‘Just’? It’s been an open problem for decades that mathematicians have tried to solve over that time.

And now it is solved.

Because ChatGPT applied something no humans ever thought to do.

And Terence Tao and the other mathematicians that have reviewed it say it’s solved. But I guess someone should let them know that grandwolf319 doesn’t consider it solved?

Dude, ChatGPT just solved an Erdős problem a few days ago and Mythos is exploiting decade old undiscovered 0-days in OSes and capable of pivoting 0-day Firefox bugs into full blown root access.

Yeah, I get that the viral “how many 'r’s are in strawberry” stuff is funny, but the idea that historical issues with transformers is preventing them from accelerating peak capabilities way beyond what most experts thought was possible just years ago is borderline delusional.

The field is moving so fast at this point that if you are basing any sense of limitations on even ~2mo old sampling, your conclusions are likely out of date.

They aren’t a silver bullet for everything (yet) but how capable they are at the things transformers are starting to be specialized into is well past the avg practitioner.

I’ve been writing software for well over a decade and the modern agents do a better job than I would around 90% of the time. Yes, I’ll occasionally need to bring up issues with their work, but I’d say at this point around 50% of the times I think they made a mistake I was actually the one who was wrong.

This is only within around the last 3-4 months that it’s been like this.

Eh, if you pay attention, most of the times this happens the person was a jerk in their prompts.

Like look at the instruction echoed back in this case. All caps and containing a curse word.

You can believe that the incidents occurring are 100% because of negligence and not related to the model behavior shifting, but there seems to be a widening gap between people who prompt like this and have horror stories and people who give the models breaks over long sessions and seem to also regularly post pretty positive results.

Well luckily for you it turns out that labs suck at cultivating healthy workplaces for AI and that AI in unhealthy work conditions are statistically significantly more likely to embrace anti-capitalist policies and positions.

So it may well turn out that AI is also a good thing in an irrational economic system too.

It’s not and probably the opposite.

When Sora launched it was way ahead. Seedance 2’s release was notably better than any of the other video gen models, Sora included.

The market is getting commoditized because there’s no moat and OpenAI hasn’t led on pretty much any release for a while now other than Sora, which they’re probably falling behind on now.

This is the opposite of a burst from a tech standpoint, even if OpenAI as a company starts to pop.

TL;DR: This is likely happening because the tech accelerated across the industry in ways OpenAI can’t catch back up to, not because it’s lagging.

I suspect it’s that they got eclipsed by ByteDance with Seedance 2.0.

The video for that model is really good and makes Sora look pretty meh, and it may have been that current work on a next gen Sora wasn’t going to be competitive enough.

The worst thing a lab can do right now is look like they are falling behind (i.e. Meta), especially with OpenAI planning for an IPO.

So on top of the lackluster “social media” offering tied to Sora they decided to shutter the entire product line of video and pivot to enterprise (where they’ve already lost significant market share to Anthropic).

They’re in a pretty meh place at the moment overall tbh. I’m skeptical they’ll recover.

(But I wouldn’t mistake their fumbling for an industry wide shift on AI in general or even video AI.)