[polyanos]

You know, this week I've been fascinated by a GPT-3 implementation writing HTML/React code, but, as you said, that is coming from a YC Alumni, so there might be actually some truth about the OpenAI press "conspiracy" theory.

Granted I was/am still sceptical, but also very curious, about the actual performance and quality of that code generator thingy and I hope he actually opens the tool this weekend this time around, as he didn't open it up at the previous announced date.

Oh, am talking about https://twitter.com/sharifshameem btw.

[gas9S9zw3P9c]

I've seen multiple profiles just like that one tweeting amazing things about GPT-3. I didn't keep track, but [0] and [1] just turned up doing a quick Twitter search. There are a lot more.

It's quite suspicious that instead of giving access to AI researchers who have the ability to evaluate the model and may be skeptical, OpenAI has largely been giving access to Silicon Valley startup people and VCs who know very little about AI but say how game-changing it is. Perhaps it's just their network with extra incentives. Gwern being the only exception that comes to mind.

[0] https://twitter.com/arram/status/1281258647566217216

[1] https://twitter.com/mckaywrigley/status/1284110063498522624