ryanmcgarvey

Good for you guys. I'm happy for you.

We joke, but I bet this will help to drastically reduce ScamInc's revenue.

Everyone that is upset about this should take note: you are not a (coding) customer at $20/mo. Their coding customers spend thousands per month (week!) on claude and it's growing faster than they can keep up with (source: I'm one of them, and I know many other like me. We're budgeting 10-20% of engineering salary spend on tokens). It sucks to no longer be able to code on the cheap anymore, but don't fool yourself into thinking you have any leverage here.

It's not the dictating, it's the clicking and the finding and addressing and the formalities. For a transactional email it's a lot, especially if there's substance to it like needing to click a link and make a choice about some options.

This is doing a bigger number on me than it has any right to.

...why are they all skewed, save for the buns that are already lopsided? Those I'll note are perfectly seated. Some are more skewed than others. Like the Big Mac is only slightly skewed.

Is there a pecking order to how skewed they are? Some social hierarchy of sandwiches?

Yes. I had a spare M1 Mini so I decided to set it up. YOLOed the entire thing and connected all the integrations, though I only ever use Opus/Sonnet. I have a dedicated Discord server I use to communicate with it.

It really is eye-opening how powerful it is once you connect your life to it. The biggest improvement is actually the trivial stuff — emailing contractors, accountants, etc. I no longer need to open another app, navigate six separate steps, and organize everything on the fly just to send one message. I stay in one place, organize my thoughts about what I’m trying to do, and the execution happens automatically.

For example, open claw will find the relevant threads, lookup the relevant details via web search, present them to me, give me a draft, I can review it, edit it, and send. 2 minutes instead of 20.

As a software engineer using AI daily, I think this is the real unlock — staying in a single context and not getting lost in the long tail of trivial details that fragment your attention before you ever get to the actual work.

The mental model shift is important too. It’s not that AI does all the work and we end up with agents emailing each other. It’s that AI automates the steps required to achieve your goals, so you can focus on the substance — like actually writing the email instead of navigating to the place where you write the email.

I also have GitHub and Vercel wired up, which means I can jot down an idea for a little productivity app for me and my family and it just appears a few minutes later, deployed and ready to use. That alone has been worth the setup.

AI didn't do the work, I did. Building up context is the part we actually have to put work into. I'm not saying it would be impossible, but boy would it be annoying to have to constantly reach a new assistant about your whole life.

Memory. I have built up so many scripts and crons and integrated little programs and memories with open claw it would be difficult to migrate to some other system.

System of record and all.

The full on dystopian take would be to require anyone receiving welfare or other public funds to fully disclose all of their private details.

You want Medicaid? Tell everyone about your hemorrhoids first.

The conversation around GPU lifecycles seems to be conflating the various shear rates within the data center. My layman understanding is that the old 3 year replacement cycle had more to do with some component, not necessarily the memory or the processor, going wrong for half of their units by 3 years, at which point GPUs were cheap enough and advancing faster enough that it was more cost effective to upgrade than to fix. However, that calculus changes completely when the GPU and the HBM are orders of magnitude more expensive than the rest of the system. I suspect that we will see repairs being done on on the various brittle bits of the system and the actual core expensive components will continue to operate much longer than 3 years.

To be clear, that is the advertised effects, not the side effects. The idea is that if you over eat you feel terrible, but also you won't want to overeat since you'll be full.

The only reason I want these things to be any smarter is because I need them to do more work over longer periods screwing up. The only reason I need them to do more work over long periods is because they are too slow to properly pair with.

If I could have it read more of my project in a single gulp and produce the 10-1000 lines of code I want in a few seconds, I wouldn't need it to go off and write the thousands of lines on its own in the background. But because even trivial changes can take minutes by the time it slurps up the right context and futzes with the linter and types, that ideal pair programmer loop is less attractive.

Your description of "little toys" and "immense power of AI" seem to be at odds in your argument. Which is it?

I don't think it will come all at once. By far the part no one wants to do and is also easiest for AI to tackle is long hauling. Humans will still be required for the last mile and delivery. I can envision a world in which we have "more" drivers because of AI not fewer.

I presume they mean that distribution is king and they make all the devices.

Still can't show me a clock that isn't 10:10.

Otherwise impressive.

Is it a political statement if it's also a statement of fact? Sure, the comment has some color to it, I'll concede that, but one can no longer post these kinds of things on Twitter and get the honest engagement from community members one used to. It's no longer a welcoming place for this kind of discussion.

In America you're free to sign or not sign terrible contracts in exchange for life altering amounts of money.

Maybe it's worth reevaluating our definition of quality?

In a world where AI can read our codebase, ingest a prompt, and quickly output "correct" if not clean and concise code, and then be able to iterate on code with more prompt, do we need all the same patterns we used to adopt when humans were painstakingly writing every line of code?

This reminds of of the CISC to RISC migration - now that computers are in the loop writing the tedious parts, we don't need to burden our codebase with patterns meant to relieve humans from the tedium.

I find myself, for instance, writing more long form, boring configuration files that once upon a time I would have built some abstraction reduce the boilerplate and verbosity. But now that co-pilot can just auto-complete the next section for me, why bother?

Isn't this (part of) the point?

We have an abundance of parking required by law and not enough housing units - removing the requirement should encourage more money to be invested into anything but parking no?

Seems like letting businesses decide whether or not it's worth opening a business with the available parking would cause one of two things to happen:

1. We wind up with the actual correct amount of parking required for a given set of businesses - instead of some arbitrary amount dictated by legislatures decades ago. 2. Everyone is tired of the lack of parking and votes to increase public transit funding.

In my experience, working from home is similar to many of the other skills we hone in our careers.

Yes, the tools you use will impact performance and quality of life. Invest in these. But also, getting better at video calls helps with performance and quality of life.

I have found that when we treat video calls as real time conversations, we get frustrated with latency and the lack of proper audio mixing.

However, if I and the people I video conference with obey certain protocols: 1 person speaks at a time, raise your hand to speak next, etc, many of the frustrations that cause fatigue go away.

It still isn't a replacement for the real time, able to talk over each other style that we might prefer, but at least it's productive and less of a drag.

I'll note that when we adhere to these protocols, we find not only are we less fatigued, but the quality of communication sky rockets. I think this is a result of everyone having more room to think and being forced to prepare your thoughts rather than reacting with whichever thought comes first.

Regarding the "they're just trying to get us back in the office" debate - I'm hopeful that genZ will come fully immunized against zoom fatigue and it'll be less of an issue.

That's the problem with software - it's decisions all the way down. Outside of prototyping, there's very little "work" to be done that doesn't contain some kind of systems decision.

In my experience so far it is much more along the lines of the horse and rider than it is a useful decision maker.

I'm still in charge of telling it where it go, but it takes me there. Knowing where to go and why is the important bit tied to wisdom.

I think the AI driven development story is more about leverage than it is about wisdom. Leverage from AI is derived from being able to do more with fewer people.

In my experience the great slowdown of growing companies comes from hitting the communication barrier on their products - the point at which the majority of effort is spent coordinating work rather than doing work. I find that the path from majority focus on product to majority focus on coordination isn't linear, but rather more of a watershed. One day you are 80/20, the seemingly overnight after some growth you are 20/80 the other way and never look back.

The advantage of being on the right side of that watershed is that you can maintain velocity and agility. Not only can you iterate quickly, but you're in a better position to change course and rebuild as needed. The left hand and the right hand require little effort to coordinate and get it done.

Larger companies live and die on their ability to either find a moat large enough to protect them, or build organizational structures that let them keep scaling. It takes decades to get the culture and processes right and baked in across the board for a large company to be able to maintain any velocity and reinvent itself.

This is where the gap is. Being small is easy, you simply don't have the coordination problems. But the moment you hit success and need to grow, you immediately are at a disadvantage compared to the big incumbents who have had decades to refine their coordination systems.

The extent to which AI can provide more leverage to smaller companies allowing them to "grow" without actually crossing that coordination watershed, they will be in a much better position take on the incumbents.

Forgive me, I may not have worded that correctly.

By empty, I simply meant that the game stopped being interesting after a while. The single player game was only so replay-able, and multi-player is a very different game that doesn't suit me.

What Civ has is a great community, but that community had no way to experience content together.

Hand crafted scenarios that people can play simultaneously and share their experiences isn't a multiplayer element, it's a community element.

My biggest gripe about the Civ series is that it felt empty after a while. I love hearing the stories other players experience when playing, but there was never a good way to recreate and share those experiences.

What Civ needs is challenges and leaderboards. Pre-configure scenarios that everyone can play and compare their experiences together. Score each play on a number of criteria (cities built, fastest tech, army size, etc), and maybe even have custom achievements for that particular scenario - like how fast you fully explored the particular valley the scenario starts you in, or how quickly you defeated the aggressive city state that starts next to you.

Not only should randomly generated maps be shareable, but Sid Meier should have weekly scenarios that involve hand crafted challenges. Not just normal games with the AI trying to act like a human, but the AI existing as part of the map. Example: at turn 50, an aggressive hoard comes down off the mountain attacks your civ. When you reach the other side of the mountain, there will be a Civ with 4 cities and an advanced army. You have certain diplomatic options available (like a quest to fulfill) or you can try and challenge them. How you handle these various challenges will impact future scenarios.

The important thing is that the AI isn't trying to play the same game. It's just there as part of the map to give the player an interesting experience.

This dilemma is precisely what people who benefit from remote work went through when most companies weren't remote friendly.

My honest opinion is that most people that enjoy the office enjoy it because it's where they've gotten most of their socialization in their adult lives. As a result, they haven't built up lives outside of work. To make matters worse, our communities haven't developed to support people having "lives" outside of work (at least out here in the suburbs).

My hope is that as remote work becomes more prolific, people will rely on their local communities more for their social lives and all of these wonderful small towns will thrive again.

Many (most?) software projects exist in environments that demand predictability. Businesses need to be able to make promises to clients and potential customers in order to make sales and garner trust. In any remotely competitive industry these are two absolutely essential ingredients to success.

I mean, think about the software you use. Of all the shops that produce that software, which ones do you trust more than others? I'd be willing to be it's the ones that delivery predictably. At the very least the ones that deliver updates reliably - as in they do what they say.

Granted, every industry is different and some rely on trust and predictability more than others. Building an organization that is good at that is very difficult - and that cost has to come from somewhere doesn't it? In my experience it's at the cost of time spent building the product. A tradeoff between reliability and quality, if you will.

Forgetting it’s a skill

Pair programming is a skill which must be learned.

You will not be good at it at first, but consistent practice will yield improvements.

Don’t give up after a difficult first experience. Don’t assume experienced developers are automatically good pairing partners. Don’t expect to be good without practice.

Consider reflecting with your pair or asking for feedback after each session. What could have been better?

In my experience (pairing on and off for about 12 years) - this is the biggest thing people misunderstand about pairing. You're not going to be good at it right away, and the team isn't going to benefit right away. It's an investment that pays off in months not days.

That being said - I find that requiring my team to be good at pairing, but not that they pair all day every day, is sufficient. It means that they pair when it's appropriate (onboarding, larger design problems, early stages of a project, etc) and they don't when it's less appropriate (fleshing out an implementation, exploring a new idea, mundane updates).

We find that if we pair all the time, especially on things that don't benefit from pairing, it exhausts us. Pairing is a useful tool, but it can be equally dangerous if overused and cause people to resent it.

This is an important point - developers have a tendency to focus on our individual performance rather than the performance of the entire organization. In my experience, while individual performance goes down for some with pair/mob programming, both quality and velocity go up for the whole organization.

And it goes without saying that we developers love our freedom because it means we can work at our own pace. Often I find when I'm pairing all day I'm putting in way more focused time than when I am soloing.