In August of 2019, reporter Karen Howe went to visit the San Francisco headquarters of Open AI.

It had started as a non-profit research lab that said it was dedicated to fundamental AI research

with no commercial focus, no commercial intent. At the time, Karen was covering AI for

MIT Tech Review. An OpenAI was the buzzy new kid on the block. It had money, big names, and lofty

ambitions to develop artificial general intelligence, AGI, for the good of society.

Then Open AI, which had really built a mission around the idea that it would be fully transparent and open source all of its research,

withheld its first piece of research for sort of unclear, ambiguous reasons. And then it restructured

as an organization. A few months earlier, the nonprofit that controlled OpenAI set up a for-profit

arm and got a billion dollars from Microsoft. Not long after, Karen walked in.

So the first meeting that I had in the offices was with Greg Brockman and Ilya

Satskeber, the CTO and chief scientist of Open AI. And I kind of went in, I went in

with a generally good view of the organization. And I wanted to just ask some

basic questions about how they thought about themselves, how they thought about their mission,

which was to ensure artificial general intelligence benefits all of humanity. So I began by

asking, you know, how do you define AGI? What do you think it will be

able to do? What are the bad things that you think could come out of it? Because another

pillar of opening eyes mythology is that they need to build good AGI before someone builds bad

AGI. And I also asked, why do you think we should spend billions of dollars on this problem?

You know, humanity is facing a lot of different challenges.

Why not spend billions directly on tackling climate change rather than on AGI that could

help tackle climate change? And I AGI that could help tackle climate change.

And I realized in that meeting that they didn't have a very good articulation at all of what

...