bcrosby95

City of Stairs

This ties into my experience: greenfield projects turn bad way faster than legacy ones, because AI has no established patterns to key off and it does increasingly dumb shit.

This is why it's good to not keep your prototype a pile of shit as it grows to 5k, 10k, 50k, 100k lines of code.

The US has papered their problems over largely because of its wildly successful service companies. Who knows what that looks like with the increasing utility of LLMs though.

ASML is selling $300 million dollar machines that are fairly small for it's price point. Regulation is going to have very little impact on producing something like that vs mass produced widgets, cars, or chips.

People used to say someday you can sit around all day and create things. Now they say you can go into the trades and build someone's house or something.

Will do. I took your advice and asked Claude what it thought and it said you're exhausting.

Pointing to luck renders all advice meaningless.

I think it's more important. AI gets a lot right, but sometimes it gets things wrong. The document might be the only human authored piece of text, and it will help future agents see that something is incorrect in the implementation.

I've never been much for design docs. But I have some projects that are on again/off again and I can lose my train of thought. It seems like having one of these could help a lot with that.

> normal people feel, reasonably, that their civic responsibility is to vote in actual elections for the policy they support, not to spend their life in City Hall jawboning.

That's not how the vast majority of this works, and the average best case scenario on this you get to vote a politician out after they already fucked people over.

Yeah, we need to go back to names like International Business Machines, fuck this immature "Google" or "Yahoo!" nonsense. Hell, Palantir is named after something in a book for children!

Give me something reasonable like Global Information and Retrieval Systems Inc or Worldwide Computational Services Holdings.

It's pretty clear once you click on it:

> Allow your content to be used to train our models, which makes ChatGPT better for you and everyone who uses it. We take steps to protect your privacy. Learn more

Yes, my favorite codebase to work in is one full of accurate and long comments that detail that making a function call it never called would cause problems because of the details of a function that gets called 3 layers deeper.

Meatspace is hard though. But if LLMs solve the virtual part, maybe we can iterate quickly on robots.

Also, if you think it's annoying when Claude goes down while coding, just wait until a robot is in the middle of fixing a leak it just caused.

Why would they do that? It wouldn't accomplish anything useful.

The first thing LLMs seem likely to automate is automation itself. I'm curious what the past few hundreds of years of industrialization would have looked like if the first thing they automated was building, designing, and running the factories themselves.

So we're going to make government weaker, and by doing so, they will stand up to corporations that are now stronger than them?

If that's true then Libertarians have a rosier view of government than I do. A regular David vs. Goliath situation. Biblical stories aside, my money is on Goliath.

Never. It would be like laughing at people writing Prolog calling themselves coders.

People have been "writing code" like this for decades. That a programmer happens to do it doesn't turn it into writing code, no one would have made that claim 25 years ago, and people who aren't programmers wouldn't make that claim today.

You claim writing is thinking, but imply writing code isn't thinking.

My opinion of LLM design review isn't that high - it seems to miss design tweaks that could vastly simplify corner cases. But if your code isn't written for human consumption maybe it doesn't matter. I'm still directly responsible for what I commit, so I can't just offload it to Claude.

They can help with it if you point them in the right direction.

Yeah, I already don't remember what I had for breakfast two days ago.

Oh that? That's just Claude being the sassy asshole it is. It loves to write in a way with maximal self-inflating impact.

Yes, OOP is hard to do right. It's not helped by examples and schools largely teaching from real world nouns. Most people in OOP languages are writing procedural code.

For all these reasons I just skip it. I still use OOP languages, but I leave the OOD at the door.

For what it's worth, you can get all that with plain old C. Declare structs up front. Allocate them all as pointers. The C file is the class, and has the actual definition of the structs.

Strangely enough, in a sense, OOP languages makes it harder to do this than in C. Not that C doesn't have other flaws.

The way I usually prevent having to scale out to a bazillion systems is never getting more than 10 users.

This post seems very concerned about what a product does - and rightly so. However, it doesn't really talk about figuring out what a product doesn't do. This is a blind spot large enough to fly a 747 full of CVEs through.

The first thing this wave of automation seems likely to replace is automation itself.

For this reason it's hard for me to give much attention to comparisons to previous industrial revolutions. It would have looked way different if the first thing they built was a factory factory.

I will give this post some slack though, it was written in 2023.

I think if they are intelligent, intrinsically motivated, and willing to pursue knowledge beyond what the class requires, it probably helps.

That last part is key. Intrinsic motivation doesn't mean you pursue it outside normal bounds. My kid loves soccer, its her second favorite thing in the world, she has an absurdly high tolerance for physical discomfort while playing, but she doesn't play it at home. There's other things she's rather do, such as play with her toys.

When you move the bar to something even less interesting to most kids like science, you're going to have a pretty huge falloff. You're basically selecting for kids who choose to do it in their spare time. I know a lot of smart kids (I run a boyscout troop, my wife a girlscout troop, both with lots of high achievers), and none of them do this.

> especially as the session drags on.

This is because these harnesses are missing a very important feature. Anything like this needs to be included with every turn, otherwise the LLM quickly drifts.

I first noticed it when I wrote a harness for D&D (because it's so damn noticeable there), but now I include this for any harness I write.