StilesCrisis

No kidding! About twelve years ago, I outfitted a house with LED bulbs for about $5 a pop (a bargain at the time!) and those lasted for six years, until I moved. In the new house, LEDs were all less expensive but now they also fail quickly. I'd HAPPILY pay $5 for eternal bulbs again.

It was a made-up number.

I'm guessing the Base and Pro chips come off the same wafer, and they get 5 Base chips (some defects) for every one Pro chip (very few defects). So they run the Base-model assembly line for six months and eventually they accumulate enough Pro chips to satisfy Pro-model demand.

Their posted videos show plenty of failures. Read the rest of the comment section.

I agree but I feel like the exaggerated physics were intentional and fitting with the video theme. If they had wanted it to topple, they wouldn't have posted the video.

The non-coherence of the world is a pretty solid tell. Any time the camera pans away and then back, the scene changes. The model doesn't remember what used to be there dozens of frames ago and so it invents something new.

It kept on going, but they have been far less stable from a technical OR leadership perspective. They had a user exodus. They had advertisers leave in droves. I suspect it was a net negative change.

It doesn't say "everyone in sales" so it's not clear.

This gives visibility to a lot more unencrypted code, so just give it time. More bugs will be found quickly now that the code is inspectable.

Definitely a risk.

Yes? Very little about Claude is deterministic. If given the same sufficiently complicated problem, but via slightly different input prompts, it can generate entirely different code. It would be shocking if they all had a single shared bug.

> in Kyoto, in a language I don't speak at all, coordinating with an obscene amount of moving parts and logistics, flawlessly handled entirely by Dot

I wonder if the folks in Japan would agree with that assessment. They would absolutely not tell you to your face if it were less flawless than you think: that would be rude!

I think RAM turned out to be the secret sauce all along!

Facebook is developed on a reimplementation of PHP!

Ironically, I'm sure Claude could have spotted all four.

> leaving a features behavior undefined

Missing an apostrophe? A human wrote this! Love to see it.

I can relate! I once joined a team that worked on "Analytics" when I was naive and didn't realize that meant "remarketing and ad targeting."

Just keep an eye out for something different. I'm sure there are jobs that you can do and feel good about.

So he's basically signing up to do your jobs once you're gone? Or he thinks he's also expendable? Either one doesn't sound too good.

Get resumes out.

If there's a treasure chest at the top of the mountain, no one cares how you got there. It's not always about the feeling of accomplishment.

If one treats all backsliding as permanent and all progress as temporary, their mind is already made up.

Look, I get it, we're clearly in the backsliding zone right now, but midterms are here.

I mean, the Flock cameras coming down is meaningful progress?

This post deserves to be flagged IMO. It's Claude output straight from the tap.

No, and that's one of the reasons mathematicians dislike automated solutions. Sure, it's verifiable, but it's too complex and messy to actually learn anything useful. So it technically answers the question, but doesn't leave you with any deeper understanding.

They were mocked because they became mentally detached from the actual rules of the system, and because they weren't in the trenches landing PRs. Trust me, I know, I worked with one. (He was also a huge proponent of AI)

One real limitation is context size. 1M tokens is enough to do a lot, but the Lean graphs they are making are way too big to fit, so an LLM will never have the complete picture.

> Imagine waking up one morning to find that an alien technology has solved the problem you had dedicated your life to.

Pretty sure everyone on HN can relate to this.

I'm past the point in my career where I need to worry about learning C++. The rules of the system are what matter.

They are also languages which humans basically won't use directly, and have almost zero representation on GitHub etc.

If the training set were loaded with Forth, LLMs would be knocking it out of the park.

In reasoning mode, if Forth were in high demand, I expect a modern LLM would blast out a python script to help it reason through the stack. That's how they seem to handle most problems that they don't innately understand.

Again, it sounds like you are against vibe coding ("stop working at the site") and so am I. There's a middle ground worth acknowledging: we've built an automatic wire pulling machine and there's no reason to hand-pull anymore unless it's a really weird bend. If you just load up the machine with a spool and come back in 24 hours, though, it'll make a mess of wire all over the place.

So they believed they had succeeded at the goal, but failed? What goal did you set? What happened instead? Did they have a mechanism to check their work?