AI is creating a generation of illiterate programmers

https://nmn.gl/blog/ai-illiterate-programmers

by namanyayg • 1 year ago

69 117 1 year ago

nmn.gl

dartos5• 1 year ago

Programming is a skill that atrophies surprisingly quickly.

I spent a year just coasting at a job and not really writing any code.

It was quite difficult to get to my same level after starting my new job.

Now that I’m back there I’m weary of losing my touch again.

dns_snek3• 1 year ago

Does it? I've had some months-long gaps where I hadn't written a single line of code and I always felt like I could get back up to speed within a few days. I suppose that depends on where you expect your level to be, are we talking about $300k+ positions at big tech or your average software job?

Re-learning a large codebase would take longer, of course, but I'm talking about just getting up to speed like if I was starting a new project.

dartos5• 1 year ago

Obviously, it's different for everyone :)

But surely you can tell it's not like riding a bike.

just getting up to speed like if I was starting a new project.

This is the best case scenario for getting back up to speed.

It may sound harsh, but there's not a lot of skill involved in starting new coding projects (it's one of the best things about the craft imo). Of course you'd get back up to speed quickly, especially if you have 10+ years of programming under your belt.

mclau156• 1 year ago

Its also creating a generation of non-programmers who can now program

worthless-trash2• 1 year ago

I belive you are incorrect. Its making a generation of people who can write prompts and sometimes get a useful program out of it.

lawlessone• 1 year ago

shhh, think of all the security jobs this will create

aredox3• 1 year ago

Don't think so, as you need to review the output to stave off hallucinations

dczomb• 1 year ago

I don't consider copy and pasting stuff I don't understand programming.

varispeed2• 1 year ago

Until an obstacle hits them and AI goes into a "reasoning" loop unable to find a solution and they don't know how to "nudge" it to the right path.

I've seen in some companies, relatively trivial tickets are bounced from developer to developer for months and management believing these are hard to do.

cricketsandmops2• 1 year ago

This happens to me and I just move to another AI program if i have 0 idea, but I have been getting better at figuring things out too.

slightwinder4• 1 year ago

How? What I see are non-programmers who can create programs, but still have hardly any ability to understand them, debug them or even create them on their own without AI-tools. This is similar, but worse, to what we used to call Stack Overflow-Programmer, or coding bootcamp-kids.

cricketsandmops2• 1 year ago

I would fall into this camp. I don't understand everything I am doing, but I have learned a lot through trial and error. I have an IT background in infrastructure and dabbled in automation, but never enough to be good. AI has allowed me to create things from my ideas that interest me. Is it good, no. Do I sell it though, no. I create things for myself. I wouldn't be able to do these things without it though because I didnt have the time and the teaching I had tried didnt interest me.

sksrbWgbfK• 1 year ago

"who can generate write-only text with a prompt" but won't understand why it is wrong.

tyho• 1 year ago

Is that actually happening? I use o1 a lot for brainstorming etc, but I don't think I could build a program of any significant size without understanding how to program myself so to guide o1 and stitch it's outputs together. It's also not at all uncommon for me to have to do stuff completely myself, it doesn't have great insights into deep problems for instance.

aunty_helen• 1 year ago

It’s also creating a generation on management, clients and non-technical people who send me outputs from Claude and suggest I implement what it said. Needless to say, this makes my life harder.

aithrowawaycomm2• 1 year ago

I would say "expands the class of simple applications one can build without expertise" in the same way spreadsheet software enabled simple financial programs, website builders meant you didn't need to do HTML/JS nitty-gritty, etc. OTOH these technologies also enable financial blunders (I am personally responsible for at least one really nasty Excel bug), horrible website design, etc etc.

Edit: Arvind Naranyan of AI Snake Oil uses LLMs to create "one-time-use" apps, mostly little games or educational things for his kids. As an AI skeptic I think this is cool and should be celebrated. But there is a serious downside when using it blindly for real work.

I have never liked the term "can program" anyway, programming is the easy part....

insane_dreamer4• 1 year ago

And that is a good thing because ....

shlomo_z• 1 year ago

Seems like everyone disagrees with you. Instead of reply to each of them individually I will reply here:

I used to be a non-programmer (although extremely technically advanced). I was very good with the terminal because I always played around with ADB and other things. I even wrote Windows batch scripts (but looking at them now is embarrassing. I used a lot of repetition instead of loops).

When ChatGPT became available, and my friend taught me how to use it to write Python scripts, I went all in. I didn't write a word of my own code. I just copy /pasted Chatgpt's output into notepad++. It was really annoying sometimes to get ChatGPT to change a simple thing. It hallucinated often and "fixed" stupid things I didn't ask for help with. I used all these scripts for various personal things. For example, it made me a GUI to edit the metadata of every .mp3 in a folder).

After several months of exposure to code, I knew the basic syntax and tried to code myself when I could, and Stack Overflow was a big help. I still had very basic skills and didnt even know what a `class` was.

Fast forward to now, I consider myself extremely good at Python, and decent in other languages. I now use classes and dataclasses all the time. I always add type hints to my code. I follow Python PEPs and try to design my code to be short, maintainable, and to the point. If a library lacks documentation, I just read the source code. I started using an IDE, and the intellisense is really a step up from using notepad++.

I still use copilot, but I refuse to paste large code blocks from any llm. It makes the code so much harder to maintain. If I really do like the code chatgpt gave me when I get stuck, I write it myself, instead of copy/paste.

I don't know if everyone has the same experience as me, but I would certainly say that ChatGPT helped me become a successful programmer.

PKop• 1 year ago

And that's a bad thing.

ladyprestor• 1 year ago

I would say it's non-programmers who THINK they can now program. As soon as the LLM stops being helpful, they're stuck.

vrighter2• 1 year ago

Programming is a mindset, more than just the ability to type (or copy/paste, nowadays) something that the compiler accepts. So no, they can't program. If something goes wrong they have no idea why it's wrong. They also have no idea why it's right, when it just happens to be.

KaiserPro• 1 year ago

I bet there are journals that have roughly the same sentiment for compiled languages vs assembler. and then again on assembler vs machine code.

However just like learning to drive, I think there is much scope for forcing the basics once in a while.

The old programming blue shell, as it were.

aiono• 1 year ago

There is a difference: when you code in higher level languages most of the time (except some domains where you still need to understand assembly) you don't need to understand how generated assembly works. This is not the case with AI generated code. You still need to be able understand, modify and validate the generated code.

dartos5• 1 year ago

I bet there are journals that have roughly the same sentiment for compiled languages vs assembler

Please find one. I highly doubt there are any at all.

AI isn’t translating idea to code. It’s guessing as to what the most likely next code is, given a context.

Compiling is exactly translating code to machine code, but hiding some (or a ton of) technical details.

Those are two wildly different mechanisms to interact with.

Assembly and compiled languages require the same precision of thought, just at different levels of technical detail.

danielbln4• 1 year ago

LLMs can absolutely translate ideas to code. How is that even a topic of debate. It might not always be good or maintainable code, but the whole notion of "it's just a next token predictor" is so trite at this point.

aredox3• 1 year ago

Using AI to become a better coder is like using Reader's Digest to become a better writer, or build one's taste in literature.

tmaly• 1 year ago

Reader's Digest is great. Do you have a subscription?

aredox3• 1 year ago

I loved reading Reader's Digest when I was young.

I shudder at the idea of what I would have become if life made it so that I would only read RD (bc of a lack of time, of support, of curiosity) for the rest of my life. A shallow half-knowledge I would have been contented with, oblivious to what literature is really about.

Another similar idea I am culpable of: watching videos of people playing games instead of playing for myself. Not the same experience at all.

sschueller• 1 year ago

I recently needed to create a Dockerfile which I have done many times before. Instead of opening the documentation I used ChatGPT.

It took 3 times longer than if I had done it by hand because I just kept copy and pasting each error back to ChatGPT ending up in a dead end having to start over from scratch. As soon as you have a requirement that is not the 90% of cases you can quickly end up in a dead end especialy if you have more than one such requirement.

Once the LLM starts hallucinating you can easily get into corner where you will need to reset and start from scratch.

rvnx• 1 year ago

I used to be like you. This is why you may want to try Cursor paired with Claude for this.

Set Cursor in "agent" mode (with Claude as a model).

Ask what you need.

It will automatically test your Docker deployment and configuration.

Then you become like me, a monkey clicking "Accept".

But a happy monkey.

chudGPT• 1 year ago

Eventually as your project grows in complexity, what you're asking of the model will exceed the intrinsic limits of what the model can do, and you won't be able to troubleshoot your problems yourself because you didn't learn anything along the way.

wseqyrku• 1 year ago

As soon as you have a requirement that is not the 90% of cases you can quickly end up in a dead end

You framed that as a criticism of ChatGPT, I think that is indeed the point the OP is making. IMO if this is something you used to do you could get a nice starting point with a single prompt and then go from there, then that would take 3 times less compared to reading the documentation from get-go.

medhir• 1 year ago

yeah, I’ve had many moments in my AI-driven development cycles looping with the LLM in a dead end, eventually giving up and falling back to regular old programming.

I think LLMs are great for reducing the friction with just starting a task, but they end up being more of a nuisance when you need to dig into more nuanced problems.

TomK32• 1 year ago

I haven't written any C++ in ages so I used a local Ollama3.2 to do a simple class and struct. The AI wanted me to use the `friend` keyword on class functions that weren't being accessed from anywhere else... I did correct itself after I asked but it took some intuition and googling to get there.

conor-• 1 year ago

This is basically the modern "google-driven" development experience as well. The copy-pasting from SO trope has existed for some time, but it's much worse in the last 5-7 years with the increase of blogspam/LLM generated content on Medium that's SEO'd to the top and effectively just rewriting the "getting started" page of some language/framework/tool/etc. for resume boosting points.

It seems like a lot of the models in ChatGPT and Copilot were trained on that content and in turn tends to produce a lot of dead end solutions for anything that isn't the 90% cases and often leads to more pain than reading documentation and building a solution through iteration/experimentation.

lgas• 1 year ago

I recently needed to create a Dockerfile which I have done many times before.
It took 3 times longer than if I had done it by hand because I just kept copy and pasting each error back to ChatGPT

You know that if you already know what you're doing you can tweak the output directly instead of mindlessly pasting it back into ChatGPT, right?

Jensson• 1 year ago

Its harder to tweak someone's else code than to tweak your own.

calmworm• 1 year ago

And high level programming languages created a generation of programmers who can’t code in assembly or machine language.

aithrowawaycomm2• 1 year ago

That wouldn't be the case if ordinary compilers outputted buggy code 5% of the time like a SOTA LLM does. (I am being quite generous to LLMs here)

ohplease• 1 year ago

5% of the time

that would be a generous invented statistic if it only was addressing the inherent stochastic nature of llm output, but you also have to factor in the training data being poisoned and out of date

in my experience the error rate on llm output is MUCH higher than 5%

atx2bos2• 1 year ago

exactly. Programming languages are all just levels of abstraction above the analog/digital interface.

While it is important to understand the fundamentals of coding, if we expected every software engineer to be well versed in assembly that wouldn't necessarily result in increased productivity.

LLMs are just the next rung up on the abstraction ladder.

There will always be people interested in the gritty details of low level languages like Assembly, C, that give you a lot more granular control over memory. While large enterprises and codebases, as well as niche use cases can absolutely benefit from these low level abstraction specialists, the avg. org doesn't need an engineer with these skills. Especially startups, where getting the 80% done ASAP is critical to growth.

dns_snek3• 1 year ago

What's the minimum set of skills that separates an idea person from a programmer?

If someone who can't read, write, or debug code can still be called a programmer, which unique skills do they still possess?

dns_snek3• 1 year ago

A programmer who doesn't know assembly is still a programmer.

Someone who delegates programming tasks to others through natural language instructions isn't a programmer, but a manager.

Tenoke• 1 year ago

They wouldn't needed to be 'literate' just how many programmers today don't need to know assembly or low level languages or compilers for their jobs.

parpfish4• 1 year ago

What does gpt dependence look like for people that have non-programming office jobs where they just email all day?

Do they just copy paste emails and office menus into an agent and send responses? Do they actually make any decisions themselves, or do they just go with whatever?

slightwinder4• 1 year ago

I saw someone recently explaining it this way: we now have tools to inflate keywords and prompts into fancy side-long text, which we then feed to tools which will shrink those text into short summaries.

So I guess office jobs are now grinding machines, where one side has keywords, the other side extracts keywords, and in the middle there is an enormous waste of resources to fit some manager's metrics.

parpfish4• 1 year ago

We spent years developing smarter compression tools, but AI has taught us that people actually want inflation tools

Havoc2• 1 year ago

What does gpt dependence look like for people that have non-programming office jobs where they just email all day?

There isn't any such dependency (yet).

Despite strong corporate push I'd say the average person I work with uses it maybe once a week - usually research or summarization.

The inroads LLMs are making on programming aren't translating to other office jobs cleanly. Think it's largely because other areas don't have an equivalent source of training data like github

It'll no doubt come, but for now all programmers seem to have mostly automated themselves out of a job not others from what I can tell.

Do they actually make any decisions themselves

There is little to no GPT decision making happening despite all the media chat about AI CEOs and similar bullshit. Inspiration for brainstorming is about as close as it gets

parpfish4• 1 year ago

would you know if they were dependent?

my assumption was that these office people were secretly using it to automate dull parts of their work without the approval or knowledge of their coworkers

lettergram• 1 year ago

We’re not becoming 10x developers with AI.
We’re becoming 10x dependent on AI. There’s a difference.

This is true, but I also don't need 10x the ability to write cursive any more. I used to have great hand writing, now it's a very poor. That said, my ability to communicate has only increased throughout my life. Similarly, my ability to spell has probably diminished with auto-correct.

Yes, you will become dependent on AI if you're a developer, because those who use AI will take your job (if you don't) and be significantly more productive than you. It sucks, but that's reality. My grandfather started programing on physical cards, he knew a lot of stuff I had no idea about. That said, I'd be able to run circles around him in the programming (and literal) sense today.

The question is really what skills do you need to know to still execute the job with AI effectively. I'd argue that doesn't change between AI and not having AI.

As a seasoned engineer, I spend probably 60% of my time designing, 30% managing/mentoring/reviewing and 10% coding. That probably wont change much with AI. I'll just be able to get a lot more done with the coding and design in the same timeframe. My skillsets likely will remain the same, though writing small functions may diminish slightly.

palmfacehn• 1 year ago

Agree here:

...and 10% coding.

Disagree here:

AI will take your job (if you don't) and be significantly more productive than you

Physically typing out the code is one of the last steps. Before that can happen the problem which is to be solved must be identified. An appropriate approach to the problem must be formulated.

Look around at some of the deranged hype in this AI cycle or look back on previous hype cycles. Many of the proposed "solutions" do not solve the problems of consumers. Solutions in search of a problem are as abundant as novel tech. These things dominate the zeitgeist. Perhaps they help inspire us, but they are not guaranteed to solve immediate problems.

There's an important distinction between problem solving skills, tools and the most probable token. To innovate is to do something new. AI, on the other hand, does the most probable thing.

grajaganDev2• 1 year ago

Leetcode has created a generation of illiterate programmers.

caseyy2• 1 year ago

Yes but we already have illiterate programmers and there are companies who hire them, and the majority that recognises what they are and don’t. I’m talking about the code copy-pasters who cannot really reason about the problems they’re given nor understand how the code they pasted works. My point is that this is not a new phenomenon and I don’t think it will fool those that weren’t fooled by it before.

__m• 1 year ago

The worst i heard this week was "we don't know what this code does" I'm pretty sure they were hired as developers.

caseyy2• 1 year ago

The worst I heard in my career so far is that the pasted code cannot be buggy because it was highly upvoted on stack overflow.

zimbatm• 1 year ago

This is true for any type of assistance. We have had the same arguments about calculators a long time ago.

sitzkrieg• 1 year ago

calculators dont spit out nonsense learners will not catch

AnimalMuppet3• 1 year ago

Oh yes they do! (Granted, it's either because learners input nonsense, or because they fat-fingered when inputting sane numbers... which I suppose are the same thing when you come down to it.)

pixxel• 1 year ago

Incorrect inputs and it will. Double-checking source numbers and correct key presses was commonplace.

tart-lemonade• 1 year ago

I've heard this argument about spell checkers in Word, browsers, etc as well.

Anecdotally, I wouldn't be surprised if there was some truth to it: in the early 2010s I saw tons of people saying "defiantly" that meant to type "definitely", and I couldn't figure out why. Then I misspelled "definitely" once and saw the spell checker suggested "defiantly", and it finally clicked.

Spell checkers have definitely improved since then (I can't remember the last time I saw the definitely/defiantly mix-up), but I can't help but wonder how bad suggestions have affected the way people understand languages.

A similar phenomenon exists on TikTok with words taking on contradictory or even opposite meanings, though this doesn't seem to be caused by spell check so much as constant incorrect usage rewriting the definition in people's minds (like "POV: you're getting mugged" when it should actually be "POV: you're mugging someone" or "POV: you're a bystander watching someone get mugged").

insane_dreamer4• 1 year ago

There is truth to it. My 12 year old is a horrible speller (like 3rd grade level) and doesn't care because he says he can "just use grammerly".

There is a dumbing down effect, for sure.

GeoAtreides• 1 year ago

Socrates and writing, calculators and LLMs, television and social media... thought terminating cliches that just crowd the discussion in the same way spam crowds an inbox

also we didn't have those discussion before, everyone forgets about abacuses and log rulers

ookblah• 1 year ago

imo we are seeing two types of programming viewports emerge.

i mean maybe it's just me but i quite like this new normal where i'm an "intermediary". i review every line that comes out and research things if they don't make sense or i'm confused. and a lot of times it helps me figure out how to approach a problem or refine an existing one. i'm still architecting everything and diving in.

i guess if you liken progamming to like i dunno, chopping your own tree down, turning it into 2x4s to build a house then yeah, it seems like your skills are atrophying.

i do not miss the "joy" of writing random boilerplate 100s of times or variations of the same function. i do however fear for the the new generation who maybe didn't cut their teeth on having minimal guidance. i can totally see people just coasting and blindly pasting things, but that's always sort of existed on stack exchange, etc.

TechDebtDevin• 1 year ago

My rule with LLMs is to use them only after reading documentation. It's easy to ask them for some boilerplate for something you haven't done in a long time, however, I think its better to just take five minutes to go read whatever spec/documentation is out there and then ask the LLM to write whatever code if you want.

remoroid2• 1 year ago

Coding will become AI dominated at many levels. A lot of people on here don't want to hear that because they have an ego about their exclusive coding skills. Your compensation will decrease. You may be fired. The coding ability of the average 'senior' developer is not impressive, and the title will not protect you.

phkahler2• 1 year ago

The only thing AI can code is glue code. But that is the majority of what people write these days. I've avoided that kind of work and AI disappoints me every time I've tried it with one exception - text to images.

philomath_mn4• 1 year ago

The only thing AI can code is X

I am curious, do you think it will stay this way? AI couldn't code anything 2.5 years ago, why should it stop with glue code?

Yesterday I fed the first 3 questions from the 2024 Putnam Exam [0] into a new open source LLM and it got 3/3. The progress is incredible and I don't expect it to suddenly stop where it is.

[0] https://kskedlaya.org/putnam-archive/

philomath_mn4• 1 year ago

I have a sinking feeling that this is correct but I'd love to understand the arguments for/against this position.

Do you have any podcast/blog/videos that you recommend which make your case clearly?

remoroid2• 1 year ago

Dont have any recommendations, I don't watch or read anyone. Think of it in two parts. You have the coding skill: knowledge of syntax, technical details, inputs, outputs, some math and logic, the infamous 'DSA' that millions of Indian people with no talent can solve. <- People pedestalize this.

Then you have the project manager / architect skill of shaping things and having an understanding of the tech stack - but does not need to be a great coder. Maybe they were in the past, but they're not grinding silly LeetCode questions and cannot solve them, and are smart enough to see that as the hamster wheel surrogate goal that it is.

I think we need less of the former, much less. I don't see it as a bad thing. I myself have severe carpal tunnel, RSI, and back problems, from sitting at a computer.

Almost all of the students that jumped into programming courses did so for money. They have no passion (less than me by far) for coding. Its a lie perpetuated by toxic positivity and invested groups that coding is some magical, super cool, super smart activity.

Its very overdue at this point. DeepSeek R1 and V3 are competing with o1, open source and local hostable. NVidia is coming out with small AI GPUs for consumers.

The winners of this AI age is everyone. EXCEPT 'coders'. Many getting 6 figures to maintain a CRUD app. I don't feel sorry for them! Cooking up garbage apps in PHP - I cannot wait for it to be overturned by energy efficient AI made apps in Rust.

jarsin• 1 year ago

AI devs are the new drag and droppers.

I remember all those devs who could only do tutorials using visual tools like VB6 and ASP forms.

Once the requirements went beyond the basics the project always got handed off to a real dev.

xena2• 1 year ago

I've ended up only using inline generative AI completions to do programming when I'm either on a strict deadline or making a demo for a conference/meetup talk. Otherwise I limit myself to pasting in/out of a chat window. This helps balance things out so that I can meet deadlines when I need to, both otherwise retain my programming/SRE skill.

That being said, one of my top google searches are things like "kubernetes mount configmap as files" because I don't quite do them often enough to have them totally internalized yet.

parpfish4• 1 year ago

I also use inline code generation, but it just seems like a more intelligent version of the code-aware autocomplete my IDE already provides. I love it, but it doesn’t feel like I’m atrophying because I’m still “driving”. It just saves a lot of keystrokes

_Algernon_2• 1 year ago

Tangent: I remember hearing of a sci-fi story where a city has been automated for generations and stuff is starting to fall apart while nobody knows how to fix it. Does anyone know what the name is? I have been unable to find it for a while.

xena2• 1 year ago

Blame! https://tvtropes.org/pmwiki/pmwiki.php/Manga/Blame

ThrowawayR2• 1 year ago

Perhaps you're thinking of Forster's "The Machine Stops"? There's been plenty of submissions to HN in the past, with links to the story text, and user comments: https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...

_Algernon_2• 1 year ago

This may be it. Thanks!

grahamj• 1 year ago

X to doubt. So far at least, to do anything nontrivial you still have to understand code. Sure it will raise the bar for what someone without such understanding can make, but a lot of that can already be done with code-free tools.

What most people who would call themselves programmers are working on is far beyond what AI can do today, or probably in the foreseeable future.

sys32768• 1 year ago

Eventually won't AI be trained on a massive corpus of source code and be tuned to actually program something as good as a lot of the existing bloatware out there that businesses pay out the nose for?

red_admiral• 1 year ago

There are a lot of questions here that also apply to teaching CS. Do we want our graduates to still know the old way? Perhaps for job interviews and days when GPT is down and for problems it can't solve yet.

The "no AI without understanding the solution" rule is a start here.

insane_dreamer4• 1 year ago

We’re not becoming 10x developers with AI.

Of course not. The ultimate goal of AI is to get rid of the developers altogether and reduce costs (more profits!). You think BigTech are spending hundreds of billions just to make developers more productive while retaining them? That's not a profitable strategy.

The copy/paste with LLM interactive stage is just a transitional stage as the LLM improves. We'll be past that in 5 years time.

No, AIs won't replace all developers -- you still need people doing the systems designs that the AI can implement code for. But it could easily reduce their numbers by some large percentage (50%? 80%?).

Edit: I would no longer advise my kids to get a degree in Computer Science.

philomath_mn4• 1 year ago

you still need people doing the systems designs that the AI can implement code for

Given the rate of improvement up to that point, how much longer do you think this would be the case?

insane_dreamer4• 1 year ago

I think that will get chipped away at incrementally, but you're right, eventually all you'll need is a list of requirements. I'd put that 30-50 years away, but some people might say sooner.

flerchin• 1 year ago

Whatever you do, you get better at. Let's not wring our hands and fret over the generation of programmers that don't write assembly anymore.

koakuma-chan• 1 year ago

First, I stopped reading documentation. Why bother when AI could explain things instantly?

Because AI doesn't have the latest documentation.

danielbln4• 1 year ago

In my editor I do /fetch <docs url> and can then happily ask the LLM questions and/or have it use the docs when writing code. We don't need purely rely on zero shot prompts and pre-training data.

EmersonL• 1 year ago

It's all in how you use it. Copy/paste driven development has long been a meme in the programming community.

For me, AI is like pairing with a very knowledgeable and experienced developer that has infinite time, infinite patience, and zero judgements about my stupid questions or knowledge gaps. Together, we tackle roadblocks, dive deep into concepts, and as a result, I accept less compromises in my work. Hell, I've managed to learn and become productive in Vim for the first time in my career with the help of AI.

ickelbawd• 1 year ago

Yes. But still you’re going to make AI tools anyway… ffs.

I think AI is just giving cover to all the incompetent people in your org, and in fact will create more incompetence via the effect you yourself noticed.

Eventually you’ll find many if not most of your coworkers are just wrappers around an LLM and provide little additional benefit—as I am finding.

matthewfelgate• 1 year ago

But is a programmer who uses AI to program the same as

  a Typescript programmer who doesn't understand JavaScript well? 
  Or a Python programmer who doesn't know C?
  Or a C programmer who doesn't know assembly code?
  Or an assembly coder who doesn't make his own PCBs?

Is AI orchestration just adding another layer of abstraction to the top of the programming stack?

weikju• 1 year ago

Not until their outputs become predictable and reproducible. Then a form of English or whatever other language becomes your programming language with its quirks and special syntax to get the right result.

abeppu• 1 year ago

I think there's a weird disconnect between the author's experience (the automated tools are good but they induce dependency) and my own experience (the tools are bad, produces code that doesn't type-check, produces explanations that don't make sense, or suggest back to me the half-broken code I already have).

I continue to believe that these tools would be more reliable if rather than merely doing next-token-prediction on code, we trained on the co-occurrence of human-written code, resulting IR, and execution traces, so the model must understand at least one level deeper of what the code does. If the prompt is not just a prose description but e.g. a skeleton of a trace, or a set of inputs/outputs/observable side effects, then simultaneously our generative tools only allow outputs which meet the constraints, with the natural and reasonable cost that the engineer needs to think a little more deeply about what they want. I think if done right, this could make many of us better engineers.

rsynnott3• 1 year ago

I think it's particularly weird that many of these tools simply do not seem to know about the type-checker at all. You'd think that'd be basic, at least for the likes of essentially fancy autocompletes like copilot; if the thing you're going to output does not type check, do not output it.

AnimalMuppet3• 1 year ago

That would require actually understanding types. GPTs don't understand anything - not really. They know the words and how the words connect to each other, and that's all.

What you want - what everybody wants - is a GPT plus. In this case, GPT plus a compiler's parsing ability. I don't think that's what we have, though. What we have is a GPT that was trained on more and better code snippets. That's not good enough.

hybrid_study• 1 year ago

Is this yet another thought that overemphasizes the immediate now? AI is developing so quickly it's difficult to see a whole generation of programmers being affected by negative transient aspects of this.

smitty1e• 1 year ago

The last guy who traditionally "knew everything" was Erasmus.

The shame is not in outsourcing the skill to, e.g. make a pencil[1]; rather, the shame is not retaining one's major skill. In IT, that is actually thinking.

[1] https://www.youtube.com/watch?v=67tHtpac5ws

tennisflyi• 1 year ago

Yes. Things can be too easy

vrighter2• 1 year ago

'Instead, I’m starting with “No-AI Days.” One day a week where...'

That's not enough. You will inevitably hit a problem which you will leave for "tomorrow" because tomorrow is not a No-AI day.

ivolimmen• 1 year ago

I have never used any AI for coding. It would feel like using side wheels on my bike. I have tested ChatGPT at one point but I did not even try coding stuff.

AI is creating a generation of illiterate programmers

Related Stories