ModernMech

Used to be they couldn't write CMU on the shirts over a dispute with Central Michigan University.

I would say probably most memeing is not really thought provoking or inducing. I agree most of the time writing can help collect one's thoughts but it can also be completely thoughtless. Probably what matters more is the specific process.

"Writing is thinking" is being repeated all the time lately to the point it seems like a meme.

Did you think about this comment or are you just repeating something you heard and that’s been said in every discussion on HN about writing and AI?

Same. 5.5 got work done then 5.6 was also fine then 6 was maybe not quite as good. Now with 6.1 they are cutting usage and raising prices and introducing ultra fast mode, but things were good enough 5 months ago.

The idea is you'll always be paying that price, they'll just shrinkflate what you get and not tell you by how much. This has already been happening, now they're being a little more explicit about it.

> 500$ is a old 200$ plan plus an option to drain it absurdly fast?

Yes.

> What is the value here of a 500$ plan I can't figure it out.

If you have unlimited money, it lets you give it to OpenAI to get faster results, presumably because you don't have unlimited time.

It’s meant for rich casuals. If you watch the presentation, everyone they depicted using it seemed to be in some high paying which collar profession. I.e. it’s for people like the people who work at open ai.

lol yeah, I opened the Uber app the other day to $100 fares, remembered a time when I could get anywhere in SF at any time for like $5.

Honestly I feel like 90% of the improvement has been connecting them to deterministic tooling that can tell them when they're wrong. Whenever I chat with one of these latest frontier bots, they still exhibit all the same problems they used to, But throw them in a tight loop with tools that connect to reality, they adjust their behavior accordingly. It's like SLAM vs. dead reckoning for anyone who knows robotics.

It's very amusing how for a decade, The Way (TM) has been "move fast and break things", but now with LLMs it's suddenly "woah, AI is moving so fast, it's going to break things!"

Very true.

lol I got flagged for using the word fuzz, not even in a security context (it was a parser so security adjacent but still).

> With your own code you reasoned about it which contributed to its stability.

Okay but to what extent? People say this but there's no way to measure it really. Did you live through the 90s? People reasoned through all that code and it was very often quite unstable. I'm sure everyone involved with Windows ME reasoned about it quite a lot, probably elements of it locally were very sound, yet in totality it was an unstable mess.

What fixed that situation wasn't that engineers today are reasoning better than engineers in the 90s, but IMO better tooling. Which brings me back to: your codebase is only as good as what it can prove. If there's any question, I just show you the proof rather than appealing to my reasoning being sound.

> With unreviewed AI gen nobody ever understood or reasoned about the code, including the AI.

And? You haven't established reasoning about it is actually necessary and it certainly isn't sufficient.

On the one hand, you're exactly right.

On the other, my local pool company is hiring a software engineer and hardware engineer because with AI, they can replace a 2 pizza team as you so succinctly put it. So no two pizza teams but that doesn't mean all the pizzas are gone, they're maybe going to be spread out and not concentrated in CA, between orgs you might not have thought as "tech" before.

> to instead modularize with succinct interfaces, so that you can reason locally

Okay but how does AI change any of that? You can still do that with AI.

> as long as it’s sufficiently documented.

AI definitely helps with that.

> What is essential is that for every part someone did reason through it with the necessary rigor at some point.

Why is that essential though? What if the person who reasoned about it dies or leaves? Moreover, why is it imperative the reasoning happens at the source code level?

Your code is only as good as what you can prove. Understanding the code is not the goal, it’s only important insofar as it helps you evolve the codebase predictably and without bugs or regressions, and understanding is not easily measurable or transferable.

Moreover, when your codebase is hundreds of thousands to millions LOC, I question how much you can ever truly understand it at the level you’re saying.

The main problem I see with RC which has prevented me from applying is there’s no stipend or housing. Most people can’t afford a summer in NYC without pay. So I’m not sure who the target audience really is, I guess wealthy SV engineers.

This started as soon as they devised the MCP. “we have this new thing called ai you can talk to it naturally. Here’s a protocol to make it work better”

We are entering a phase where ai creates clean, informative interfaces, so humans trying to differentiate make theirs “interesting” (unreadable), as in this case: “Mhat Js Glop”

This is the fault of capitalism, under which you are required to sell your time and skills for cash to survive. Under this system, no one who isn’t independently wealthy can be an artist without commoditizing their art. So that’s the lens through which people view art: just another commodity to sell so one can survive, not unlike grains or oil. The quote also shows that the artist is fungible in this conception, only the output matters no matter how it got there, as long as you can sell it.

Are they changing or converging? If you can pinpoint where they are going you can get there before they do.

True, it’s also trained on millions of papers of programming language research, so you can write a language to operationalize that research and it does great with it. The millions of lines of typescript scaffold the syntax, the PL papers scaffold the semantics.

No, it’s not different now because you’re conflating two different things. One is the first amendment cited in Twitter case. The other is free speech cited here.

Did Twitter violate the principles of free speech by shutting Trump down? Well he was using his speech to foment an insurrection and caused a mob to sack the Capitol so yes, Twitter, within the bounds of the first amendment, for an eminently justifiable reason (to preserve democracy), attacked Trump’s speech, which is against free speech, yet preserves free speech, because the alternative was allowing insurrection to continue. The first amendment cannot hold under a dictatorship, as we are seeing right now as all press are (were? Are they complying with the order?) illegally banned from the White House.

It’s not straightforward but it’s consistent.

You’re conflating the first amendment with free speech generally. They are different (related) concepts. The person you’re replying to did not appeal to the first amendment.

"But I still took time to thumb it out on my phone keyboard"

Is that the part that matters when it comes to writing, the mechanical work of pressing the keys? Not the thinking? What if someone generates some AIslop text, and then re-types it out with their thumbs, is that then worthy of your time to read it, simply because some person felt strongly enough to retype it? Because that's basically what you did by re-posting memeslop.

Moreover, let's say you read something and you learned something from it and actually enjoyed reading it, but only afterwards find out it was written by AI. Does that mean you had wasted your time?

"Long walls of AI slop flip the equation"

Not really, there's always been long walls of human slop text to figure out if there's thinking behind it. You ever been manic or known someone with mania? Got ADHD and racing thoughts? The vast output of corporations and politicians? Nuff said.

Is there more now? Probably but it doesn't "flip the equation"; that was true after radio, tv, the internet, heck even the printing press. It's always been incumbent upon the reader to figure out the intent and quality of a piece of writing, no matter the source, and that's never been an easy or automatable task. Inb4 "but it used to be that writing something at least indicated a modicum of thought" memeslop, because that actually isn't true for the reasons I stated already.

You don’t, that’s why in agentic world your codebase is only as good as what you can prove. For this reason, you’re going to see more languages evolving feature like capability permission, effect/coeffect types, refinement types, formal verifiers, strict type checkers, static analysis and so forth. Tests are only a small part of the verification. These ideas are old and have sat outside of the mainstream coding world, but the value proposition in the ai age has changed enough their relevance is renewed.

It’s equally good at catching mistakes as it is at making them. So you put it in a loop, write code, make mistakes, catch mistakes, write code, make mistakes… etc. Eventually it converges and stabilizes. That process can take weeks. So the question is: when is it worth it versus not?

I think that movie is Robocop.

These days I only use AI once a week the way Astra chews through usage. “Congrats your usage has reset you can get back to… aaaaand it’s gone”