OpenAI: DeepSeek "found some of the core ideas that we did on our way to o1"

twitter.com

9 points·modeless··6 comments

I feel like this tweet and Altman's tweet (https://x.com/sama/status/1884066337103962416) have a vibe of "our investors are mad and told us we had to say something". Are people getting cold feet over the $500B?

> we will obviously deliver much better models

It's not obvious Sam, that's the problem.

If OpenAI is so smart, how come they got beat? Altman can't stop making excuses.

They were first, which is a huge accomplishment. History will remember the ChatGPT launch no matter what happens next. However, first is not necessarily best.

> On March 8th, after an independent investigation into his sudden firing, OpenAI reinstated Altman as a member of the board, along with three other additions.

> The removal was caused by concerns about his handling of artificial intelligence safety, and allegations of abusive behavior.

Something happened at OpenAI when it became a for-profit company.

The models were build with public data because it was for scientific research (and allowed legally to do so). All changed when they got the dollar sign eyes. Since then they are just another company trying to hype the market instead of building better products. The CEO leaving and coming back was just part of the turmoil. If he has abusive behavior that is the best way to lose talent and reduce the output of the talent left in favor of marketing and hype.

This feels a bit like the Teller-Ulam design / Sakharov's third idea.

Reply on news.ycombinator.com