“Not consistently candid”

Maybe OpenAI’s board had a point, after all

Remember how OpenAI’s board said Sam Altman wasn’t “consistently candid”? For a long time, nearly everyone brushed them off.

One wildly popular opinion on X was that the OpenAI drama was caused by a conflict of interest on the part of Adam D’Angelo, in popular tweets like these:

Kara Swisher’s narrative, meanwhile, was essentially that the board was a disaster, and that there was a palace coup, over a mere difference of opinion. The more the drama continued, the more Swisher dug in — eventually going so far as to block me on Twitter when I didn’t fully buy it.

The first time I pushed Swisher see how the board’s statement fit in, Swisher brushed me off, and took another whack at the board.

She continued to hammer on that the board for days:

For a little while she even had me going. But then on the Sunday when Sam walked and everyone reported negotiations would be done by 5pm, as the deadline approached, Emily Chang reported that the negotiations were held up because Sam wanted a statement “absolving him of wrongdoing”:

Absolution? Hmmm. Could be standard operating procedure, but the fact that it was a sticking point drew my attention.

Mike Solana’s theory, meanwhile, and I kid you not, was essentially that board member Helen Toner was a communist:

§

None of these theories— that D’Angelo was driven by conflict of interest, that the board was clod, or Solana’s ridiculous but popular communist theory—acknowledged, even a tiny bit, that the board might have a legitimate point.

The more that comes out, the less likely any of these three theories seems to be correct. (Altman himself debunked the Po theory, after the dust settled.)

There was of course, all along, a fourth theory, a really dopey one, that most people dismissed: that the board spoke truthfully when it said it that it was concerned about Altman’s candor

For a while, there wasn’t much to support it beyond this post on X, which initially felt to me like hearsay

Since then then, more facts have trickled in, and Irving’s post has resonated more and more.. To my knowledge, each and every new fact that has come in, since November 20, speaks further in favor of the board’s original claims:.

• Even after all the chaos was said and done, D’Angelo remained on the board. And as mentioned, after the dust was settled, Altman himself backed D’angelo up on X. Would that have happened if D’Angelo had no legitimate reason for voting to fire Sam?

• Altman gave up his board seat. If there was really nothing that even looked like legitimate beef against him, I doubt he would have conceded this.

• Altman signed off on his friend and ally and cofounder Greg Brockman giving up his board seat. Again, if there neither smoke nor fire, it is doubtful that he would conceded this.

• Altman agreed to an internal investigation. He could be innocent of the charges, whatever they were, if there was not at least the appearance of a conflict of interest, it’s not clear that he would not have agreed to this.

Two more scoops further this sense.

First, a few days ago Charles Duhigg just got huge scoop at The New Yorker (famous for its fact-checking, so presumably legit), the first direct insight beyond the board’s own initial statement into what they are thinking. It fits exactly with the dull, un-conspiratorial theory that the board said what it meant:

Wow.

Second, I just stumbled on something else that in hindisght feels, well, not completely candid.

It started when was struck by a report yesterday at WIRED that OpenAI had agreed to buy $51M worth of chips from a company that Altman had a stake in (Rain).

That reminded me of a key moment where we in the Senate together under oath. Senator Kennedy asked him about his stake in OpenAI.

After poking around, I found out that “I have no equity in OpenAI” was only half the truth; while Altman to my knowledge holds no direct equity in OpenAI, he does have an indirect stake in OpenAI, and that fact should have been disclosed.

In particular, he own a stake of Y Combinator, and Y Combinator owns a stake in OpenAI. It may well be worth tens of millions of dollars; even for Altman, that’s not trivial. Since he was President of Y Combinator, and CEO of OpenAI; he surely was aware of this¹

#notconsistentlycandid.²

§

Geoffrey Miller just said it pretty well on X:

The fact that 750 employees who were about to get to sell stock at a $86 billion valuation felt differently doesn‘t change that, not one bit.

§

Most of you by now realize that OpenAI was set up to be a nonprofit. Fewer realize that it still is a nonprofit, receiving tax benefits at least as recently as couple years ago.

Even fewer have probably read the original papers. They are striking in current context.

OpenAI swore to the State of California that their goal was to build safe AI technology and to sure that AI’s benefits would be “widely and evenly distribution as possible” and that their goal was “to benefit humanity as a whole”, and that they would “openly share … plans and capabilities along the way”.

By any reasonable measure they have strayed, very far, from that mission. That’s primarily on Sam.

If the Board was making a step towards bringing them back to their original mission, of AI for human interest rather than for profit, we should be praising them, not burying them.

Gary Marcus spent three decades as an academic, but has also worked in three startups, founding and CEO’ing two, one of which was acquired by Uber.

1. OpenAI characterizes Altman’s indirect stake as a “small investment”, but OpenAI and Altman should nonetheless disclose the value of that stake, My guess is that it is at least worth many tens of millions, at the new $86B valuation, possibly more, more than many people will earn in a lifetime.

2. If he was already working on the chip deal in May at the Senate, that should have been disclosed, too.