The cost of fragmentation

Well, this is not a hypothetical.

Late last week, three former OpenAI safety researchers published a letter addressed to OpenAI's board and its safety committees, and they say that the reason they were given for their firings were simply not adding up.

Of course, OpenAI sees it very differently.

The company says they were let go for mishandling sensitive information.

There was a significant breach of trust and that this wasn't about raising safety concerns.

Very much not here to referee that discussion.

Certainly don't have the facts.

But what interests me is what it shows about where governance is under strain.

And we've talked a little bit about this on some of the other episodes, since obviously AI is the hottest issue around for us and many others.

Yeah, I mean, it's really interesting.

I read both the some coverage of the event, the letter that the three former employees had sent to leadership and to the board.

And then I read OpenAI's tweet responding to all of that.

And like you said, definitely very hard to speculate.

So we won't, considering we don't really know everything involved.

But I think a lot of folks out there who read all of these materials are probably wondering, well, what was the significant breach of trust?

Definitely.

But even without those facts, I mean, this is not necessarily a standalone issue, if you will, related to concerns about safety.

And I think we touched upon this even in our last episode.

A couple of weeks ago, the anthropic researcher Jacob Cox had resigned and publicly warned about where AI is heading.

And Axios reported that several colleagues still inside the company very much endorsed his concerns.

Anthropic is also reportedly headed toward an IPO with a perspective that reportedly devotes around 80 pages to risk factors.

There's also been reporting in Axios as well that more than a thousand employees across OpenAI, Anthropic, Google signed a petition called Pacing the Frontier and described researchers as having unusual leverage over their employers.

So When you have people closest to this technology speaking up in public, on Twitter or X, in open letters and in resignations, I mean, if you're a board member, it comes back first here, I think, to your comment.

If there's dissent inside the company, does it reach the board through a proper channel or does the board find out the way everybody else does? And the second one, which is very much top of mind for me, is when a company tells its investors, The downside could actually be catastrophic, but it keeps going.

It keeps competing full speed anyway.

Well, this is not a hypothetical.

Late last week, three former OpenAI safety researchers published a letter addressed to OpenAI's board and its safety committees, and they say that the reason they were given for their firings were simply not adding up.

Of course, OpenAI sees it very differently.

The company says they were let go for mishandling sensitive information.

There was a significant breach of trust and that this wasn't about raising safety concerns.

Very much not here to referee that discussion.

Certainly don't have the facts.

But what interests me is what it shows about where governance is under strain.

And we've talked a little bit about this on some of the other episodes, since obviously AI is the hottest issue around for us and many others.

Yeah, I mean, it's really interesting.

I read both the some coverage of the event, the letter that the three former employees had sent to leadership and to the board.

And then I read OpenAI's tweet responding to all of that.

And like you said, definitely very hard to speculate.

So we won't, considering we don't really know everything involved.

But I think a lot of folks out there who read all of these materials are probably wondering, well, what was the significant breach of trust?

Definitely.

But even without those facts, I mean, this is not necessarily a standalone issue, if you will, related to concerns about safety.

And I think we touched upon this even in our last episode.

A couple of weeks ago, the anthropic researcher Jacob Cox had resigned and publicly warned about where AI is heading.

And Axios reported that several colleagues still inside the company very much endorsed his concerns.

Anthropic is also reportedly headed toward an IPO with a perspective that reportedly devotes around 80 pages to risk factors.

There's also been reporting in Axios as well that more than a thousand employees across OpenAI, Anthropic, Google signed a petition called Pacing the Frontier and described researchers as having unusual leverage over their employers.

So When you have people closest to this technology speaking up in public, on Twitter or X, in open letters and in resignations, I mean, if you're a board member, it comes back first here, I think, to your comment.

If there's dissent inside the company, does it reach the board through a proper channel or does the board find out the way everybody else does? And the second one, which is very much top of mind for me, is when a company tells its investors, The downside could actually be catastrophic, but it keeps going.

It keeps competing full speed anyway.