They might also be focusing all their work on beating Grok 4 now, since xAi has a significant edge in accumulating computing power and they opened a considerable gap in raw intelligence tests like ARC and HLE. OpenAI is in this to win the competitive race, not the open one.

> They might also be focusing all their work on beating Grok 4 now,

With half the key team members they had a month prior

I'm starting to think talent is way less concentrated in these individuals than execs would have investors believe. While all those people who left OpenAI certainly have the ability to raise ridiculous sums of venture capital in all sorts of companies, Anthropic remains the only offspring that has actually reached a level where they can go head-to-head with OpenAI. Zuck now spending billions on snatching those people seems more like a move out of desperation than a real plan.

Meta/Facebook waiting until they're out of step, then trying to hire the leaders has been their modus operandi since they missed on mobile photo sharing and acquired Instagram.

They've kind of played themselves making "genius engineers" their competitive advantage. Anyone can hire those engineers!

At this point, it seems to be more engineering throughput that will decide short to medium term outcomes. I’ve yet to see a case where an IC who took a position only because of X in fact outrageous compensation package (especially if not directly tied to longterm company performance through equity) was ever productive again. Meta certainly doesn’t strike me as a company that attracts talent for their “mission.”

TLDR Zuck’s recent actions definitely smell like a predictable failure driven by desperation to me.

Dont be mechahitler?

simonw is going to force every competitive LLM to over-ingest cartoon svg pelicans before this is over!

Cue unpublished battery of '{animal} riding {motiveDevice}' real benchmarks behind the scenes.

You forgot the on a bicycle part.

This could be it, especially since they announced last week that it would be the best open-source model.

Technically they were right when they said it, in their minds. Things are moving so fast that in a week, it will be true again.

Btw why is there no k2 discussion on HN? Isn’t it pretty huge news?

There is, but it’s not on the front page so you don’t find it unless you go through multiple pages or manually search it up.

Moonshot ai has released banger models without much noise about it. Like for example Kimi K1.5, it was quite impressive at the time

probably because maybe 1 or 2 folks on here can run it? It's 1000B model, if 16bit training then you need 2000b of GPU vram to run it. Or about 80 5090s hooked up to the same machine. Or 20 of them to run it in Q2.

Still more accessible then fully closed models.

...but not widespread enough impact/interest to reach/stay on the HN-frontpage.

Infinity minus one is smaller than infinity?

I think there are alot, like alot of alot, of shills pumping up hype around commercial SaaS models. Deepseek was ignored for like forever until it became embarrassing.

I even made a post trying to break the surreal silence.

https://news.ycombinator.com/item?id=38505986

had to search for the discussion, it's here, seems like nobody noticed it and it only couple hundred upvotes.

Here: https://news.ycombinator.com/item?id=44533403

Every openai model since gpt4 has been behind the curve by miles