Hacker News Reader

It's possible that it's only one strong personality and some money away but my guess is that OpenAI-rosoft have the best stack for doing inference "seriously" at big, big, scale e.g. moving away from hacky research python code and so on.

View on HN

Its pretty hard to ignore Google in any discussion on big scale

Completely right. Was basically only thinking about OpenAI versus Anthropic. Oops

Google in their corporate structure, is to cautious to be a serious competitor.

I'm not so sure about that. They have kind of opposite incentives to OpenAI. OpenAI starting without much money had to hype the AGI next year stuff to get billions given to them. Google on the other hand is in such a dominant position with most of the search market, much of the ad market, ownership of Deepmind, huge amounts of data and money and so on probably don't want to be seen as a potential monopoly to be broken up.

Also Sergey Brin is back in there working on AI.

They seem to have managed to do so just fine :)

Reply on news.ycombinator.com