IshKebab

> The hardware team at Jane Street designs FPGAs and ASICs that run some of the fastest trading systems in the world, and the day-to-day work is full of puzzles just like this one: staring at a sea of gates, timing reports, or waveforms and slowly teasing out what’s really going on.

Although not really any more. AI is really good at debugging failing tests so I no longer have to spend hours - sometimes it was even days - staring at waves and instruction traces.

Tbh that's one thing I won't miss.

Tbh it doesn't even matter if humans turn out to have a soul, or quantum microtubules or whatever other magic LLMs can't have.

The normal definition of the word "thinking" definitely includes what LLMs do. Hell people used to say computers were thinking even before AI. It's super weird to get all uppity about the semantics of the word now.

There is no argument. It's just people's reflex reactions.

The technical excuses they come up with (e.g. that the server can detect it and send different content) are just post-hoc justifications for their instinct.

Just ignore them.

Uhm... emit the errors and fail the compilation? What else would you do?

Maybe there's some work to do to make that happen cleanly but the desired behaviour seems pretty obvious.

> If I make an AI capable of self-sustainment on the internet

I'm really surprised nobody has done that yet. With how cheap AI is to run these days it would only need to make a small amount of money (e.g. through hacking).

Someone should set one up with the long term goal of getting egg on LeCun's face.

> These models aren’t thinking.

They are for any definition of the word that makes any kind of sense. I'm sure you have a contorted definition that magically only includes humans though...

> I'm guessing b/c it can be used to make merchandise

I doubt it. How much merchandise relies on high quality 3D scans?

I think the real reason is that a lot of people are naturally paranoid control freaks. Their instinct is that these scans should be secret and they never really think beyond that.

Seems like an easy win! I'm kind of surprised nobody did this already. I guess someone will need to reimplement this by hand given Rust's AI policy, but this is still great because it demonstrates that it's a really good idea.

Yeah AI is extremely good at reverse engineering compared to humans. It's one of those things were there's a huge amount of tedious but not super complex work. I think if you can set up a good harness for the agent it can probably write complete drivers.

The most difficult part is likely that a lot of hardware is easily brickable if you do the wrong thing, so I'm not going to let it loose on e.g. my solar inverter.. even though their software is shit and I'd love to replace it. (Don't buy QCells.)

Ugh guys... The front page of your project is the most important thing there is. If you can't even be bothered to write it yourself then I don't have much hope for the rest of your project.

I would imagine they're thinking it increases sympathy with the victim, which doesn't benefit the killer.

But yeah it would be funny if this backfires. 10 years for murder seems like a great deal for the murderer!

This is what let me understand lambdas. Great tool.

Is this a unikernel? Your ASCII art diagram is broken.

Fax was never cool.

And `cp`.

I think the way to handle this is to just feed it into another AI agent (a better one) and ask it how serious the issue actually is. Fight fire with fire!

Gotta admit it's weird that they haven't set up Claude to fix them! Tbf my threshold for "AI now writes good enough code" was Astra or Opus 5.5 and they only just released that.

Ha funny but no.

Nobody is one-shotting big projects. Even humans can't do that.

Wow I thought surely they wouldn't object to using AI to confirm bugs, but they really do.

Tbf I guess as a popular open source project not using AI to fix bugs, they probably already have more open bugs than they can ever fix so it doesn't really help them for people to find more.

I would imagine his bug was actually ignored just because Zig has 2700 open bugs, rather than some AI policy violation.

Uhm I'm pretty sure a severity score is supposed to be a score of severity.

Yeah maybe it made sense in the past, but not in the age of AI.

Don't forget the training data! Legal copies of all the books in the world, the entire web scraped, and all of YouTube.

Astra also really loves reverse engineering binaries. I guess it's one of those things that isn't that complicated but is super tedious, and tedium means nothing to AI.

IMO TLA+ is not very good. It has super weird syntax, and a whole separate DSL (PlusCal) to give it workable syntax for programs. You can really tell it was created by the same mind as LaTeX.

What is the Typst of formal modeling?

Another issue is that you end up with a formal model that passes, but then you have still have to convert that to a real language by hand and not make any mistakes.

How do managers build software?

The fact is when LLMs get good enough you WILL be able to build software without reading/understanding the code.

Whether or not you think we are already at that point is kind of an unimportant detail.

I would say we are quite close, depending on the type of software you are building.

I would say the ability to actually use software is kind of relevant to said software. There's definitely software I wanted to use but gave up on because it was just impossible to actually install. Hell I just spent half of today coaxing Claude to fix some Kotlin software so that it worked with the version of Java I have. Let me tell you that would have taken a good few weeks without AI, even if I was unfortunate enough to be an expert in Maven and Gradle. "Written in Rust" would have saved me a huge amount of (pre-AI) time for that software.

Rust matters because it means you can actually deploy it without going insane. I wonder if the anti-Rust zealots have actually ever used pip, especially on Windows.

He could have said "... not written in Python" - would that have been acceptable?

Yeah likewise. Pytorch needs to die.

That said I don't know what's wrong with using a Rust AI framework like Candle.

Maybe ideas aren't so cheap anymore...