This should tell you how unserious these "companies" are when they consider themselves the grand harbinger of the one and true "AGI" because their frontier models were able to solve (with great assistance) some sort of conjecture that was never taken seriously in the mathematical world but somehow is significant now because ???

I was wondering where this goal post would shift to next. So now instead of AI being incapable of doing something not in its training data it needs too much external help doing what no human could do, and what it came up with isn't even important anyways.

The companies can suck and their product can do amazing things at the same time.

I think a lot of what Open AI has done feels insecure and looking for some sort of business validation.

I posted this in another comment but this was weird: https://openai.com/sam-and-jony/

Just more mismanagement by Sam.

It is interesting to see CEOs of big tech companies are so convinced they are always right that they start sidestepping any internal review of communication. Often to the detriment of the company.

My first thought was: did Sam write this?

they assume if someone wants context, they'll just ask chatGPT for it

“Nothing in my training data covers this, but if I had to guess, OpenAI may have released information about GPU support in Mac OS X early, pissing off Steve Jobs.”

sadly they are probably right

I find it refreshing that they are skipping the typical hedged corporate speak.

I think this is why companies are doing it. But interestingly it causes us to remember why we have corporate speak in the first place.

I think the pendulum will then swing back to the expectation that companies act and behave like adults rather than cringy kids trying to be cool.

Straight to unhinged speak is definitely refreshing /s