-complete, IDEs are all terms that describe types of automation in software development.

LLM-generated code is another form of automation, but it certainly isn't the first. Right now most of the problems are when it is inappropriately used. The same applies to other forms of automation too - but LLMs are really hyped up right now. Probably it will go through the normal hype cycle where there'll be a crash, and then a plateau where LLMs are used to drive productivity but expectations of their capability are more aligned to reality.

lloeki2• 1 year ago

In french the two fields that are Computer Science and Information Technology are under the same moniker: "informatique", a portmanteau of information and automatic.

The whole field is about automating yourself out of a job, and it's right in the name.

jampekka3• 1 year ago

Gloating about hardships was and is a great way to ensure things will get worse for workers as efficiency and automation increases.

Another option would be to join forces to collectively demand more equitable distribution of the fruits of technological development. Sadly it doesn't seem to be very popular.

wkat42428• 1 year ago

Sadly it doesn't seem to be very popular

Strange enough the people that have the most to gain from keeping things the same, are really successful at convincing the masses who have the most to benefit from change in this regard to vote against it.

https://pjhollis123.medium.com/careful-mate-that-foreigner-w...

chobeat• 1 year ago

Join tech workers coalition

zaptheimpaler• 1 year ago

Yes! In a fair world, we would all be very excited about jobs becoming 30% more efficient or fully automated with AI, because that surplus would come back to us. Producing more with less work is a good thing! It's only been distorted to become an economic anxiety become the gains from automation are so unevenly distributed.

smokel• 1 year ago

I have been in this industry for some time, and over the years I have only seen more people being glued to electronic devices, not less.

It might have been a selling point, but the status quo is that we are inventing new jobs faster than phasing out old ones. The new jobs aren't necessarily more enjoyable, though, and there are no more smoking breaks.

aatd86• 1 year ago

not necessarily. the economies of scale might be increasing as dev productivity increases. the goal of many businesses being to do more with less.

devnullbrain6• 1 year ago

The person the headline refers to is a webdev. What job is that getting rid of?

garretraziel• 1 year ago

Cashiers, some officials, a lot of the “personal contact with a customer” gets transferred to web. I am not complaining, just answering the question.

sokoloff4• 1 year ago

Web dev doesn’t exist in a vacuum.

Web dev for e-commmerce displaced brick and mortar retail. Web dev for streaming displaced video rentals and audio sales.

happytoexplain• 1 year ago

Savoring suffering is uniquely hideous, and one of the grand hallmarks of almost every facet of the decline of the US. It's a clear, bright sign of the death of one's humanity, and the foundation of all evil.

Is that dramatic? No.

More specifically: Things can be inevitable and also horrible. It is not some kind of cognitive dissonance to care about people losing their livelihoods and also agree that it had to happen. We can structure society to help people, but instead we hate the imaginary stereotypes we've generalized every conceivable grouping of real people into, politics being the most obvious example, but we do it with everything, as you have.

The electrician doesn't "deserve" punishment for "advocating" away the jobs replaced by electricity. The engineer doesn't "deserve" punishment for "advocating" away the jobs replaced by engineering. A person isn't an asshole who deserves his family to suffer because he committed his life to learning to write application servers, or whatever.

closewith14• 1 year ago

If in the process of automating away people's livelihoods, you do not simultaneously advocate for the destruction of the capitalist system that ties the well-being of people and their families intrinsically to those jobs, then you do in fact deserve what's coming to you. History shows that retribution against class traitors is not limited to financial hardship, either.

layer84• 1 year ago

The problem I see is less that of losing jobs, but the fact that the jobs get less enjoyable, less deep work, more mindlessness and less reflection, and possibly also the quality of the produced software decreasing.

Modern AI encroaches upon what software engineers consider to be interesting work, and also adds more of what they find less enjoyable — using natural language instead of formal language (aka code) for detailed specification — which creates a conflict that didn’t previously exist in software technology.

SpaceNoodled• 1 year ago

LLMs can't do creative or unique work. They're really only useful for boilerplate, which is the tedious part.

spacemadness• 1 year ago

To be fair, all of corporate has grated against deep work and well written software way before the dawn of LLMs. Tech debt is one of the things that modern software engineering produces in spades.

casenmgreen2• 1 year ago

I may be wrong, but I think every job creates wealth overall (or it would not exist), and that software engineering has been making some jobs more efficient and others not necessary, and then the wealth which formerly had to be employed where those jobs were inefficient, or had to exist at all, is then employed elsewhere.

If you are the person who lost their job, you get all the downside.

Overall, over the whole of the economy, the entire population, and a reasonable period of time, this increasing efficiency is a core driver of the annual overall increase in wealth we know as economic growth.

When an economy is growing, there is in general demand for workers, and so pay and conditions are encouraged; when an economy is shrinking, there is less demand than supply, and pay and conditions are discouraged.

closewith14• 1 year ago

Overall, over the whole of the economy, the entire population, and a reasonable period of time, this increasing efficiency is a core driver of the annual overall increase in wealth we know as economic growth.

This is only true while wealth inequality is decreasing, which it is not.

wkat42428• 1 year ago

Hmm on the other hand, there isn't much resistance against genAI in software development (unlike other creative industries) because ours is founded in collaboration and continuing others' work. It's where open source came from, and the use of libraries. Using stackoverflow was never frowned on. AI is just the same but more efficient. Nobody invents the wheel from scratch.

It will change the job yes but it also can mean the job can go in new directions because we can do more with less.

bgwalter10• 1 year ago

There isn't much open resistance because most of open source developers are bought and paid for. So they continue the path of destruction in the hopes that they will not be obsolete.

This is naive of course. Once you have identified yourself as corporate servants (like for example the CPython developers) the companies will disrespect you and fire you when convenient (as has happened at Google and Microsoft).

oulipo• 1 year ago

Agreed, although it's slightly beside the point. The goal of building tools and robots has always been to alleviate work. And this is fine. There's still plenty of stuff to do if machines work for us to give us basic housing and food.

Now what needs to be done is to give back the profits to everyone, inclusively, as a kind of "universal basic income", so that we all enjoy it together, and not just the billionaires

picardo• 1 year ago

Sure, AI makes writing code easy, but code review remains a bottleneck. The individual coders are finding ways to use AI to improve their workflows, but as long as organizational culture remains stuck in the old ways of reviewing code, the overall throughput will not improve significantly. That's been my personal experience at Amazon. I get significant pushback when I submit a CR more than a few pages long. You need to have high degrees of trust within a team for this to work at scale, and that's very rarely the case in my group.

iLoveOncall5• 1 year ago

I get significant pushback when I submit a CR more than a few pages long.

Hum yeah, because it's insanely hard to properly review a CR that's more than a few pages long?

nssnsjsjsjs4• 1 year ago

Username checks out. Bio even more so. Don't want to be oncall when a LGTM 1000-liner got dumped on prod.

Anamon4• 1 year ago

Stuck in the old ways of reviewing code? In other words, what you're suggesting should be the new way is to simply accept the LLM's code salad and push it to production as soon as the pipeline lights up green?

rvz2• 1 year ago

Well if the majority of developers are mostly using web-based technologies such as JS, TS, HTML and CSS, and all you're doing is modifying the site and shuffling around elements + a11y addition, etc that is something an AI can do in seconds.

Seems indeed to be like Warehouse work, which is why Web developers will be the first to be affected by AI.

Doesn't matter if you are "senior" or "staff" in Web development. AI is already senior staff level in that.

timkam4• 1 year ago

Is it really that LLM-based tools make developers so much more productive or rather that organizations have found out they can do with less -- and less privileged -- developers? What I don't really see, especially not big tech-internally, are stories of teams that have become amazingly more productive. For now it feels we get some minor productivity improvements that probably do not off-set the invest and are barely enough to keep the narrative alive.

codr73• 1 year ago

[flagged]

Pet_Ant3• 1 year ago

I see it more as replacing shitty code monkeys because it leaves the hard parts behind.

closewith14• 1 year ago

But you of course with your superior skills are above that risk?

dang• 1 year ago

shitty code monkeys

Please don't put others down like that on HN. It's mean and degrades the community.

https://news.ycombinator.com/newsguidelines.html

add-sub-mul-div• 1 year ago

For this narrative to make sense you would have to believe that Amazon management cares more about short-term profit than the long-term quality of their work.

locococo2• 1 year ago

Management has different layers with different goals. A middle manager and a director certainly care a lot about accomplishing short term goals and are ok with tech debt to meet the goals.

layer84• 1 year ago

So it does make sense?

timkam4• 1 year ago

The narrative reflects a broader cultural shift, from "we are all in this together" (pandemic) to "our organizations are bloated and people don't work hard enough" (already pre-LLM hype post-pandemic). The observation that less-skilled people can, with the help of LLMs, take the work of traditionally more-skilled people fits this narrative. In the end, it is about demoting some types of knowledge workers from the skilled class to the working class. Apparently, important people believe that this is a long-term sustainable narrative.

Jeff_Brown3• 1 year ago

Caring is part of it. Having good measures is another. Older measures that worked might need updating to reflect the new, higher spaghetti risk. I expect Amazon to figure it out but I don't see why they necessarily already would have.

locococo2• 1 year ago

A lot of it is perception. Writing software was long considered somewhat difficult and that it required smart people to do so. AI changes this perception and coding starts to be perceived as a low level task that anyone can do easily with augmentation from AI tools. I certainly agree that writing software is turning more into a factory job and is less intellectually rewarding now.

datavirtue• 1 year ago

It's been like this for decades.

cmiles74• 1 year ago

When I started working in the field (1996), I was told that I would receive detailed specs from an analyst that I would then "translate" into code. At that time this idea was already out of fashion, things worked this way for the core business team (COBOL on the AS/400) but in my group (internal tools, Delphi mostly) I would get only the most vague requirements.

Eventually everyone was expected to understand a good deal of the code they were working on. The analyst and the coder became the same person.

I'm deeply skeptical that the kind of people that enjoy software development are the same kind of people that enjoy steering and proofing LLM generated code. Unlike the analyst and the coder, this strike me as a very different skill set.

pjmlp3• 1 year ago

It has been a factory job for decades.

Not everyone gets to code the next ground breaking algorithm at some R&D department.

Most programming tasks are rather repetitive, and in many countries there is hardly anything to look up to software developers, it is another blue collar job.

And in many cultures if you don't go into management after about five years, usually it is seen as a failure to grow up on their career.

catigula2• 1 year ago

What you're describing doesn't sound like something that requires a lot of foreign laborers.

billy99k2• 1 year ago

It's been like this for awhile now. Aside from companies like Google and Facebook, most companies are using some CRUD web app where the development consists of gluing code together for multiple third-party services and libraries.

It's these sorts of jobs that will be replaced by AI and a vibe coder, which will cost much less because you don't need as much experience or expertise.

zkry2• 1 year ago

Even before AI I've always had the perception that writing software felt more intellectually on the level of plumbing. AI just feels like a having one of those fancy new tools that tradespersons may use.

Nasrudith• 1 year ago

Organizations have long had a preference for 'deskilling' to something reliable through bureaucratic procedures, regardless of the side effects or even if it results in it costs more due to needing three people where one talented could do it before. Because it is more dependable, even if it is dependably mediocre. Even though this technique may lead to their long-term doom and irrelevance.

specialist• 1 year ago

Yes and (adjacently):

Seeing Like a State by James Scott

https://en.wikipedia.org/wiki/Seeing_Like_a_State

Explains a lot of the confusing stuff I've experienced, in that eureka sort of way.

xkjyeah• 1 year ago

The number of organizations that continue to use tedious languages like Java 8 and Golang...

Like, they hadn't realized they were turning humans into compilers for abstract concepts, yet now they are telling humans to get tf out of the way of AI

gerdesj3• 1 year ago

Please give some worked examples.

I'm not sure what: "'deskilling' to something reliable through bureaucratic procedures" ... means.

I'm the Managing Director of a small company and I'm pretty sure you are digging at the likes of me (int al) - so what am I doing wrong?

stock_toaster• 1 year ago

I wonder about codebase maintainability over time.

I hypothesize that it takes some period of time for vibe-coding to slowly "bit rot" a complex codebase with abstractions and subtle bugs, slowly making it less robust and more difficult to maintain, and more difficult to add new features/functionality.

So while companies may be seeing what appears to be increases in output _now_, they may be missing the increased drag on features and bugfixes _later_.

greyadept• 1 year ago

I’m concerned that it might not be easy to vibecode a security fix for a complex codebase, especially when the flaw was introduced by vibecoding.

doug_durham• 1 year ago

Up until now large software systems required thousands of hours of work and efforts of bright engineers. We take established code as something to be preserved because it embeds so my knowledge and took so long to develop. If it rots then it takes too long to repair or never gets repaired.

Imagine a future where the prompts become the precious artifact. That we regularly `rm -rf *` the entire code base and regenerate it with the original prompts perhaps when a better model becomes available. We stop fretting about code structure or hygiene because it won't be maintained by developers. Code is written for readability and audibility. So instead of finding the right abstractions that allow the problem to be elegantly implemented the focus is on allowing people to read the code to audit that it does what it says it does. No DSLs just plain readable code.

mrheosuper4• 1 year ago

I wonder whether we have the same talk when the C compiler first came out.

People may worry that the "ASM" codebase will be bit-rot and no one can understand the compiler output or add new feature to the ASM codebase.

aprilthird20214• 1 year ago

Yes, big-tech-internally I also see a lot of desire to get us to come up with some great AI achievements, but they are so far not achieving far far more than already existing automations and bots and code generators can do for us

inadequatespace• 1 year ago

Right. What the article is unsurprisingly glossing over (per usual) is that just because AI is perceived (by higher-ups that don’t actually do the work) to speed up coding work doesn't mean it actually does.

and that probably to some extent all involved (depending on how delusional they are) know that it's simply an excuse to do layoffs (replaced by offshoring) by artificially so-called raising the bar to what is unrealistic for most people

dpflan• 1 year ago

How is the impact of this assessed? How is system quality / performance changed? Is there an increase in high severity defects as more code gets pushed out more quickly?

pbw5• 1 year ago

Companies will always try to capture the productivity gains from a new tool or technique, and then quickly establish it as the new standard for everyone. This is frustrating and feels Sisyphean: it seems like you simply cannot get ahead.

The game is to learn new tools quickly and learn to use them better than most of your peers, then stay quietly a bit ahead. But know you have to keep doing this forever. Or to work for yourself or in an environment where you get the gains, not the employer. But "work for yourself" probably means direct competition with others who are just as expert as you with AI, so that's no panacea.

jampekka3• 1 year ago

Another game is to distribute the gains from increased productivity more equally. E.g. in Europe as late as early 2000s working hours were reduced in response to technological development. But since then the response even from workers seems to be to demand increasingly shittier bullshit jobs to keep people busy.

almostgotcaught4• 1 year ago

Bro lol. You were this close - you're channeling Marx (literally saying the same stuff he was) and instead of coming to the obvious conclusion (unions) you're like nah I'm just gonna alienate myself further. It's just amazing how thoroughly people have been brainwashed. I'm 100% sure nothing will ever improve.

pbw5• 1 year ago

“I'm 100% sure nothing will ever improve.” Nothing? Ever? Brainwashed?

pessimizer2• 1 year ago

you're channeling Marx (literally saying the same stuff he was)

Marx is the originator of precisely none of those thoughts, you couldn't find an economist that disagrees with them. "Unions" is also not the obvious solution for the problems of an individual. Unless you have a specific, existing union with a contact phone number that you're referring to, one that has a track record of making sure that individuals are not affected negatively by technological progress over the span of their entire careers, you're just lazily talking shit.

If it's the solution, so much easier than keeping ahead of the technology treadmill, and it's so obvious to you, it's strange that you haven't set up the One Big Union yet and fixed all the problems.

this15testingg• 1 year ago

right; the "ancap" mentality in computing could only last for so long. Eventually, and especially with the refusal of incorporating any ethics or humanity into it, it's now an established industry affecting all walks of life just like every other that has preceded it, and the belief that its technological superiority/uniqueness was a good reason to essentially exempt it from regulation (TV broadcasts for children are required to have "bumper" sections that would clearly define the show vs the advertisement; Why was computing/the internet treated differently? A high-horse mentality that stemmed from "complexity olympics"? no child could ever use or comprehend a sophisticated machine like this!!) has really fucked us. The labor is decentralized at such a scale that I also have a hard time believing anything could be rectified; open source software is mostly just corporate welfare, putting anything at all on the internet has become corporate welfare, and there is no real purpose or goal for building all of this. The computer was supposed to allow us to do less work, right?

ath3nd14• 1 year ago

But know you have to keep doing this forever. Or to work for yourself or in an environment where you get the gains, not the employer

Or, you know, being a member of society, you can find other members of society who feel like you, and organize together to place demands on employers that...you know...stops them from exploiting you.

- That's how you got the weekend: https://www.bbc.com/worklife/article/20200117-the-modern-phe...

- And that's how you got the 8-hour working week: https://en.wikipedia.org/wiki/Eight-hour_day_movement

- And that's how you got children off the factories: https://en.wikipedia.org/wiki/Child_labour

But, you know, you can always hustle against your fellow SEs, and try to appease your masters. Where others work the bare minimum of 8 hours, why not work 12, and also on the weekend? It's also fine.

Generating shareholder value is very important for the well-being of society! /s

pbw5• 1 year ago

By all means “organize together to place demands on your employers”. I didn’t say don’t do that. But there are 24 hours in a day — maybe strive to be good at your job AND organize instead of doing just one or the other?

wiseowise6• 1 year ago

Not everything is about society.

TheOtherHobbes2• 1 year ago

This can only go three ways.

The first is that the entire global codebase starts to become an unstable shitpile, and eventually critical infrastructure starts collapsing in a kind of self-inflicted Y2k event. Experienced developers will be rehired at astronomical rates to put everything back together, and then everyone will proceed more cautiously. (Perhaps.)

The second is that AI is just about good enough and things muddle along in a not-great-not-terrible way. Dev status and salaries drop slowly, profits increase, reliability and quality are both down, but not enough to cause serious problems.

The third is that the shitpile singularity is avoided because AI gets much better at coding much more quickly, and rapidly becomes smarter than human devs. It gets good enough to create smart specs with a better-than-human understanding of edge cases, strategy, etc, and also good enough to implement clean code from those specs.

If this happens development as we know it would end, because the concept of a codebase would become obsolete. The entire Internet would become dynamic and adaptive, with code being generated in real time as requirements and condition evolve.

I'm sure this will happen eventually, but current LLMs are hilariously short of it.

So for now there's a gap between what CEOs believe is happening - option 3. And what is really happening - option 1.

I think a shitpile singularity is quite likely within a couple of years. But if there's any sane management left it may just about be possible to steer into option 2.

entropi• 1 year ago

I agree with you three scenarios. But I would assign different probabilities. I think the second option is the most likely. Things will get shittier and cheaper. Third option might not ever come to pass.

Just like clothing and textile work. They are getting cheaper and cheaper, true, but even with centuries of automation, they are still getting shittier in the process.

dragandj4• 1 year ago

There are many more scenarios, though. One of them is that AI slop is impressive looking to outsiders, but can't produce anything great on itself, and, after the first wave of increased use based on faith, it just gets tossed in the pile of tools somewhere above UML and Web Services. Something that many people use because "it's the standard" but generally despise because it's crap.

JeremyNT• 1 year ago

The whole thing with gen AI is so depressing to me.

For the first time now I can feel the joy of what I do slipping away from me. I don't even mind my employer capturing more productivity, but I do mind if all the things I love about the job are done by robots instead.

Maybe I'm in the minority but I love writing code! I love writing tests! If I wanted to ask for this stuff to be done for me, I would be a manager!

Now, I'll need to use gen AI to replace the fun part of the job, or I'll be put out to pasture.

It's not a future I look forward to, even if I'm able to keep up and continue working in the industry.

CuriouslyC3• 1 year ago

The fun part for me was coming up with ideas for new things, then architecting those things, creating the high level systems to implement them, and iterating on them to make them better. Figuring out why some test harness wasn't mocking some random thing correctly, remembering which api call had which syntax, or just writing a bajillion almost-boilerplate endpoints was always drudgery and I'm glad to be rid of it.

nssnsjsjsjs4• 1 year ago

The game is live within your means and max out your retirement fund with index funds. Then you own a slither of that production.

This will work until the capitalists realize the stock market let's plebs do well and they'll unlist the best companies.

iLoveOncall5• 1 year ago

I'm a software engineer at Amazon and I'm directly and indirectly involved a lot in programs related to GenAI tooling.

I can safely say this article is bullshit. While there are a lot of programs ongoing to allow builders to adopt GenAI tooling, and while there is definitely a lot of communication ongoing around those programs, nobody is, at all, forced to use any of the AI tools. None are installed by default or enabled by default anywhere, and everyone is free to completely ignore them.

That said, is there an increase in expectations? Yes. But that's just normal Amazon in an employer's market, and has nothing to do with LLMs and GenAI.

The comparison to Microsoft where we can witness in public the .Net maintainers fighting with the shit code generated by Copilot on their repos is ridiculous. Amazon is probably one of the companies pushing the least and being the most prudent about GenAI adoption.

placardloop2• 1 year ago

If you truly believe what you’re saying, then you’re uninformed as to what is going on outside your own team. And just because it’s not happening in your team does not mean it isn’t happening.

Q is installed by default in all browsers on Amazon laptops now, and literally cannot be uninstalled. If you don’t have it installed in your IDE, you get a non-dismissible popup nagging you to install it until you do. Many teams are being told they must use AI every single day (some VPs have sent out org-wide emails saying that AI must be used), and engineers have to tell their managers how they are making use of it day-to-day. In my org, OP1 docs must include at least one section about how the team will increase use of AI. Hackathons aren’t allowed to happen anymore unless they are AI-themed. I could keep going. Amazon is absolutely forcing AI usage, and the article undersells how egregious it is.

iLoveOncall5• 1 year ago

Somehow you're dismissing my post saying it's based on my own anecdotical experience while providing your own anecdotical experience and just hearsay.

There is absolutely no company-wide mandate to use GenAI. If some SDM is pushing it on his SDEs, that's an outlier and on that person alone.

bgwalter10• 1 year ago

Amazon is going down. 20 years ago they had excellent customer service, now they are violating EU laws and cheat the gmp developers out of a CPU replacement:

https://gmplib.org/

Granlund's gcc optimizations probably save Amazon millions in electricity each year. But evidently they don't care about real programmers.

mananaysiempre• 1 year ago

20 years ago they had excellent customer service, now

... you will use them anyway because, customer service or no, there’s a good chance you don’t have a choice that doesn’t cost half again as much. (Regional availability may vary.)

bgwalter10• 1 year ago

For the EU, use this search engine:

https://geizhals.eu/

Amazon is not nearly the cheapest or most reliable one for hardware.

kjkjadksj2• 1 year ago

They aren’t cheaper anymore. The “real” branded stuff is usually more expensive on amazon than straight from the manufacturers website. The “cheap” fake aliexpress stuff is just that charged at a premium. You can find the same exact product images even on the same products on ebay from probably the same merchants listing on both marketplaces.

It has changed dramatically over the past 5 or so years into this.

rwyinuse2• 1 year ago

At least for stuff like electronics, random plastic household items etc EU Amazon isn't particularly cheap. Branded stuff costs about the same in many smaller stores, and random Chinese no-name brands can be bought from Aliexpress cheaper.

baxtr2• 1 year ago

This shift from writing to reading code can make engineers feel as if they are bystanders in their own jobs. The Amazon engineers said that managers have encouraged them to use A.I. to help write one-page memos proposing a solution to a software problem and that the artificial intelligence can now generate a rough draft from scattered thoughts.

This feels like we are forcing people who rather look at code to start talking in plain language, which not every dev likes or is proficient in.

Devs won’t be replaced by AI. Devs will be replaced by people that can (and want to) speak to LLMs.

gortok• 1 year ago

Companies seem to be persuaded that, like assembly lines of old, A.I. can increase productivity. A recent paper by researchers at Microsoft and three universities found that programmers’ use of an A.I. coding assistant called Copilot, which proposes snippets of code that they can accept or reject, increased a key measure of output more than 25 percent.
The engineers said that the company had raised output goals and had become less forgiving about deadlines.

There are two issues this article brings to mind:

1. Feels like we are back when lines of code was a measure of productivity.

2. We’ve heard this tune before, and the problem we have now is that you don’t understand what you didn’t write. Software development is in large part about understanding the system from the highest level to its smallest details, and this takes away a key part of how our world works, in favor of a “cattle not pets” view of code.

Now, if you don’t expect your programmers to have an understanding of the system they built, and you treat code as disposable, then you’ll center around a world where folks aren’t trained to learn a system, and I don’t see that as a world that is compatible with increased reliance on A.I.

Havoc• 1 year ago

Feels like this could be part of a broader shift towards dis-empowering knowledge workers.

The cog in machine effect has always been there in the corporate world, but somehow it feels like the technique has been refined in the last couple of years.

bgwalter10• 1 year ago

The solution is a no-commercial use enhanced GPL license. Corporations are actively using our open source against us, so we have to fight back.

All these narratives about user freedom, for any purpose etc. are just propaganda these days.

giantg2• 1 year ago

I'm seeing the speed up and it's forcing out people with disabilities who are able to do the work at the previous slower speeds. I wonder if there are any solutions to this or if people like me are just expected to be walmart greeters.

osigurdson2• 1 year ago

Why would any company want to pay a software engineer to do repetitive / factory like work? It would be better to automate that and not have any engineers from the company's perspective.

hbartab• 1 year ago

Because engineers train the AI that'll replace them.

osigurdson2• 1 year ago

Probably better just to replace them now if they are just doing simple / repetitive tasks.

telesilla2• 1 year ago

Will AI also start developing creative tools such as new VST plugins or photoshop filters? What about low-level low-latency tools needed for some industries or for aerospace. I guess at some point we won't need so many humans to run our Kubernetes clusters or maintain our WordPress sites but won't there always be something to do that pushes the boundaries of human needs and desires that can't be done by AI?

sidibe• 1 year ago

won't there always be something to do that pushes the boundaries of human needs and desires that can't be done by AI?

No why would there be? Unless you are spiritual, there isn't any reason any of the physical processes that make up human thought can't be done artificially, probably much more efficiently. Society needs to confront the myth that automation is going to always open up more jobs that need human labor. It's comforting for people who hate the idea of UBI or other safety nets that people can keep "retraining". Eventually there's going to be nothing to retrain to (at least of nothing of economic value)

TheOtherHobbes2• 1 year ago

I keep meaning to experiment with vibe-coded VSTs. The shell and UI parts should be easy to automate, because they're basically boilerplate. But I doubt AI knows enough about DSP to design an incredible new reverb algorithm, or is an expert on bandwidth-limited oscillator design.

Or maybe it is? When I have some time, I'll find out.

telesilla2• 1 year ago

I'm sure it could replicate any existing algorithm or riff on combinations, but a good plugin is made carefully by hand by an artisan who knows what they are looking to achieve, who knows what to listen for. I'm skeptical that a pure AI algorithm would be anything but plagiarism on existing designs.

jwitthuhn• 1 year ago

You might be interested in this video, which is about a physical guitar pedal someone made using a chatgpt-based design. Same idea but all analog.

https://www.youtube.com/watch?v=J-PTzq1bv9M

softwaredoug• 1 year ago

I am optimistic longer term, pessimistic near term

What needs to happen is the education of "junior programmers" needs to be revamped to embrace generative AI. In the same way we embraced google or stackoverflow. We're at a weird transition state where the juniors are being taught to code with an abacus, while the industry has moved on to different tools. Generative AI feels taboo in education circles instead of embraced.

Now there will eventually be a generation of coders just "born" into AI, etc, and they will do great in this new ecosystem. Eventually education will catch up. But the cohort currently coming up as juniors will feel the most pain.

xtracto• 1 year ago

You raise a good point. Ba k in the early 90s when i started with programming, my source of knowledge were magazines with code, a "programming with C/C++" book, and a lot of time.

Then the internet came, and it felt like 'cheating'.

Then forums came and it felt like cheating, Then SO, and so on and so forth.

Now AI is eating the [software] world, and to a lot of people, it feels like cheating. I am just amazed of what i can build.

In 10-15 years software will become a commodity, along with books/stories and maybe even music/art. I don't know how it looks like. But darn im excited to be here to experience it.

physicles3• 1 year ago

I was there for those changes too (I taught myself programming a couple years before we got internet), and I disagree that the internet and SO felt like cheating. No matter what resources you had, you could never get very far if you didn’t understand what your code was doing.

That’s no longer true. And that democratizes these skills, which I agree could be a great thing.

But do you agree that it’s important for kids to learn to think critically and systematically? Because it’s super hard to stay motivated to learn those things when LLMs do that for you (and you’re too young to tell when they’re doing a bad job of it).

crabl• 1 year ago

As a potential solution, do you think formal/semi-formal software development education (undergrad programs, colleges/polytechnics, dev bootcamps, etc) should lean super heavily into AI? To the extent that it's not just "use ChatGPT to help you complete this assignment" but rather "complete this assignment using *only* ChatGPT: you're not allowed to write any of the code by yourself".

SoftTalker6• 1 year ago

CS degree programs have never been about learning to code. They are learning about computers, data structures, machine structures, algorithms. The code was always done on your own time at least that's how my school did it. I never had a class in "Java" or "Python" or "C" or any other languge, that was always incidental to the particular course. I could have used ChatGPT (had it existed) or hired a friend to write the code but that wouldn't help me on the exams (written on paper, at that time). Dev bootcamps? Yeah they should probably be leaning hard into AI as that's just what junior devs are going to be asked to do from here on out.

userbinator2• 1 year ago

What needs to happen is the education of "junior programmers" needs to be revamped to embrace generative AI.

No they don't. They need to actually learn how to use their brains first.

_ink_2• 1 year ago

I agree. I always wondered at what point I will feel old and that I can't keep up with technology. Afterall I grew up with Internet. But this might be it.

mrheosuper4• 1 year ago

Generative AI feels taboo in education

Nah, only the teachers feel this, Gen AI is extremely popular in students. In fact, you would look weird if you don't use Gen AI for school work.

What we need is teaching them how to use gen AI effectively.

drysine• 1 year ago

And how to use forklift in a gym to lift weights.

mrweasel• 1 year ago

I'm optimistic long long long term, but long term our industry is screwed.

The current LLM based "AI" isn't good enough and we're already seeing way to many unable to code without the assistance of an AI agent. Sure, many of these people couldn't code at all before, or only very poorly, but at least their output was limited. We're producing way to much code (and to much content in general). The heavy leaning into AI at this point is going to set us back 10 - 15 years, for a short term profit. It's the dotcom bubble all over again in that respect. Way to many unskilled people are producing garbage code, and there aren't enough skilled people around to fix it, because the output volume is to high.

Google recently told employees it would soon hold a companywide hackathon in which one category would be creating A.I. tools that could “enhance their overall daily productivity,” according to an internal announcement. Winning teams will receive $10,000.

If it's really that great, why the competition? Shouldn't this happen pretty organically? Companies are pushing "AI" hard, why to hard, it's not yet there where it can realistically deliver what is expected on the business side. I think even Google developers know this, but hey, $10,000 is $10,000.

I'm very concerned that we eroding trust, safety and quality long term, for a short term profit. It's not that LLMs can't be helpful, save money or improve quality, but you have to be a fairly skilled developer to get those advantages safely.

fzeroracer2• 1 year ago

You should try interviewing a few people fresh out of college. That'll change your opinion incredibly quickly.

vjvjvjvjghv• 1 year ago

With all the changes coming up, I am happy that I am retiring soon. Since I started in the 90s, SW dev has become more and more tightly controlled and feels more like an assembly line. When I started, you could work for weeks and months without much interruption. You had plenty of time for experimentation and creativity. Now everything is ticked based and you constantly have to report status and justify what you are doing. I am sure there will always be devs who are doing interesting work but I feel these opportunities will be less and less.

In a way it's only fair. Automation has made a lot of jobs obsolete or miserable. Software devs are a big contributor to automation so we shouldn't be surprised that we are finally managing to automate our own jobs away,

pbw5• 1 year ago

I also started in the 1990’s and agree the evolution has been as you describe it. It does highly depend on where you work, but the tightly managed JIRA-driven development seems awfully popular.

But I fall short of declaring the 1990s or 2000s or 2010s were the glory days and now things suck. I think part of it is nostalgia bias. I can think of a job I spent 4 years and list all the good parts of the experience. But I suspect I’m forgetting over a lot of mediocre or negative stuff.

At any rate I still like the work today. There are still generally hard challenges that you can overcome, people that depend on you, new technologies to learn about. Generically good stuff.

pbw5• 1 year ago

Another datapoint is working earlier eras sound bad to me: punchcards, assembly, COBOL, FORTRAN. Yes I suspect those people had a blast.

Krei-se2• 1 year ago

Thanks for pointing out JIRA. I think the problem comes from needing to keep the codebase running next month while trying to up scalars / numbers, not thinking years ahead or how to improve both inside culture and outside image of a company which are more complex structures with lots of little metrics and interdependent components than a win/loss output or an issue tracker that ignores the fact issues solved != issues prevented.

I guess these strategies boil down to having some MBA on top or an engineer that has no board of MBAs to bow down to. I strive to stay with private owned companies for this reason but ofc these are less loud on the internet, so you can easily miss them while jobhunting.

dymk• 1 year ago

It’s less automating the job of the engineer away, and replacing management with a panopticon AI

nyarlathotep_3• 1 year ago

Since I started in the 90s, SW dev has become more and more tightly controlled and feels more like an assembly line. When I started, you could work for weeks and months without much interruption. You had plenty of time for experimentation and creativity. Now everything is ticked based and you constantly have to report status and justify what you are doing

Yeah the consistent "reporting" of "status" on "stand-ups" where you say some filler to get someone incapable of understanding what it is that you're doing off your back for 24 more hours has consistently been one of the most useless and unpleasant parts of the job.

neom9• 1 year ago

I tried, but I couldn't get a job as a web dev in the 90s because nobody in my area knew what the web was, I'm not sure anyone in my country did.

mji2• 1 year ago

you say some filler to get someone incapable of understanding what it is that you're doing off your back for 24 more hours has consistently been one of the most useless and unpleasant parts of the job

This sucks for the 50% or so who are like you, but there's another 50% who won't really get much done otherwise, either because they don't know what to do and aren't self-motivated or capable enough to figure it out (common) or because they're actively cheating you and barely working (less common)

georgeecollins2• 1 year ago

> Yeah the consistent "reporting" of "status" on "stand-ups" where you say some filler to get someone incapable of understanding what it is that you're doing off your back

In my experience it is human nature to think you are doing something that people around you can't or don't understand. The graveyard is full of irreplaceable people is an old saying. Sometimes the people you report to are morons, but if you consistently report to a moron its time for introspection. There's more that you can do than just suffer that. One place to start is to have some charity for the people you work with.

Krei-se2• 1 year ago

Knowing Amazons backends this is just their technical debt creeping up. Coders do repetetive work there because extending from their architecture means lots of code simply to keep the machine running. I would be surprised if they did not assign ASIN (which come from the need to keep ISBN compatibility everywhere) to not only Apps but other virtual goods lol.

From my integrations pov Ebay was ahead of their time with their data structure and pushed for deprecation fast to not keep the debt. Amazon ooth only looks more modern through acquiring new market fields instantly followed by throwing a lot of money to facade up the mess. Every contact like key account managers there were usually pushed for numbers, this has nothing to do with coders being coders.

Bosses always look for ways to instantly measure coders output which is just short-sighted way of thinking. My coworkers were measured by lines of code obviously. I wonder how you measure great engineering.

So no, this has not changed, you can still work uninterrupted on stuff for months or years if you want and skip these places, maybe even proved over your career that previous designs are stable for years to come.

dominicrose• 1 year ago

From what I've seen, devs though that COVID = full remote forever. It's no wonder there's more control now. So it started before AI IMO.

bamboozled2• 1 year ago

Now everything is ticked based and you constantly have to report status and justify what you are doing.

The weird thing about this is, many developers wanted this. They wanted the theater of Agile, JIRA tickets, points etc.

I'm in the same boat yet still need to squeeze out another 10 years or so but personally working om multiple-side projects so I can get out of this boring, mundane shit.

georgeecollins2• 1 year ago

I think what you are describing is the difference between being among the people to engineer the first car to being in the factory trying to make the 100 millionth car as cheaply as possible. Forty years ago people mostly learned to program because they were interested in it. People starting software / tech companies were taking a chance on something most didn't understand. That self selected for a type. Now its the most common major in school. I am sure its possible to recreate what you felt in the '90s but probably in a different field or a subset of this one.

elzbardico2• 1 year ago

The plague of the two-week sprint.

curiousgal• 1 year ago

One Amazon engineer said his team was roughly half the size it had been last year, but it was expected to produce roughly the same amount of code by using A.I.

I am sick of these verbose articles that boil down to nothing basically. What the f does it mean to "produce code"? Like are we just churning out LoCs daily just for the sake of doing so?

ManBeardPc3• 1 year ago

I feel my job in the future will be more secure than ever. Tons and tons of AI generated garbage code (trained on more and more existing garbage code) that the „developer“ will at a certain point no longer be able to maintain or fix. Not even speaking about trusting the output. Feels similar to all the outsourced development to cheap suppliers that inevitably collapse or create horrible maintenance overhead.

nlitened• 1 year ago

Do you feel you would be able to “maintain and fix” LLM-generated 100-megabyte source code blobs? And if you could, do you think it would be a job you’d want to do?

ManBeardPc3• 1 year ago

No, writing a replacement. Either part by part or a whole new software system.

petesergeant• 1 year ago

I spent 2000-2015ish as a Perl programmer, which meant I spent most of that time working for a small number of e-commerce and SaaS companies who had large codebases that had mostly been written around 2005 with some hair-raising tech practices. Nursing a nasty codebase to health bit by bit can be very rewarding.

WD-42• 1 year ago

Secure, sure. But this sounds like a crappy job.

ManBeardPc3• 1 year ago

Not so crappy if the pain is big enough for the customer. Once a certain threshold is reached they are often very open to changes and giving freedom to develop a proper solution.

talles2• 1 year ago

It amazes me how immature our field can be. Anyone that worked for big corporations and in humongous codebases know how 'generating new code' is a small part of the job.

AI blew up and suddenly I'm seeing seasoned people talking about KLOC like in the 90s.

SonOfKyuss3• 1 year ago

Agreed. I spend maybe 20% of my time writing code. The rest is gathering requirements, design, testing, scheduling. Maybe if that 20% now takes half as long, I might have time to actually write some tests and documentation

mk89• 1 year ago

Yes, because from my observations these are the people who cannot write anymore code today - they really don't understand the newer paradigms or technologies. So Copilot enabled a lot of people to write POCs and ship them to production fast, without thinking too much about edge cases, HA, etc.

And these people have become advocates in their respective companies, so that everyone is actually following inaccurate claims about productivity improvements. These are the same people quoting Google's ceo who say that 30% of newly generated code at Google is written by AI, without possibility to deny or validate it. Just blindly quote a company that has a huge conflict of interests in this field and you'll look smarter than you are.

This is where we're at today. I understand these are great tools, but all I am seeing madness around. And who works with these tools on a daily basis knows it. Knows what it means, how they can be misleading, etc.

But hey, everyone says we must use them ...

Shocka12• 1 year ago

Not that I disagree, but Anthropic put out some usage statistics for their products a few months ago and IIRC, something like 40% was used for software engineering.

maxnevermind2• 1 year ago

It amazes me how immature our field can be. Anyone that worked for big corporations and in humongous codebases know how 'generating new code' is a small part of the job.

Exactly, but I would go further, anyone who worked in big corps know that other non 'generating new code' part is usually pretty inefficient and I would argue AI is going to change that too. So there will be much less of that abstract yapping in endless meetings or there will be less people involved in that.

bigfatkitten3• 1 year ago

If that yapping went away, what are mid to senior managers going to do all day?

mji2• 1 year ago

'generating new code' is a small part of the job

I think this attitude has been taken too far to the point that people (especially senior+ engineers) end up spending massive amounts of time debating and aligning things that can just be done in far less time (especially with AI, but even without it). And these big companies need to change that if they want to get their productivity back. From the article:

One engineer said that building a feature for the website used to take a few weeks; now it must frequently be done within a few days. He said this is possible only by using A.I. to help automate the coding and by cutting down on meetings with colleagues to solicit feedback and explore alternative ideas."

wiseowise6• 1 year ago

and by cutting down on meetings with colleagues to solicit feedback and explore alternative ideas.

At least something good comes out of this.

guax3• 1 year ago

What used to take a week now can be done in just 5 days.

Altern4tiveAcc• 1 year ago

How is the shorter deadline better for the worker? Ultimately, that devolves to a race to the bottom with people choosing between overworking or being laid off. Surely AWS is profitable enough by now, that all those employees could get their work hours reduced, receive a raise, and have both the organization and the product keep existing just fine.

CuriouslyC3• 1 year ago

We don't have meaningful metrics that are objective. KLOC is bad but at least it means something concrete, issues closed or PRs submitted is essentially meaningless. This is a software engineering problem not an AI problem.

talles2• 1 year ago

KLOC is as objective as your other examples and as meaningless as well IMO.

julkali• 1 year ago

I think there is a fundamental misconception of the benefit / performance-improvement of LLM-aided programming:

Without sacrificing code quality, it only makes coding more productive _if you already know_ what you're doing.

This means that while it has a big potential for experienced programmers (making them push out more good code), you cannot replace them by an army of code monkeys with LLMs and expect good software.

neom9• 1 year ago

I keep reading this but I feel it really ignores gaussian, where are your lines? What is good enough for what where? What is the base level of already know? I'm churning out a web app for fun right now with a couple of second year comp sci students from Sri Lanka + LLMs, they charge me around $1000 a month and my friend who is a SRE at appl looks at the code every week, said it's quality, modern and scalable. I do think they're a bit slow, but I'm not looking for fast.

philipwhiuk• 1 year ago

You're paying $1000 a month to build a web app for fun?

This seems like a crazy solution to a situation.

nyarlathotep_3• 1 year ago

you cannot replace them by an army of code monkeys with LLMs and expect good software.

"Good" software only matters in narrow use cases--look at how much money and how many contracts companies like Deloitte and Accenture make/have.

Sure, you can't "vibe" slop your way to a renderer for a AAA title, but the majority of F500s have no conception of quality and do not care nor know any "better."

tom_m• 1 year ago

Hasn't that always been the case at Amazon?

levocardia• 1 year ago

Indeed, I have heard that Amazon is a grindhouse many times before. It's understandable, in a sense: unlike other big tech companies, Amazon lives and dies on low margins. They're more like Walmart than Google, in terms of the business model at least. So they need to push devs hard and get lots of productivity out of them.

neom9• 1 year ago

From the business side of the house, I only ever look at it as what can do the right job for the right money to very quickly deliver continuous value to customers. In an employee base: Experience around is always good for growing and improving my people, but it isn't necessarily tied directly to the work to be done. The shift I feel is happening is how I place the value here. Where I can imagine I would land if I was actively running a startup right now is making sure I have amazing production teams (whatever that looks like, age, location, LLM powered or otherwise) book ended by great people growers as all cogs start to squeak if they're not oiled, and experience is the best lubricant on teams, you need both.

bookofjoe• 1 year ago

I note less hostility to AI/LLM etc in the comments on this post than in previous months. Is the needle shifting?

bgwalter10• 1 year ago

There were many more throwaway accounts before (I think) the new moderation efforts. Many people cannot speak freely on this topic.

000ooo000• 1 year ago

Can't think of a post I've seen that was in any way dubious or critical of AI that wasn't also flagged. HN is heavily biased toward AI.

QuiEgo• 1 year ago

Amazon’s entry level SDEs are hired to answer pages and grind sev2s at 3am. AI seems set to make that worse, not better.

tushar-r• 1 year ago

20 years ago, when I first started working, Wipro and HCL Tech had been talking about making the outsourcing model into "factory-like assembly lines." This fizzled out very quickly back then, but I guess it is now coming to fruition :-|

overgard6• 1 year ago

You know, the point of AI in coding seems to be so that you can code in English, instead of a formal language. (And for some reason we're pretending like these formalisms are hard and the people that understand them evil gatekeepers, to which I'd say: nope to both propositions. It has never been easier to learn how to code.)

The thing I don't understand is why anyone thinks this is an improvement. I think anyone that's written code knows that writing code is a lot more fun than reading code, but for some reason we're delegating to the AI the actual enjoyable task and turning ourselves into glorified code reviewers. Except we don't actually review the code, it seems, we just go on with bug ridden monstrosities.

I fail to see why anyone would want this to be the future of code. I don't want my job to be reviewing LLM slop to find hallucinations and security vulnerabilities, only to try to tweak the 20,000 word salad prompt.

HPsquared• 1 year ago

I personally find debugging and fixing code to be the most rewarding.

That way you know you're (usually) strictly making an improvement.

overgard6• 1 year ago

I know I'm making an improvement by features I've shipped to customers. Bugs prevent me from shipping more features until they're fixed. I do not enjoy bugs.

cj• 1 year ago

I fail to see why anyone would want this to be the future of code.

IME there's an inverse relationship between how excited the person is about AI coding and their seniority as an engineer.

Juniors (and non-coders) love it because now they can suddenly do things they didn't know how to do before.

Seniors don't love it because it gets in their way, causes coworkers to put up low-quality AI-generated code for peer review, and struggles with any level of complexity or nuance.

My fear is that AI code assistants will inadvertently stop people from progressing from Junior --> Senior since it's easier to put out work without really understanding what you're doing. Although I guess I could have said the same thing about Stack Overflow 10 years ago.

Anamon4• 1 year ago

It's a very different thing from SO, in my opinion, because SO involves human discussion. One of the most valuable, educational things about using SO is the comments underneath some answers saying that it's a bad idea, and explaining why. This critical reflection is completely missing from LLM coding. Good suggestion, horrible suggestion – there's no distinction that would be noticed by someone not already experienced in the topic.

jevndev• 1 year ago

I think the key to understanding why people want this is that those people care about results more than the act of coding. The easy example for this is a corporation. If the software does what was said on the product pitch, it doesn’t matter if the developer had fun writing it. All that matters is that it was done in an efficient enough (either by money or time) manner.

A slightly less bleak example is data analysis. When I am analyzing some dataset for work or home, being able to skip over the “rote” parts of the work is invaluable. Examples off the top of my head being: when the data isn’t in quite the right structure, or I want to add a new element to a plot that’s not trivial. It still has to be done with discipline and in a way that you can be confident in the results. I’ll generally lock down each code generation to only doing small subproblems with clearly defined boundaries. That generally helps reduce hallucinations, makes it easier to write tests if applicable and makes it easier to audit the code myself.

All of that said, I want to make clear that I agree that your vision of software engineering Becoming LLM code review hell sounds like… well, hell. I’m in no way advocating that the software engineering industry should become that. Just wanted to throw in my two cents

overgard6• 1 year ago

If you care about the results you have to care about the craft, full stop.

Wowfunhappy3• 1 year ago

(And for some reason we're pretending like these formalisms are hard and the people that understand them evil gatekeepers, to which I'd say: nope to both propositions. It has never been easier to learn how to code.)

I am not a professional programmer, and I tend to bounce between lots of different languages depending on the problem I'm trying to solve.

If I had all of the syntax for a given language memorized, I can imagine how an LLM might not save me that much time. (It would still be helpful for e.g. tracking down bugs, because I can describe a problem and e.g. ask the AI to take a first pass through the codebase and give me an idea of where to look.)

However, I don't have the syntax memorized! Give me some Python code and I can probably read it, but ask me to write some code from scratch, and LLMs I would have needed to dive into the language documentation or search Stack Overflow. LLMs, and Claude Code in particular, have probably 10x'd what I am capable of, because I can describe the function I want and have the machine figure out the minutia of syntax. Afterwards, I can read what the produced and either (A) ask it to change something specific or (B) edit the code by hand.

I also do find writing code to be less enjoyable than reading/editing code, for the reason described above.

skydhash6• 1 year ago

No one have the language syntax memorized, unless you're working with the language daily. Instead we store patterns, and there isn't a lot (checkout the formal grammar for any programming language). For any C like language, they overlap a lot, the difference are mostly in syntax minutia (which we can refresh in an afternoon with a reference) and the higher abstractions (which you learn once, like OOP, pattern matching).

Generally you spend 80% of the time wrangling abstractions, especially in mature project. The coding part is often a quick mental break where you just doing translation. Checking the syntax is often a quick action that no one mind.

overgard6• 1 year ago

I mean, I'm not a mathematician but I don't expect that I should be able to write a proof.

You talk of memorizing syntax like its a challenge but excluding a small number of advanced languages no programmer thinks syntax is hard. But if you don't understand the basics how can you expect to be able to understand if the solution an LLM presents is decent and not rife with bugs (security and otherwise)

I guess my issue is people are confusing a shortcut with actually being able to do the thing. If you can't remember syntax I don't really want your code anywhere I care about

overgard6• 1 year ago

This might be a harsh thing to say, but I don't think the talent at Amazon is top notch because (from everything I've heard anyway) Amazon is such a nightmarishly dystopian place to work that, well, if you have a choice you probably avoid working there? I say that with the recognition that there are always diamonds in the rough, but I'm not sure this really comments on AI usage at a place that treats their employees well.

(That said I'm all for more dystopian stories so we can get past this AI-replacing-coders fad)

nagisa• 1 year ago

Amazon pays really well. Amazon also has a number of interesting projects (e.g. a number of contributors to the rust compiler are employed by Amazon to work on rustc.) It also looks nice enough on a resume to give people a nice stepping stone towards even better opportunities.

lazystar2• 1 year ago

Amazon pays really well

maybe 10 years ago. HR is like a rabid dog, fighting for every dollar.

newAccount2025• 1 year ago

I really like a lot about Amazon. Friendly and creative people, good projects with high ambiguity and autonomy. The bad parts are the penny-pinching, careless and unaccountable senior leaders who just wants to exploit everyone and everything. But really, aside from things like RTO, they aren’t too impactful on the day to day.

raxxorraxor• 1 year ago

There are probably some spots with interesting engineering problems or projects. But overall I agree, I don't want to work for big tech.

As for developing becoming more factory-like. Not sure about that, you have the same leverage as a developer and you can just work for the other guy.

OptionOfT• 1 year ago

To me it feels like part of the hype train, like crypto & VR.

I recently had the (dis)pleasure of fixing a bug in a codebase that was vibe coded.

It ends up being a collection of disorganized business problems converted into code, without any kind of structure.

Refinements are implemented as super-narrow patches, resulting in complex and unorganized code, whereas a human developer might take a step back to try and extract more common patterns.

And once you reach the limit of the context window you're essentially suck, as the LLM can no longer keep track of its patches.

English (or all spoken human language) is not precise enough to articulate what you want your code to do, and more importantly, a lot of time and experience precedes code that a senior developer writes.

If you want to have this senior developer 'vibe' code, then you'll need to have a way to be more precise in your prompts, and be able to articulate all learnings from your past mistakes and experience.

And that is incredibly heavy. Remember, this is opposite from answering 'why did you write it like this'. This is an endless list of items that say 'don't do this, but this, in this highly specific context'.

torben-friis4• 1 year ago

Counterpoint: AI has help me refactor things where I normally couldn’t. Things like extracting some common structure that’s present in a slightly different way in 30 places, where cursor detects it, or suggesting potential for a certain pattern.

The problem with vibe coding is more behavioral I think: the person more likely to jump in the bandwagon to avoid writing some code themselves is probably not the one thinking about long term architecture and craftsmanship. It’s a laziness enhancer.

codr73• 1 year ago

And once AI is not there anymore, you're back on square one.

It only looks effective if you remove learning from the equation.

It's the wrong tool for the job, that's what it is.

greesil2• 1 year ago

What do you think is going to happen when management is pushing on deadlines?

podviaznikov2• 1 year ago

100% this. I did this many times too. Often I wouldn’t bother with cleanup or refactor before, but now it’s way easier, faster and cheaper to do it.

And it’s better in the long run.

winrid3• 1 year ago

I have the same observation. I've been able to improve things I just didn't have the energy to do for a while. But if you're gonna be lazy, it will multiply the bad.

disqard• 1 year ago

I'll quote one of the wisest people in tech I know of, Kentaro Toyama:

"Technology acts as an amplifier of human intentions"

...so, if someone is just doing a sloppy job, AI-assisted vibe coding will enable them to do it faster.

makeitdouble• 1 year ago

AI has help me refactor things where I normally couldn’t.

Reading "couldn't" as, you would technically not be able to do it because of the complexity or intricacy of the problem, how did you guarantee that the change offered by the AI made proper sense and didn't leave out critical patterns that were too complex for you to detect ?

Your comment makes it sound like you're now dependent on AI to refactor again if dire consequences are detected way down the line (in a few months for instance), and the problem space is already just not graspable by a mere human. Which sounds really bad if that's the case.

adonese• 1 year ago

Is cursor _that_ good on monorepo? My use with AI so far has been the chat interface. I provide a clear description of what i want and manually copy paste it. Using copilot, I couldn't buy their agentic mode nor adding files to the chat' session context. Gemini' large context has been really good handling large context, but still doesn't help much in refactoring.

throwaway20375• 1 year ago

    > Counterpoint: AI has help me refactor things where I normally couldn’t. Things like extracting some common structure that’s present in a slightly different way in 30 places, where cursor detects it, or suggesting potential for a certain pattern.

I have a feeling this post is going to get a lot of backlash, but I think this is a very good counterpoint. To be clear: I am not here to shill for LLMs nor vibe coding. This is a good example where an "all seeing" LLM can be helpful. Whether or not you choose to accept the recommendation from the LLM isn't my main point. The LLM making you aware is the key value here.

Recently, I was listening to a podcast about realistic real world uses for an LLM. One of them was a law firm trying to review details of a case to determine a strategy. One of podcasters (sp?) recoiled in horror: "An LLM is writing your briefs?" They replied: "No, no. We use it generate ideas. Then, we select best." It was experts (lawyers, in this case) using an LLM as a tool.

bamboozled2• 1 year ago

I've found it can help one get the confidence to get started, but there are limits and there is a point where a lack of domain knowledge and a desired target (sane) architecture will bight you hard.

You can have AI almost generated anything, but even AI has limited to understanding requirements, if you cannot articulate what you want very precisely, it's difficult to get "AI" to help you with that.

jajko2• 1 year ago

If you couldnt do the task youself (a very loaded statement which I honestly dont't believe), how could you even validate what llm did was correct, didnt miss anything, didnt introduce a nasty corner case bug etc?

In any case a very rare and specific corner case you mention, a dev can go on a decade or two (or lifetime or two) without ever experiencing similar requirement. If it should be a convincing argument for almighty llm it certainly isnt.

larodi3• 1 year ago

well, in all honesty, a lot of code that actually works in production, and delivers... goods, services, etc. is an over patched mess of code snippets. that sad part is that cpus are so powerful that it works.

mystified5016• 1 year ago

Yeah, this isn't new, but it certainly seems worse now

amrocha• 1 year ago

This argument ignores the fact that AI makes it so much easier to ship a ton of garbage.

It’s like comparing a machine gun to a matchlock pistol and saying they’re the same thing.

mullingitover• 1 year ago

Software is a gas that expands to fill its container.

noiv• 1 year ago

I feel you and would personally pay a bit more for LLMs trained on 'senior code'.

ModernMech8• 1 year ago

Yeah, I feel like a lot of the code they spit out is tutorial-level code you find on beginner websites.

Flemlo6• 1 year ago

A hype is over promising and under delivering.

AI is extremely new, impressive and is already changing things.

How can that be in any way be similar to crypto and VR?

gavinward• 1 year ago

Agreed. I can't think of another area where AI could amplify the effect of my existing level of knowledge as it does with coding. It's far exceeding my expectations.

ModernMech8• 1 year ago

Yes, it's new and impressive and changing things, but nevertheless it's still underdelivering because the over-promising is out of control. They are selling these things as Ph.D. level researchers that can ace the SAT, pass the MCAT, pass the Bar, yet it still has trouble counting the R's in "strawberry".

malfist2• 1 year ago

You can see that in this article referencing Jassy's 4,500 years of effort. Which is said "sounds crazy but it's true!"

It isn't true.

The tool he's bragging about most went about changing JDK8 to JDK17 in a build config file, and if you're lucky tweaking log4j versions. 4,500 years my ass. It was more regex and AI

bigstrat2003• 1 year ago

A hype is over promising and under delivering.

Indeed, which is exactly what's happening with AI.

ModernMech8• 1 year ago

English (or all spoken human language) is not precise enough to articulate what you want your code to do

Exactly. And this is why I feel like we are going to go full circle on this. We've seen this cycle in our industry a couple times now:

"Formal languages are hard, wouldn't it be great if we could just talk in English and the computer would understand what we mean?" -> "Natural languages are ambiguous and not precise, wouldn't it be great if we could use a formal langue so that the computer can understand precisely what we mean?"

The eternal hope is that someday, somehow, we will be able to invent a natural language way of communicating something precise like a program, and it's just not going to happen.

Why do I think this? Because we can't even use natural language to communicate unambiguously between intelligent people. Our most earnest attempt at this, the law, is so fraught with ambiguity there's an entire profession dedicated to arguing in the gray area. So what hope do we have controlling machines precisely in this way? Are future developers destined to be equivalent to lawyers, who have to essentially debate the meaning of a program before it's compiled, just to resolve the ambiguities? If that's where this ends up, I will be very sad indeed.

Nathanba• 1 year ago

it doesn't have to be precise enough anymore though because a perfect AI can get all the details from the existing code and understand your intent from your prompt.

jazzyjackson2• 1 year ago

I hear formal Sanskrit has precise logical semantics

CapricornNoble• 1 year ago

> The eternal hope is that someday, somehow, we will be able to invent a natural language way of communicating something precise like a program, and it's just not going to happen.

I think the closest might be the constructed language "Ithkuil". Learning it is....difficult, to put it mildly.

https://ithkuil.net/

RossBencina4• 1 year ago

Are future developers destined to be equivalent to lawyers, who have to essentially debate the meaning of a program before it's compiled, just to resolve the ambiguities?

Future developers? You sound like you've never programmed in C++.

bwfan1234• 1 year ago

The eternal hope is that someday, somehow, we will be able to invent a natural language way of communicating something precise like a program, and it's just not going to happen

My take is more nuanced.

First, there is some evidence [1] that human language is neither necessary nor sufficient to enable what we experience as "thinking".

Second, our intuition, thinking etc are communicated via natural languages and imagery which form the basis for topics in the humanities.

Third, from communication via natural language - slowly emerges symbolism, and formalism which codifies intuitions in a manner which is operational, and useful.

As an example, socratic dialog was a precursor to euclidean geometry which operationally codifies our intuitions of space around us in a manner which becomes useful.

However, formalism is stale as there are always new worlds we experience which cannot be captured by any formalism. The genius of the human brain which is not yet captured in LLMs is to be able create symbolisms of these worlds almost on demand.

ie, if we were to order in terms of expressive power, it would be something like:

1) perception, cognition, thinking, imagination 2) human language 3) formal languages and computers codifying worlds experienced via 1) and 2)

Meanwhile, there is a provocative hypothesis [2] which argues that our "thinking" process lies outside computation as we know it.

[1] https://www.nature.com/articles/s41586-024-07522-w [2] https://www.amazon.com/Emperors-New-Mind-Concerning-Computer... [3] https://www.cs.utexas.edu/~EWD/transcriptions/EWD06xx/EWD667...

bwfan1234• 1 year ago

The eternal hope is that someday, somehow, we will be able to invent a natural language way of communicating something precise like a program, and it's just not going to happen

what we operationally mean by "precise" involves formalism. ie, there is an inherent contradiction between precise, and natural languages.

bdangubic• 1 year ago

It ends up being a collection of disorganized business problems converted into code, without any kind of structure.

this matches the description of every codebase (except one) I came across in my 30-year career

salomonk_mur• 1 year ago

Then you had some very shitty jobs. Good teams put time and effort in architecture before development.

LLMs will produce code with good structure, as long as you provide that architecture before hand.

maeln• 1 year ago

While I have a similar experience with, hurm, "legacy" codebase, I gotta say, LLM (in my experience) made the "legacification" of the codebase way, way faster.

One thing especially, is the loss of knowledge about the codebase. While there was always some stackoverflow-coding, when seeing a weird / complicated piece of code, I used to be able to ask the author why it was like that. Now, I sometimes get the answer "idk, its what chatgpt gave me".

gitaarik• 1 year ago

But that is for different reasons. There are definitely human created codebases that are clean, that no AI could match.

throwaway20375• 1 year ago

    > except one

Please share more!

tome2• 1 year ago

I call this the Peter Principle of Programming:

"Code increases in complication to the first level where it is too complicated to understand. It then hovers around this level of complexity as developers fear to touch it, pecking away here and there to add needed features."

http://h2.jaguarpaw.co.uk/posts/peter-principle/

Seems like I'm not the first one to notice this either:

https://nigeltao.github.io/blog/2021/json-with-commas-commen...

jmyeet2• 1 year ago

You're not wrong that AI is hyped just like crypto and VR were. But it's also true that automation will increasingly impact your job, even for jobs that we have considered to be highly technical like software engineering.

I've noticed this over the last decade where tech people (of which I am one) have considered themselves above the problems of ordinary workers such as just affording to live. I really started to notice this in the lead up to the 2016 election where many privileged people did not recognize or just immediately dismissed the genuine anger and plight of working people.

This dovetails into the myth of meritocracy and the view that not having enough money or lacking basic necessities like food or shelter or a personal, moral failure and not a systemic problem.

Tech people in the 2010s were incredibly privileged. Earnings kept going up. There was seemingly infinite demand for our services. Life was in many ways great. The pandemic was the opportunity for employers to reign in runaway (from their perspective) labor costs.

Permanent layoff culture is nothing more than suppressesing wages. The facade of the warm, fuzzy Big Tech employer is long gone. They are defense contractors now. Google, Microsoft or Amazon are indistinguishable from Boeing, Lockheed Martin and Northrop Grumman.

So AI won't immediately replace you. It'll start by 4 engineers with AI being able to do the job that was previously done by 5. Laying off that one person saves that money directly but also suppresses the wages of the other 4 who won't be asking for raises. They're too afraid of losing their jobs. Then it'll be 3. Then 2.

A lot of people, particularly here on HN, are going to find out just how replaceable they are and how aligning with the interests of the very wealthiest was a huge mistake. You might get paid $500K+ a year but you are still a worker. Your interests align with nurses, teachers, baristas, fast food workers and truck drivers, not the Peter Thiels of the world.

sham12• 1 year ago

There have always been segments of the working class, which have deluded themselves into believing that by co-operating with the capitalists, they could shield themselves from the adverse effects of what is happening to the rest of the working class.

And it's an understandable impulse, but at some point you'd think people would learn instead of being mesmerized by the promise of slightly better treatment by the higher classes in exchange for pushing down the rest of the working class.

Now it's our turn as software engineers to swallow that bitter pill.

aprilthird20214• 1 year ago

I think engineers are more like doctors / lawyers, who are also both contracted labor, whose wages can be (and have been) suppressed as automations and tactics to suppress wages arrived.

But these groups also don't have strong unions and generally don't have the class consciousness you are talking about, especially as the pay increases.

guax3• 1 year ago

Whoosh noises, flashback bells...

In the future no one will have to code. Well compile the business case from UML diagrams!

wesselbindt• 1 year ago

One simile I've heard describing the situation where fancy autocomplete can no longer keep track of its patches is that you'll be sloshing back and forth between bugs. I thought it was quite poetic.

major5052• 1 year ago

In 20 yeas in this market I saw a lot of this. A about 6 years back it was the blockchain crazy.

My boss wanted me to put blockchain in everything (so he could market this to our clients). I printed a small sign and left on my desk. Everytime someone asked me about blockchain, I would point the sign "We don't need blockchain!"

wg0• 1 year ago

This is professionalism. Appeasement is dishonest.

ryandrake2• 1 year ago

Every 5 years or so, there's some new thing that every uncreative product lead decides they absolutely must have, for no other reason than that it's the "in" thing. Web is hot. We need to put our product on the web. Mobile is hot. We need a mobile app. 3D is hot. We need to put 3D in our product. VR is hot. We need to put VR in our product. IoT is hot. We need to put IoT in our product. Blockchain is hot. We need to put blockchain in our product. AI is hot. We need to put AI in our product. It'll keep going on long after we're out of the business.

noobermin• 1 year ago

If you have to be so precise in your prompts to the point you're almost using a well specified, domain-specific language, you might as well program again because that's effectively the same thing.

jb3689• 1 year ago

Makes me wonder if we’ll see more emphasis on loosely coupled architecture as a result of this. Software engineers maintain the structure, and AI codes it chaos at the leaf. Similar to how data engineers commoditized data via the warehouse

rwyinuse2• 1 year ago

I'm wondering whether most other white collar jobs will turn like this as well. I see lots of articles that have been at least partially written by AI - so there goes most of your average journalist's job. Then there's stuff like marketing and communications, where your output is mostly text and media. Again, something AI can mostly handle based on initial requirements.

Personally I find babysitting AI quite boring. It's easy to just stop caring about the quality of the output when one just is wage slaving it, and the process itself is no longer satisfying.

tehjoker6• 1 year ago

This article features a study claiming 25% productivity gains form co-pilot. I wonder what a similar study would say about developers with and without access to internet search. In the study, it seems, as I expected, less experienced people found more value from it. My experience has been AI tools are more helpful when I don't know the domain, and then their benefit declines as I familiarize.

I also looked at the study and noticed a few aspects that were surprising:

(a) some of the 95% CIs crossed zero, meaning no benefit is a possible interpretation in figures 6 and 7

(b) did anyone account for what happens when two workers are in different experimental groups and sit next to each other? I imagine it was likely common people in the experimental group to use co-pilot queries for their friends.

There are a lot of other subtleties in the interpretation of this paper.

"...developers who were less productive before the experiment are significantly more likely to accept any given suggestion...."

...curious lol

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4945566

zeroonetwothree7• 1 year ago

Yeah out of all the studies I’ve seen about actual productivity gains from LLMs the median estimate is probably in the low single digits (possibly including 0). Not exactly earth shattering yet.

neilv3• 1 year ago

You can speed up software development dramatically by simply copying the code that someone else was foolish/idealistic enough to store on GitHub, stripping out the copyright and licensing notices, and adapting it to your needs

The problem with that is that you don't want to get caught doing that directly.

So you need to hire a sleazy offshore firm to launder that for you.

Or (faster and cheaper) use a sleazy "AI" to launder it.

Google had too much class (and comfortable market position) to launch the sleazy "AI" gold rush.

But plenty of upstarts currently see an irresistible opportunity to get rich, by leveraging automated mass copyright violation, while handwaving "AI".

tkzed492• 1 year ago

I'm curious what industry you work in. Rarely have I run into a large problem that was tempting to solve with Stackoverflow/Github-driven development. I find the limiting factor is never "we don't have the code" and more "we don't know what code we need".

neilv3• 1 year ago

I certainly agree that there are many important aspects to software development, other than just code.

But I don't want to confuse the point, and accidentally be thinking "plagiarism of code isn't important, because code isn't important" (or the limiting factor).

For one reason, code is important, and valuable...

There's plenty of evidence throughout many types of software development to suggest that a lot of money has to be plowed into simply producing bulk of code -- excluding domain or market understanding, requirements analysis, holistic system design, etc.

A typical growth startup wasn't in a hiring frenzy for great analytical minds, but simply to scale up production of raw bulk of code (and connecting together legally-obtained off-the-shelf pieces). And they were often tracking quantitative metrics on that bulk, and consciously aware of how much all those code-monkey salaries were costing on the balance sheet. They paid it because code bulk was very necessary, the way they were working.

There's also plenty of evidence of the popularity of copy&paste reuse for many years. Including from StackOverflow.

Copying entire projects and subsystems has been much-much less common, partly due to the aforementioned not wanting to get caught doing it. But "AI" laundering is paving the way. See all the blog posts and videos about "I created an app/game/site in an hour with AI".

We can also look at startup schools of thought, where execution is widely regarded to be everything (ideas are a dime a dozen). There is plenty of thinking that churning out code is one of the most time-consuming or frequently-blocking parts of execution.

In both startups and established corporate environments, there are normally big lags between when a need for specific code is identified, and when that code can be delivered. Doesn't that look like a limiting factor, or at least it must be important.

I think that's enough reason that we not dismiss the value of code, nor dismiss the importance of plagiarism of code.

catigula2• 1 year ago

Yep.

ChatGPT was hugely irresponsible in many ways and rests solely on the shoulders of Sam Altman, someone allegedly implicated in various schemes.

AstroBen• 1 year ago

Google had too much class? They've been scraping sites content for years before all this LLM stuff - using it to answer questions directly in search results and sending no clicks

No, I think Google was just slow this time

neilv3• 1 year ago

I think Google has ethics in their self-image. Albeit more flexibility in practice.

But decisions might've come down to what keeps the ad revenue machine fed. Or it might not have made it to decisions.

If they were slow, wasn't it only in being prepared for when the boat got rocked? Until the boat got rocked, they had a monopoly, so why rock it themselves?

I wouldn't write in my promotion packet, "Disrupted our golden cash cow, and put our dominance and entire company at risk, before someone else did, because it would happen eventually, you'll thank me later, pls promo kthx."

cobertos• 1 year ago

If that is the case, isn't the future more private code? The producers realize that public release is no longer in their favor and will keep it in their groups (until they leak or whatever)?

gruez• 1 year ago

You can speed up software development dramatically by simply copying the code that someone else was foolish/idealistic enough to store on GitHub, stripping out the copyright and licensing notices, and adapting it to your needs

That's not what people who are using AI for though? The typical use case is "write a program that does [insert business problem]". I highly doubt there's a codebase that solves that exact business problem.

dabockster8• 1 year ago

that someone else was foolish/idealistic enough to store on GitHub

That hits hard from my youth (and I'm only in my early 30s). Can we all please actively communicate to people that open source is not a business model, and any large firms promoting it are likely just looking for free work? Like communicate that in an empowering way?

Yeah, it would probably decimate open source contributions in the short term. But honestly, our field would be so much healthier overall in the long term.

billy99k2• 1 year ago

I argued this point in the early 2000s when Richard Stallman was seen as a god in the software community and tried to convince everyone that this would be the new business model for software.

The same could be said about copyrights/patents (another point argued over the years). If you get rid of these protections for everyone, the only people that will end up on top are the large companies with lots of resources that can just take your ideas and not compensate you.

userbinator2• 1 year ago

Ironically, I've heard a few people saying they wish the big companies would just copy their publicly available code on GitHub even without attribution instead of coming up with their own overengineered and inefficient solutions which take a lot more time and effort, and which are then forced upon those former people as users.

dehrmann• 1 year ago

Three Amazon engineers said that managers had increasingly pushed them to use A.I. in their work over the past year.

Three?

_pdp_• 1 year ago

AI coding assistants are a force multiplier. It can multiply good things and bad things equally and without discrimination.

pessimizer2• 1 year ago

The Amazon site search and at least some product pages seem to be down right now, although I'm hesitant to say that because I don't remember Amazon ever breaking.

I'm getting 503s ("Sorry! Something went wrong on our end") on searches and "We're sorry, an error has occurred. Please reload this page and try again" on product pages.

The AIs should probably fix that.

0x445442• 1 year ago

Fast, good, cheap; pick two. It's the single immutable law in software development and the one thing management has been trying to circumvent for decades. This is just their latest incantation.

canxerian• 1 year ago

With or without LLMs, the culture of Amazon would eventually lead to all of their positions feeling like warehouse work.

Source: I worked there.

shadowdev1• 1 year ago

Bugs. Bugs everywhere.

tengbretson• 1 year ago

I think the author simply wanted to write this story now. Like when in the last 10 years could you not have found a handful of Amazon devs to corroborate this claim?

agarren• 1 year ago

[Harper Reed] cautioned against being overly precious about the value of deeply understanding one’s code, which is no longer necessary to ensure that it works.

That just strikes me as an odd thing to say. I’m convinced that this is the dividing line between today’s software engineers and tomorrow’s AI engineers (in whatever form that takes - prompt, vibe, etc.) Reed’s statement feels very much like a justification of “if it compiles, ship it!”

“It would be crazy if in an auto factory people were measuring to make sure every angle is correct,” he said, since machines now do the work. “It’s not as important as when it was group of ten people pounding out the metal.”

Except that the machines doing that work aren’t regularly hallucinating angles, spurious welding joints, etc.

lolinder2• 1 year ago

Also, you know who did measure every angle to make sure it was correct? The engineers who put together the initial design. They sure as hell took their time getting every detail of the design right before it ever made it to the assembly line.

Who's filling that role in this brave new world?

alpaca1282• 1 year ago

They took their time precisely specifying the robot movements so they're always repeatable. Almost like...some sort of carefully crafted code.

mrheosuper4• 1 year ago

it would be the SA guy, but look like he will also be replaced by a bunch of "vibe coder"

throwup238• 1 year ago

Who's filling that role in this brave new world?

Us?

(Yeah, we’re fucked)

b112• 1 year ago

This also seems, to me, like composer/npm issues.

An entire generation of devs, who grew up using unaudited, unverified, unknown license code. And which at a moments notice, can be sold to a threat actor.

And I've seen devs try to add packages to the project without even considering the source. Using forks of forks of forks, without considering the root project. Or examing if it's just a private fork, or what is most active and updated.

If you don't care about that code, why care about AI code? Or even your own?

d0gsg0w00f2• 1 year ago

This is an underrated comment. Who's job is it to do the thinking? I suppose it's still the software engineer, which means the job comes down to "code prompt engineer" and "test prompt engineer".

xeonmc• 1 year ago

Cory Doctorow described this cart-before-the-horse arrangement as Reverse Centaurs[0][1]

[0] https://pluralistic.net/2022/04/17/revenge-of-the-chickenize...

[1] https://pluralistic.net/2024/08/02/despotism-on-demand/

nothercastle• 1 year ago

Not true anymore, lots of places want to reduce engineering hours and let production eat the cost of defects.

wat100002• 1 year ago

More to the point, there are people who carefully ensure that those angles are correct, and that all of the angles result in the car’s occupants arriving at their destination instead of turning into a ball of fire. It’s just that this process takes place at design time, not assembly time.

Software the same way. It’s even more automated than auto factories. Assembly is 100% automated. Design is what we get paid to do, and that requires understanding, just like the engineers at Ford need to understand how their cars work.

tdeck• 1 year ago

You're just not visionary enough to see the potential in vibe-CADing an automobile!

majormajor• 1 year ago

One of the reasons outsourcing for software fizzled out some compared to manufacturing is because in a factory you don't have "measuring to make sure every angle is correct" because (a) the manufacturing tools are tested and repeatable already and (b) the heavy lifting of figuring out all the angles was done ahead of time. So it was easy to mechanize and send to wherever the labor was cheapest since the value was in the tools and in the plans.

The vast majority of software, especially since waterfall methods were largely abandoned, has the planning being done at the same time as the "execution". Many edge cases aren't discovered until the programmer says "oh, huh, what about this other case that the specs didn't consider?" And outsourcing then became costly because that feedback loop for the spec-refinement ran really slowly, or not at all. Spend lots of money, find out you got the wrong thing later. So good luck with complex, long-running projects without deeply understanding the system.

Alternately, compare to something more bespoke and manual like building a house, where the tools are less precise and more of the work is done in the field. If you don't make sure all those angles are correct, you're gonna get crappy results.

(The most common answer here seems to be "just tell the agent what was wrong and let it iterate until it fixes it." I think it remains to be seen how well "find out everything that is wrong after all the code is written, and then tell the coding agent(s) to fix all of those" will work in practice. If nothing else, it will require a HUGE shift in manual testing appetite. Maybe all the software engineers turn into QA engineers + deployment engineers.)

anonzzzies4• 1 year ago

outsourcing for software fizzled out

Any data on that? I see everyone trying to outsource as much as they can. Sure, now it is moving toward AI, but every company I walk into have 10-1000s of FTEs in outsource countries.

I see most fortune 1000 companies here doing some type of agile/planningexecution which is in fact more waterfall. The people here in the west are more management and client facing, the rest is 'thrown over the fence'.

bgwalter10• 1 year ago

Amazing. What has Harper Reed written? Does he have any significant output?

This is frankly the most idiotic statement I have heard about programming yet.

pchristensen2• 1 year ago

This made the rounds recently: https://harper.blog/2025/02/16/my-llm-codegen-workflow-atm/

That quote really misrepresents his writing.

agentultra2• 1 year ago

It’s almost like he doesn’t know what a programming language is or what a compiler is for.

Agree, this comment makes no sense.

intended2• 1 year ago

The fact that they used the auto industry as an example is funny, because the Toyota way and six sigma came out of that industry.

So you are telling me, your AI code passes a six sigma grade of quality control?

I have a bridge to sell you. No, Bridges!

lightedman2• 1 year ago

"The fact that they used the auto industry as an example is funny, because the Toyota way and six sigma came out of that industry."

It's even funnier when you consider that Toyota has learned how bad of an idea lean manufacturing/6-Sig/5S can be thanks to the pandemic - they're moving away from it in some degrees, now.

bigfatkitten3• 1 year ago

six sigma came out of that industry

Six Sigma came out of Motorola, who still practice it today.

It was then adopted by the likes of GE, before finding its way into the automotive and many other manufacturing industries.

bartread3• 1 year ago

And except that the factory analogy of software delivery has always been utterly terrible.

If you want to draw parallels between software delivery and automotive delivery then most of what software engineers do would fall into the design and development phases. The bit that doesn’t: the manufacturing phase - I.e., creating lots of copies of the car - is most closely modelled by deployment, or distribution of deliverables (e.g., downloading a piece of software - like an app on your phone - creates a copy of it).

The “manufacturing phase” of software is super thin, even for most basic crud apps, because every application is different, and creating copies is practically free.

The idea that because software goes through a standardised workflow and pipeline over and over and over again as it’s built it’s somehow like a factory is also bullshit. You don’t think engineers and designers follow a standardised process when they develop a new car?

It would be crazy for auto factory workers to check every angle. It is absolutely not crazy for designers and engineers to have a deep understanding of the new car they’re developing.

The difference between auto engineering and software engineering is that in one your final prototype forms the basis for building out manufacturing to create copies of it, whereas in the other your final prototype is the only copy you need and becomes the thing you ship.

(Shipping cadence is irrelevant: it still doesn’t make software delivery a factory.)

This entire line of reasoning is… not really reasoning. It’s utterly vacuous.

fsloth• 1 year ago

It’s not only vacuous. It’s incompentent by failing to recognize the main value add mechanisms. In software delivery, manufacturing, or both. I’m not sure if Dunning-Krueger model is scientifically valid or not but this would be a very Dunning-Krueger thing to say (high confidence, total incompentence).

shakna2• 1 year ago

You don’t think engineers and designers follow a standardised process when they develop a new car?

Apparently the Cybertruck did not. And that sort of speaks for itself.

kloop• 1 year ago

The idea that because software goes through a standardised workflow and pipeline over and over and over again as it’s built it’s somehow like a factory is also bullshit.

I don't think it's bs. The pipeline system is almost exactly like a factory. In fact, the entire system we've created is probably what you get when cost of creating a factory approaches instantaneous and free.

The compilation step really does correspond to the "build" phase in the project lifecycle. We've just completely automated by this point.

What's hard for people to understand is that the bit right before the build phase that takes all the man-hours isn't part of the build phase. This is an understandable mistake, as the build phase in physical projects takes most of the man-hours, but it doesn't make it any more correct.

fisf• 1 year ago

The “manufacturing phase” of software is super thin, even for most basic crud apps, because every application is different, and creating copies is practically free.

This is not true from a manager's perspective (indoctrinated by Taylorism). From a manager's perspective, development is manufacturing, and underlying business process is the blueprint.

baxtr2• 1 year ago

His comment is a reflection of the future Software operating model they’ll have.

pchristensen2• 1 year ago

Harper talks a lot about using defensive coding (tests, linters, formal verification, etc) so that its not strictly required to craft and understand everything.

This article (and several that follow) explain his ideas better than this out of context quote.

https://harper.blog/2025/02/16/my-llm-codegen-workflow-atm/

HenryBemis2• 1 year ago

The issue is that, (the way I see it happening more and more in the real world): - tests are ran by machines - linters are (being) replaced by machines/SaaS services (so.. machines) - formal verification: yes - SHure, 5 people will review the thousands lines of code written every day, in a variety of languages/systems/stacks/scripts/RPAs/etc. or they will simply check that the machines return "green-a-OK" and ask the release team to push it to production.

The other thing that I have noticed, is that '(misplaced) trust erodes controls'. "hey the code hasn't broke for 6 months, so let's remove ABC and DEF controls", and then boom goes the app (because we used to test integration but 'come on - no need for that).

Now.. this is probably the paranoid (audit/sec) in me, but stuff happens, and history repeats itself.

Also.. Devs are cost center, not a profit center. They are "value enablers" not "value adders". Like everything and everyone else, if something can be replaced with something 'equally effective' and cheaper, it is simply a matter of time.

I feel that companies want to both run for this new gold-rush, while at the same time do it slowly and see if this monster bites (someone else first).

andyferris• 1 year ago

The correct analogy is that software engineers design and build the _factory_. The software performs the repeatable process as defined by code, and no person sits and watches if each machine instruction is executed correctly.

Do you really want your auto tool makers to not ensure the angle of the tools are correct _before_ you go and build 10,000 (misshaped) cars?

I’m not saying we don’t embrace tooling and automation as appropriate at the next level up, but sheesh that is a misguided analogy.

brundolf2• 1 year ago

Yes and this is just garden-variety abstraction and toolmaking, which is what programmers have done since the very beginning

lolbert6• 1 year ago

works for tesla

p_v_doom• 1 year ago

The correct analogy is that software engineers design and build the _factory_.

This is, I think very important especially for non-technical managers to grasp (lol, good luck with that).

brundolf2• 1 year ago

Something I've been thinking about is that most claims of AI productivity apply just as well (and more concretely and reliably) to just... better tooling and abstractions

Code already lets us automate work away! I can stamp out ten instances of a component or call a function ten times and cut my manual labor by 90%

I'm not saying AI has nothing to add, but the "assembly line" analogies - where we precisely factor out mundane parts of the process to be automated - is what we've been doing this whole time

AI demands a whole other analogy. The intuitions from automating factories really don't apply, imo.

Here's one candidate: AI is like gaining access to a huge pool of cheap labor, doing tasks that don't lend themselves to normal automation. Something like when manufacturing got offshored to China in the late 20th century

If you're chronically doing something mundane in software development, you're doing something wrong. That was true even before AI.

fhd22• 1 year ago

100%. I keep thinking this, and sometimes saying it.

Sure, if you're stuck in a horrible legacy code base, it's harder. But you can _still_ automate tedious work, given you can manage to put in the proverbial stop for gas. I've seen loads of developers just happily copy paste things together, not stopping to wonder if it was perhaps time to refactor.

Cthulhu_2• 1 year ago

Exactly that. Software development isn't about writing code, never was, it's about what code to write. Doesn't matter if I type in the code or tell an AI what code it should type.

I'll admit that assuming it's correct, an AI can type faster than me. But time spent typing represents only a fraction of the software development cycle.

But, it'll take another year or two on the hype cycle for the gullible managers being sold AI to realise this fully.

kamaal• 1 year ago

>That just strikes me as an odd thing to say.

This is along the same lines as why I don't doubt the syntactical features to break. I assume they work. You have to accept some abstraction work, and build on top of it.

We will reach some point where we will have to assume AI is generating the correct code.

PicassoCTs• 1 year ago

The lead poisoning of our time, companies getting high on hype-tech, killed off because the "freedom from programmers- no code tools" create gordion project nods.

And all because the MBAs yearn for freedom from dependencies and thus reality.

GrumpyNl• 1 year ago

In the beginning there was a saying, "nobody cares what your code looks like, as long as it works". We went full circle.

soraminazuki5• 1 year ago

It's a reckless stance that should never ever come from a software professional. "Let's develop modern spy^H^H^Hsoftware in the same way as the 737 Max, what could possibly go wrong?"

elmt35• 1 year ago

To me the biggest difference between a software engineer and an assembly worker is the worker makes cars, the software engineer makes the assembly line.

fock• 1 year ago

also software is a factory and the LLM is a workshop building factories. Also I strongly believe that people building factories still do a lot of a) ten people pounding out the metal AND b) measuring to check

Cthulhu_2• 1 year ago

That's a weird / wrong analogy. Software engineers design and program the robots, they don't build the cars.

discreteevent• 1 year ago

That analogy wouldn't give a person much confidence in the rest of his ability/judgement.

captainbland• 1 year ago

That's wild. You can't say this statistical, hyper-generalised system of "AI" is in any way comparable to the outputs of highly specific, highly deterministic machinery. It's like comparing a dice roll to a clock. If anything reviewing engineers now need to "measure the angles" more closely than ever.

namaria• 1 year ago

What strikes me the most is not even that people are willing to do that, to just fudge their work until everything is green and call it a day.

The thing that gets me is how everyone is attaching subsidized GPU farms to their workflows, organizations and code bases like this is just some regulated utility.

Sooner of later this whole LLM thing will get monetized or die. I know that people are willing to push bellow par work. I didn't know people were ready to put on the leash of some untested new sort of vendor lock-in so willingly and even arguing this is the way. Some may even have the worst of the two worlds and end up on the hook for a new class of sticker shock, pay down and later have these products fail from under them and left out to dry.

Someone will pay for these models, the investors or the users so dependent they'll pay whatever price is charged.

zelphirkalt2• 1 year ago

Seems like the more distanced we get from actual engineering methods, the more fucked up our software and systems become. Not really surprising to be honest. Just look at the web as an example. Almost everyone is just throwing massive frameworks and a hundred libraries as dependencies together and calls it a day. No wait, must apply the uglyfier! OK now call it a day.

fireflash38• 1 year ago

I'd love to blame shitty web apps on outsourcing of computing power to your users... But you'll hear countless stories of devs blowing up their AWS or GCP accounts with bad code or architecture decisions (who cares if this takes a lot of CPU to run, throw another instance at it!) , so maybe it's just a lazy/bad dev thing.

ThrowawayR23• 1 year ago

There's no incentive for engineering methods because there's no liability for software defects. "The software is provided “as is”, without warranty of any kind, express or implied, including but not limited to the warranties of merchantability." It's long overdue but most people in the software industry don't want liability because it would derail their gravy train.

never_inline• 1 year ago

And yet people here keep arguing proper cs education is not important.

Without proper understanding of CS, this is what we get. Lack of rigour.

ashoeafoot• 1 year ago

Companies with lack of engineering rigor are basically pre filtered customers packaged for a buyout by a company/vc able to afford more rigorous companies structure.

otikik• 1 year ago

Let them do it. It will naturally explode on their faces. Let Amazon be the early adopters of this trend.

RossBencina4• 1 year ago

The whole auto factory thing sounds completely misinformed to me. Just because a machine made it does not mean the output isn't checked in a multitude of ways.

Any manufacturing process is subject to quality controls. Machines are maintained. Machine parts are swapped out long before they lead to out-of-tolerance work. Process outputs are statistically characterised, measured and monitored. Measurement equipment is recalibrated on a schedule. 3d printed parts are routinely X-rayed to check for internal residue. If something can go wrong, it sure as hell is checked.

Maybe things that can't possibly fail are not checked, but the class of software that can't possibly fail is currently very small, no matter who or what generates it.

red_admiral2• 1 year ago

On the auto factory side, the Toyota stuck gas pedal comes to mind, even if it can happen only under worst-case circumstances. But that's the (1 - 0.[lots of nines]) case.

On the software side, the THERAC story is absolutely terrifying - you replace a physical interlock with a software-based one that _can't possibly go wrong_ and you get a killing machine that would probably count as unethical for executions of convicted terrorists.

jasonwatkinspdx• 1 year ago

Additionally production lines are all about doing the same thing over and over again, with fairly minimal variations.

Software isn't like that. Because code is relatively easy to reuse, novelty tends to dominate new code written. Software developers are acting like integrators in at least partly novel contexts, not stamping out part number 100k of 200k that are identical.

I do think modern ML has a place as a coding tool, but these factory like conceptions are very off the mark imo.

lambdasquirrel• 1 year ago

It’s been said that every jetliner that is in the air right now has a 1in hairline fracture in it somewhere? But that the plane is designed for the failure of any one or two parts?

Software doesn’t exactly work the same way. You can make “AI” that operates more like [0,1] but at the end of the day the computer is still going to {0,1}.

jcims• 1 year ago

A buddy of mine was a director in a metrology integration firm that did nothing but install lidar, structured light and other optical measurement solutions for auto assembly plants. He had a couple dozen people working full time on new model line build outs (every new model requires substantial refurb and redesign to the assembly line) and ongoing QA of vehicles as they were being manufactured at two local Honda plants. The precision they were looking for is pretty remarkable.

0xEF4• 1 year ago

Any manufacturing process is subject to quality controls.

A few things on this illusion:

* Any manufacturer will do everything in their power to avoid meeting anything but the barest minimums of standards due to budget concerns

* QA workers are often pressured to let small things fly and cave easily because they simply do not get paid enough to care and know they won't win that fight unless their employer's product causes some major catastrophe that costs lives

* Most common goods and infrastructure are built by the lowest bidder with the cheapest materials using underpaid labor, so as for "quality" we're already starting at the bottom.

There is this notion that because things like ISO and QC standards exist, people follow them. The enforcement of quality is weak and the reach of any enforcing bodies is extremely short when pushed up against the wall by the teams of lawyers afforded to companies like Boeing or Stellantis.

I see it too regularly at my job to not call out this idea that quality control is anything but smoke and mirrors, deployed with minimal effort and maximum reluctance. Hell, it's arguably the reason why I have a job since about 75% of the machines I walk in their doors to fix broke because they were improperly maintained, poorly implemented or sabotaged by an inept operator. It leaves me embittered, to be honest, because it doesn't have to be this way and the only reason why it is boils down to greed and mismanagement.

Neil44• 1 year ago

The ultimate irony being that they actually do measure all kinds of things against datum points as the cars move down the line as the earlier you junk a faulty build the less it's cost you.

red_admiral2• 1 year ago

It would be crazy if in an auto factory people were measuring to make sure every angle is correct

They are.

Mechanical engineers measure more angles and measurements than a consultant might guess - its a standard part of quality control, although machines often do the measuring with the occasional human sampling as a back-up. You'd be suprised just how much effort goes into getting things correct such as _packs of kitkats_ or _cans of coke_.

If getting your angles wrong risks human lives, the threat of prosecution usually makes the angles turn out right, but if all else fails, recalls can happen because the gas pedal can get stuck in the driver-side floor carpet.

Assembly-line engineering has your favour that (A) CNC machines don't randomly hallucinate; they can fail or go out of tolerance, but usually in predictable ways and (B) you can measure a lot of things on an assembly line with lasers as the parts roll through.

It was thankfully a crazy one-off that someone didn't check that _the plugs were put back into the door_, but that could be a sign of bad engineering culture.

ignoramous• 1 year ago

Reed’s statement feels very much like a justification of "if it compiles, ship it!"

Not really. More like, if the fopen works fine, don't bother looking how it does so.

SWE is going to look more like QA. I mean, as a SWE if I use the webrtc library to implement chat and it works almost always but just this once it didn't, it is likely my manager is going to ask me to file a bug and move on.

bumby4• 1 year ago

It would be crazy if in an auto factory people were measuring to make sure every angle is correct

To someone who used to automate assembly plants, sounds to me as a rationalization of someone who has never worked in manufacturing. Quality people rightly obsess over whether or not the machine is making “every angle” correct. Imagine trying to make a car where parts don’t fit together well. Software tends to have even more interfaces, and more failure modes.

I’ve also worked in software quality and people are great at rationalizing reasons for not doing the hard stuff, especially if that means confronting an undesired aspect of their identity (like maybe they aren’t as great of a programmer as they envision). We should strive to build processes that protect us from our own shortcomings.

AlexandrB• 1 year ago

Imagine trying to make a car where parts don’t fit together well.

Don't have to imagine, just walk over to your local Tesla dealership.

lifeisstillgood• 1 year ago

It would be crazy if in an auto factory people were measuring to make sure every angle is correct

They need to read “the code is the design”. When you are cutting the door for the thousandth car, you are past design and into building.

For us building is automatic - take code turn into binary.

The reason we measure the door is we are at the design stage abs you need to make sure everything fits.

kriro• 1 year ago

Car factory is a funny example because that's also one of the least likely places you will see AI assisted coding. Safety critical domains are a whole different animal with SysML everywhere and thousand page requirements docs.

notyourav• 1 year ago

Agreed. Real world is governed by physics, which tend to work well very well with interpolations, predictability and statistics

Hilift• 1 year ago

That just strikes me as an odd thing to say.

Is it though? It could be interpreted as an acknowledgement. Five years from now, testing will be further improved, yet the same people will be able to take over your iPhone by sending you a text message that you don't have to read. It's like expecting AI to solve the spam email problem, only to learn that it does not.

It's possible to say "we take the security and privacy of our customers seriously" without knowing how the code works. That's the beauty of AI. It legitimizes and normalizes stupid humans without measurably changing the level of human stupidity or the quality or efficiency of the product.

Sold! Sold! Sold!

InsideOutSanta• 1 year ago

It would be crazy if in an auto factory people were measuring to make sure every angle is correct

Yeah, but there's still something checking the angles. When an LLM writes code, if it's not the human checking the angles, then nothing is, and you just hope that the angles are correct, and you'll get your answer when you're driving 200 km/h on the Autobahn.

javcasas• 1 year ago

It would be crazy if in an auto factory people were measuring to make sure every angle is correct

That cannot be any furthest from the truth.

Take a decent enterprise CNC machine (look in youtube, lots of videos) that is based on servos, not the stepper motor amateur machines. That servo-based machine is measuring distances and angles hundreds of times per second, because that is how it works. Your average factory has a bunch of those.

Whoever said that should try getting their head out of their ass at least every other year.

skeeter20206• 1 year ago

It's wild to think that someone who purports to be an expert would compare an assembly line to AI, where the first is the ultra-optimization of systems management with human-centric processes thoughtfully layered on top and the latter a non-deterministic black box to everybody. It's almost like they are willfully lying...

Timpy• 1 year ago

I keep feeling like there's a huge disconnect between the owners/CTOs/managers with how useful they think LLMs _are supposed to be_, vs the people working on the product and how useful they think LLMs _actually are_. The article describes Harper Reed as a "longtime programmer", so maybe he falsifies my theory? From Wikipedia:

Harper Reed is an American entrepreneur

Ah, that's a more realistic indicator of his biases. Either there's some misunderstanding, or he's incorrect, or he's being dishonest; it's my job to make sure the code that I ship is correct.

harper• 1 year ago

It was not a strong metaphor! Can't win them all.

You are correct tho. I do think that we are approaching the point of "If it compiles, ship it"

nicetryguy3• 1 year ago

Andy Jassy ... said working faster was essential because competitors would gain ground if Amazon doesn’t give customers what they want “as quickly as possible” and cited coding as an activity where A.I. would “change the norms.”

Bruh, no competitor is going to set up an army of datacenters and warehouses on the gargantuan scale of amazon anytime soon... What do customers want? How about fixing your search function accuracy and policing the disgusting influx of scam products with fake reviews!!! Ugh. How about you get AI working on THAT jackknife...

echelon• 1 year ago

Bruh, firstly, no competitor is going to set up an army of datacenters and warehouses on the gargantuan scale of amazon anytime soon.

Google, Microsoft, OpenAI, Oracle, and xAI covet what AWS has. They are absolutely coming for AWS.

nicetryguy3• 1 year ago

AWS? Possibly. They have a huge head start and haven't lost much ground yet. The warehouses and physical delivery? No way in hell. Any competing physical infrastructure would take a decade from the first shovel hitting the ground today.

Izikiel43• 1 year ago

AWS has competition in Azure and Google Cloud, and AWS is the money maker for Amazon, everything else is incidental.

nicetryguy3• 1 year ago

AWS is the money maker ... everything else is incidental.

Considering Amazon kills and steals all kinds of small businesses and ideas that would be a godsend for Americans.

tho234j2342343• 1 year ago

What has been the experience with using these tools for folks on HN.

I'm generally amazed at the JS stuff it can come up with from scratch, but asking it to implement features is often not quite great.

Far more often it can't even use the damn edit tool given to the LLM! I have to take some of the "excrement" left behind; may be do a git reset; and then carefully fix the code.

I'm not surprised the managerial class has gotten high on the kool-aid of replacing workers...

_ink_2• 1 year ago

My experience has been poor. My last attempt was with copilot. I gave it the structure of a psql table and a DTO and asked it to create java code to write to the dB using a prepared statement. It failed and the errors were hard to spot. Not sure if it saved time. That was 4o.

ozim• 1 year ago

As a software dev I always felt like working warehouse - sprints, strony points being held on productivity metrics.

people think all devs work FAANG like companies when there is loads of companies where devs are treated like dirt only now one FAANG company catches up with reality.

Where business people seem to never be measured in any way.

If requirements were shitty - well dev team did a bad job - not that business made up stupid decisions on stupider time lines.

tho234j2342343• 1 year ago

I thought devs were always treated like dirt at Amazon ?

I've tried interviewing at that place, and have't felt that kind of hostility ever - one guy even wanted me to prove that K-means in 2d is in P. This I later found was a was the topic of a key paper in this niche of ML theory.

bubblyworld• 1 year ago

That can't be right - planar k-means is well known to be NP-complete.

rs1863• 1 year ago

Amazon is a well known sweatshop -- instead of producing clothes they produce code.

rand17• 1 year ago

I'm afraid software development will soon be about fixing code hallucinated up by AIs - and standups, retrospectives, reviews, demos, plannings etc.

microtherion• 1 year ago

Yes, it's all the wrong way around. I'd rather have an AI attending standups and planning meetings for me. It's supposed to understand my code perfectly, so why can't it report on it for me?

baalimago• 1 year ago

Seems to me like FANG is primarily sacrificing QA to increase productivity, with AI as the reason. Will be interesting to see if they too are 'peeing their pants' (warm and nice -> cold and not so nice), or if it actually will be sustainable.

LLM's were trailed as customer service bots, which backfired (for the most part). Now they're exploring different areas.

tho234j2342343• 1 year ago

May be it's easier in FAANG because a lot of the work comes from the sheer (and unnecessary IMO) complexity of the stack/tools.

In fact, a common meme within Google is that Engineers are mere plumbers routing protos from one place to another.

(Given how brittle these systems are though, I'm not sure if I trust LLMs to be better; but they may be more suited for dealing with un-interactive delayed-feedback baroque systems that are the norm in such places.)

surfingdino• 1 year ago

I'm so glad I wound down all of my AWS accounts and pay ~$1.5 for them per month. I'll close the rest this month.

misja111• 1 year ago

My experience with AI so far is exactly the opposite. It can help me with finding configurations to set up software packages, with syntax of programming languages that I use only occasionally so am not too familiar with, or it can give suggestions for improvement for code that I have written. It's like having a personal assistant.

But the interesting work: thinking about software architecture, about implementation strategies for large projects, or about finding a good debugging strategy for bugs in large code bases, that work is still for me.

dragandj4• 1 year ago

In other words, AI is just a replacement for google search and Stack Overflow, automated and served as an IDE plugin...

looofooo0• 1 year ago

Also serves as a manual on steroids, where you can find example of library usage for your problem.

Simon_O_Rourke• 1 year ago

While the writing seems to be on the wall for entry level developer positions, there's going to have to be a pathway for Amazon and others to have junior devs become senior devs, and be able up fill in whenever the machines fail.

cess11• 1 year ago

'“It’s more fun to write code than to read code,” said Simon Willison, an A.I. fan who is a longtime programmer and blogger, channeling the objections of other programmers. “If you’re told you have to do a code review, it’s never a fun part of the job. When you’re working with these tools, it’s most of the job.”'

I don't recognise this at all, not in myself, neither in the well functioning teams I've been a part of.

Executing and reviewing code is the job. That's the core thing we do. Writing code is just a fluent activity, like walking or absentmindedly spinning a pen between the fingers.

simonw• 1 year ago

Yeah, I agree. I said that because I've heard other programmers say it - which I guess is hinted at by the "channeling the objections of other programmers" part of that quote.

whatever14• 1 year ago

An example from today: I have some scientific calculations that run in a Django app.

One specific endpoint has a function that takes 2-3 minutes to execute. When I was hitting the endpoint I was getting a timeout error. I asked Claude to fix it, and it came up with a huge change in the front end (polling) and backend (celery) to support asynchronous operations.

I instead increased the nginx timeout from 1 minute to 5 minutes. Problem solved without introducing 100x the complexity.

rafaelmn• 1 year ago

You didn't specify how you asked Claude to fix it here so hard to evaluate if it's giving you what you asked for or just choosing the wrong approach. It's one of the things these models will do - they will narrow down on what you asked them to do - so for example if you just gave it your code without the deployment and said fix it on code level that would bias it to go in that direction, etc.

Disclaimer: not one of those "you didn't prompt right whenever LLMs fail" people, but just something I've noticed in my use of LLMs - they usually don't have the problem context so can't really tell you "should we even be going down this path"

whatever14• 1 year ago

I am pretty sure that if I hinted the direction of the solution the model would come up with the simple proxy config fix. I will try it for fun.

But I was trying to simulate the vibe coding experience where you are completely clueless and rely exclusively on the LLM. It works, but it creates even more work (we call it debt) for the future.

PS I was using Claude within Copilot, so it could definitely see the docker & nginx configuration.

wordofx2• 1 year ago

“I instead put in a hacky solution because I’m not a programmer and didn’t understand the solution or why it might be better”

whatever14• 1 year ago

Explain to us non-programmers why a 60" request (the default nginx timeout) is ok, but a 61" request is not.

grewsome• 1 year ago

I've read previously that certain parts of the software industry at one point was all in on move fast and break things, but then after a while the quality of their code was so bad that they took a different course. Maybe some people didn't get their fill the first time round. I can see that LLMs are going to be quick, I can't see how their code bases are going to do anything but turn into a very bad state, probably much more so this time round.

rs1863• 1 year ago

Sounds like Claude is doing the right thing, at least for the general problem.

You are just waiting for this to deteriorate, and at some point that 5min timeout is not long enough, and you'll have to come up with something else again.

And more generally, there are multiple issues with properly handling long running actions -- you want a good UX for these actions in your app, they may need to be queued or be allowed to cancel etc. I don't know the exact situation, but I assume this is a small app so you don't care much about it. But in any serious application where UX is important and you don't want someone to submit 10000 such requests (or 10000 users submitting at the same time) to blow up your backend, the sane design is to do this asynchronously with other mechanism to manage the actions.

geraneum• 1 year ago

Sounds like Claude is doing the right thing, at least for the general problem.

No not really. The engineer’s job is work within the constraints first and when a big change is really warranted, pursue it.

It’s having a structural issue that can be fixed with a reinforcement, and an automated system suggests to demolish and rebuild part of the most structure altogether. It’s not that clear cut. You maybe right but I wouldn’t jump into conclusions like this.

wolvesechoes• 1 year ago

Some software "engineers" are discovering they never actually did engineering and are surprised.

gsky• 1 year ago

Investing in AI companies is the best thing one could do now

nssnsjsjsjs4• 1 year ago

You just need to figure out which companies capture the value, if any. My bet is on Amazon, funny enough.

thih9• 1 year ago

The Amazon engineers said that managers have encouraged them to use A.I. to help write one-page memos proposing a solution to a software problem and that the artificial intelligence can now generate a rough draft from scattered thoughts.

Why generate that memo at all - maybe scattered thoughts would be enough and the AI could help managers understand them. Then again, if the goal is to replace the engineers and not the other way round, then this makes sense.

physicles3• 1 year ago

My hot take is if you’re writing a memo or a spec for internal consumption, using AI is an anti-pattern. The doc isn’t the important product, the act of writing the doc is, because that’s how you clarify your own thinking.

We know AIs are fine for bouncing ideas off of, but they’re not senior-level architects yet.

fedeb95• 1 year ago

we'll see if this increased return for companies will stay there in the long term, or create unforeseen consequences.

alkonaut• 1 year ago

I love that I never even jumped on the whole "internet" thing when it arrived to software development in the 90s. After that, dodging blockchains and later AI was even simpler.

n_ary• 1 year ago

Andy Jassy, the chief executive, wrote that generative A.I. was yielding big returns for companies that use it for “productivity and cost avoidance.” He said working faster was essential because competitors would gain ground if Amazon doesn’t give customers what they want “as quickly as possible” and cited coding as an activity where A.I. would “change the norms.”

Like what exactly is Amazon giving us here? I don't get it. Also, I want to see Andy Jassy start writing some codes or fixing issues/bugs for next 5-10 years and have those reviewed by anonymous engineers before I take any word from him. These marketers/sales sleezy dudes claim garbage about things they do not do or know how to do but media picks up everything they say. It is like my grandmother who never went to school starts telling me about how brain surgery is slow and needs productivity else more people will die and those doctors need to adapt. Shameless behavior of these marketing/sales idiots as well as the dark side of media has reached new extreme in this AI bubble.

Meanwhile, I can see from comments how a lot of HNers totally agree everything this salesy guy says as holy bible verse and my colleague is sending me freaked out texts about how he is planning to switch career as Amazon Super Boss is talking about vibe coding now but became calm after I told him, these dudes are mostly sales/MBA who never wrote code of fixed issues, same way our PO doesn't know the diff between var and const.

wavemode• 1 year ago

Jassy is one of the most anti-employee CEOs I've ever seen. Which is remarkable, considering who he is the successor of.

akudha• 1 year ago

Why is this surprising? Isn't being anti-employee and being a psycho the most basic requirement for being a CEO of huge companies these days?

LightBug1• 1 year ago

Do they have timed toilet breaks?

rybosworld• 1 year ago

Hard to feel bad for any dev that agrees to work at Amazon.

They’ve opted into shitty working conditions for years and played a pretty big part in spreading those conditions to other places.

I have found that former Amazon employees can get jobs at other FAANG companies. But smaller/medium companies that compete on culture rather than pay, don’t tend to want to hire those folks.

ken47• 1 year ago

As expected, shallow analysis from people who don’t understand our craft and/or are all too happy to feed a hype train.

jbs789• 1 year ago

“It would be crazy if in an auto factory people were measuring to make sure every angle is correct,” he said, since machines now do the work.

Am I accountable for the angles or not? Don’t think so! My job is just to bolt it on.

Is a developer not accountable for the code, regardless of how it was created?

chaosbolt• 1 year ago

This is good, it means they'll optimize for this kind of coding to have less coders to save pennies here and there and please the shareholders short term, and get left behind in the long term once an old school startup that tries to do everything better comes up.

Of course I'm fantasizing and we sadly don't live in this kind of world (Amazon or Google would stomp you if you come for their users) but hey a man can dream.

leibnitz27• 1 year ago

If it compiles, ship it. Even if it's subtly broken.....

I, for one, look forward to the contracting gigs fixing garbage code. :)

The pull back on this will be brutal.

jongjong• 1 year ago

The trend towards increasingly low-quality, mass-produced code started long before AI but AI has greatly accelerated the trend and taken it to levels not previously possible. Short-term, this is destroying software dev jobs, long-term, it's going to create a lot of software dev jobs when people have to maintain or rewrite the mess (unless AGI can fill the gap). It will create a lot of jobs in cyber-security also because complex code, rushed out hastily, with less context awareness, opens the door to more potential vulnerabilities.

It's interesting how the trend which took over manufacturing is now also taking over software development. This trends towards low-quality, mass-production is taking over everything.

snowstormsun• 1 year ago

So who will be held accountable if things break? "Sorry dear customer, it's not our fault, this code was AI generated and it worked in our tests."

zettapwn• 1 year ago

No one is free until everyone is free. Thinking you’re insulated because you do white collar work is idiotic.

major5052• 1 year ago

This seens like a recipe for disaster on the making. Well, thats why we need brave people. To show us the way, or serve as a cautionary tale.

yapyap• 1 year ago

Sounds like employees have just started turning into accountability sinks, you let the AI do all the work, give the employee not enough time to really read / understand it all and then blame the employee who “pushed” it for not catching a bug instead of being accountable yourself if you let an AI code it all.

rassalas• 1 year ago

This kind of thinking is exactly why a lot of Americans buy only foreign made vehicles.The American auto industry is a failure propped up by government subsidies. Please don't use them as an example of anything good.

vijucat2• 1 year ago

Not directly related, but an anecdote: well before AI, I was talking to a Portfolio Solutions Manager or something from JP Morgan. He was an MD at the firm and very full of himself. He told me, "You guys, your job is....you just Google search your problem and copy paste a solution, right?". What I found hilarious is that he also told me, "The quants, I hate that they keep their C++ code secret. I opened up the executable in Notepad to read it and it was just gibberish". Lesson: people with grave incompetence at programming feel completely competent to judge what programming is and should be.

game_the0ry• 1 year ago

Lesson: people with grave incompetence at programming feel completely competent to judge what programming is and should be.

As an employee at a company with a similar attitude, I cannot agree more with this.

JKCalhoun7• 1 year ago

My own tangential gripe (a bit related to yours though): the factory work began when Agile crept into the workplace. Additionally, lint, unit tests, code reviews... all this crap just piled on making programming worse still.

It stopped being fun to code around that point. Too many i's to dot to make management happy.

thinkingtoilet• 1 year ago

What's wrong with writing tests? I sleep well at night when we push to production because of our robust test suite.

rwmj• 1 year ago

Agile yes, it's micromanagement at scale. But writing tests and doing code reviews is good practice.

myvoiceismypass• 1 year ago

I never found linting or writing unit tests to be particularly un-fun, but I generally really really value correctness in my code, and both of those things tend to help on that front.

FooBarBizBazz• 1 year ago

If you give up on unit tests and code review then the code is "yours" instead of "ours" and your coworkers will not want to collaborate on it with you.

However, this has to be substantive code review by technical peers who actually care.

Unit tests also need the be valued as integral to the implementation task. The author writes the unit tests. It helps to guide the thought process. You should not offload unit tests to an intern as "scutwork".

If your code is sloppy, a stylistic mess, and unreviewed, then I am going to put it behind an interface as best I can, refer to it as "legacy", rely on you for bugfixes (I'm not touching that stinking pile), and will probably try to rally people behind a replacement.

HDThoreaun2• 1 year ago

It turns out that doing a good job at work is more important than having fun.

varjag• 1 year ago

It doesn't help that most "tech visionaries" or people considered tech bros these days more often come from accounting or legal background than anything technical. They are widely perceived as authorities but come without the expertise. This is why it's so perplexing for the techies when the industry gets caught up in some ridiculous hype cycle apparently neglecting the physical realities.

bumby4• 1 year ago

I used to work in aerospace R&D. The number of times I heard some variant of “it’s just software” to disregard a safety critical concern was mind boggling. My favorite is a high-level person equating it to writing directions on a napkin.

dwaltrip• 1 year ago

Hubris and fragile egos run amok

A burning need to dominate in a misguided attempt to fill the gaping void inside

Broken and hurting people spreading their pain as they flail blindly for relief

Our world creaks, cracks splintering out like spider thread

The foundations tremble

vijucat2• 1 year ago

Poetic

tlogan• 1 year ago

At the end of the day, coding with AI is still coding.

LLMs need to get better at it - more precision, less hallucination, larger context windows, and stricter reasoning. We also need a clearer and more domain-specific vocabulary when working with them.

In short LLM will end up as one more programming language - but one that’s easier for humans to understand, since it operates using the terminology of the problem domain.

But humans will still need to do the thinking: the what-ifs, the abstractions, the reasoning.

That’s where AI becomes unsettling - too many people have been trained not to think at all.

assimpleaspossi• 1 year ago

I'm glad I'm a programmer and not just a coder.

coliveira• 1 year ago

A recent paper by researchers at Microsoft and three universities found that programmers’ use of an A.I. coding assistant called Copilot, which proposes snippets of code that they can accept or reject, increased a key measure of output more than 25 percent.

Study paid by MS finds that MS tool is improving productivity...

awinter-py• 1 year ago

ah yes, how the kiva robot lifts a mobile shelf full of Codes and brings them to me for my inspection. I use my left hand to reach for the Code I need and place it in the tote, because my right shoulder is injured from repeating this motion.

The tote goes onto a conveyor where it is shot forwards, backwards and sideways, all around the warehouse until it reaches the compiler station. There, another worker will lint each Code with a lint roller and compile them by selecting the appropriately sized cardboard box.

Next, this executable bundle will be linked with its customer, possibly by means of a plane or 18-wheeler truck, with last-mile linking performed by a smaller vehicle suitable for urban traffic ...

icedogma• 1 year ago

I worked at Amazon Robotics low level software team, as a contractor, for a few months. This article sounds about right. Crazy pace of work with no room for reflection or — especially — groundbreaking invention, is given. And that was a year ago, when AI was totally optional.

The engineers are smart and fast folks, which helps a great deal. But don’t seek true creativity and deep expertise at Amazon, unless we’re talking about legacy L6+ (or likely even L7+) level engineers who get to have the leeway needed. The picker robots I had been working with were created, originally, by a Netherlands company, not Amazon. Amazon’s job was scaling it all up for their massive warehouse operation. It was a bunch of hacks on top of hacks. Heck, not even Harel state machines (which is kind of what one would expect to have on a properly designed event-driven robot that’s not overly tied to its PLC implementation, where things can be modularized, in and out, as needed). Just a bunch of switch statements and a menagerie of globals, with only two key people intimately familiar with most of its key code structure, one of whom decided to leave the group in frustration.

In fairness, the mess didn’t originate just at Amazon, but there never was a technical push at Amazon to have the initial work done in the right way. Not the software part anyway.

I can see that the craziness not only didn’t abate but has gotten worse. The money was good but it’s not worth the stress and other frustrations IMO, unless you’re a noncontractor and at least a technical program manager or above, where you won’t be coding as much as reviewing others’ code anyway.

Otherwise, I can totally understand why people just work there for a couple years, to get their resume looking good and then leave

storus• 1 year ago

The same complaints come from my doctor friends - they are now also being run as a factory where each procedure has a certain allotted time and they need to fit each patient into it. Like you have 10 minutes for a GP task, 15 minutes for psychotherapy, 3 minutes for X-Ray etc. They are complaining about it a lot, it makes them feel like robots.

iwontberude• 1 year ago

This is why I quit Google many years ago

cleandreams• 1 year ago

This type of thing really incentivizes founding a startup. If you are a very senior developer, who needs the corporate stupid factory? You can do a lot of work with half the people and work for yourself.

DrNosferatu• 1 year ago

But they mean general warehouse work, or “Amazon Warehouse” work?

Because it’s quite different…

Suppafly• 1 year ago

At Amazon, some coders say their jobs have begun to resemble warehouse work

Only because the people responding have never done warehouse work.

Shocka12• 1 year ago

As someone who packed pears for my first job in a warehouse at 15 1/2 years old for 10 hours a week, this comment is great. Working in a warehouse for pennies is about as worse as it gets. Now, as a well paid software engineer, a very bad day doesn't even come close to that experience.

bustling-noose• 1 year ago

I want to know if I can use an LLM during problem solving at an interview for SDE jobs at Amazon, Google and Microsoft who have boldly claimed lot of their work is done by LLMs. I will be able to point out the algorithm and explain how it works including the time complexity, but I will need the LLM first to solve it.

Will SDE interviews change ? Are these companies gearing up to let AI engineers in ? I highly doubt this is ever going to happen.

There is a quote by buffet that I think applies to a lot of scenarios not just investing : 'to be fearful when others are greedy and to be greedy only when others are fearful.'

ML IoT 5G blockchain etc. so many technologies are great that had their gaussian curve moment during greed. But these things take a back seat after that.

Anamon4• 1 year ago

"In a memo to employees in April, the chief executive of Shopify, a company that helps entrepreneurs build and manage e-commerce websites, announced that “A.I. usage is now a baseline expectation” and that the company would “add A.I. usage questions” to performance reviews."

Usage of a tool becoming a performance metric is a clear sign that the tool isn't doing what it's promised to do. This is managers desperately trying to prevent their investment in GenAI from blowing up in their faces.

This is as clear an admission as any that the supposed productivity or quality gains from LLM coding aren't showing up in the results. If a tool is effective, management generally doesn't have to micromanage and force employees into using it.

At Amazon, some coders say their jobs have begun to resemble warehouse work

Related Stories