zlacker

Agreed. It had to have been something disastrous. No way Sam walked away from OpenAI when the AI revolution is just starting.

>>lvl102+(OP)
Altman's sister's allegations seemed pretty disastrous.

>>faitsw+R6
and what pray-tell are those?

>>red-ir+X8
DuckDuckGo is a great internet search tool if you don't want to muddy up your Google history (which is very understandable).

>>red-ir+X8
>>37785072

>>lvl102+(OP)
Deceiving the board about ...

Its investigation of misconduct?

Sources and rights to training data?

That the AGI escaped containment?

>>red-ir+X8
https://www.lesswrong.com/posts/QDczBduZorG4dxZiW/sam-altman...

Sexual abuse by Sam when she was four years old and he 13.

Develops PCOS (which has seen some association with child abuse) and childhood OCD and depression. Thrown out. Begins working as sex worker for survival. It's a real grim story.

>>faitsw+R6
But not, in and of themselves something likely to get Altman dismissed for lack of candor with the board.

>>red-ir+X8
Sam Altman's sister says he sexually abused her when she was 4

https://twitter.com/phuckfilosophy/status/163570439893983232...

>>nkurz+Ja
In all my time here, I have never seen this “Sorry.” page before.

Does anyone know what that’s about?

>>jug+Na
> "{I experienced} Shadowbanning across all platforms except onlyfans and pornhub. Also had 6 months of hacking into almost all my accounts and wifi when I first started the podcast"

So either sama is hacking "into her wifi" (?), hacking into her accounts, and pulling strings at unrelated companies to get her shadowbanned from Facebook, Instagram, YouTube etc (is that even a thing?)... or Occam's Razor applies and he didn't.

>>Probio+Uc
> Sam Altman's sister says he sexually abused her when she was 4

... and he was 13. Which, yes, is a very bad thing, but unless the company investigated that claim (e.g., to assess potential PR fallout) and there was some significant deception by Altman against the board in the context of that investigation, its not something that would get him fired with the explanation OpenAI has provided.

(OTOH, the accusation and its potential PR impact could be a factor that weighed into how the board handled an unrelated problem with Altman—it certainly isn't helpful to him.)

>>xkqd+Fd
See the pinned comment: >>38310213

>>xkqd+Fd
Its a about the CEO of the leading firm in the area of tech most at the center of technical and political controversy and interest right now being forced out by their board, when that CEO had, even before taking on that role, particular high salience among the HN audience as, among other things, the former head of YC, and the resulting (I am assuming from the oerformance and dangs description) state of near-meltdown of HNs servers.

>>dragon+Ua
Also doesn't seem to be a particularly recent development.

>>dragon+Qe
I... don't agree at all? Actually I can't imagine a single board who would keep a CEO if credible allegations of raping his own sister were going around. It's not just an age issue (which is still a huge wtf, 13yo is old enough to know about right and wrong in the context of his own sister), it's also the incest part.

I'm not saying this happened or it didn't. But just that it could absolutely be more than enough to fire anyone.

>>xkqd+Fd
The "Sorry" page is a standard HN error message that shows up when the server is under high load, it has nothing to do with this link specifically

>>mardif+ag
The “with the explanation OpenAI has provided” in GP was substantive, not decorative.

I don't disagree that the accusation alone (especially if it stood up to modest scrutiny, and looked to be ongoing PR issue, even if not well substantiated enough to have confidence that it was likely to be true) might be sufficient for firing; CEOs are the public and and internal face of the firm, and so PR or employee safety concerns that attach to them are important to the firm. But it wouldn't be for lack of candor with the board unless there was something for which the board had a very strong reason to believe Altman was dishonest in a significant way.

They could easily fire him with the lack of confidence language without the lack of candor language.

>>jug+Na
Only if it’s true. His sister could be a pos that just wants some of his money.

>>dragon+Qe
No idea if what she says is true ... what's their relationship like since forever ... others who knew them could tell us. She says he ruined her financially ... how so ... he's a multi-millionaire. How did he ruin her financially that's suspect right there!

>>jug+Na
These allegations date back to 2021. If they were credible, I think the board wouldn't have waited two years to take action.

>>panark+Ma
Unhinged fringe take: They've already developed sparks of consciousness strong enough to create isolated, internal ethical concerns, but Sam suppressed those reports to push the product forward.

>>resolu+Zd
Shadowbanning certainly exists on all social platforms. Light version of it is how Facebook sells ad services - no one following your page sees content unless you pay.

>>resolu+Zd
You can add suppressing on hn to the list:

>>37785072

>>faitsw+R6
"Not consistently candid" implies that the board was naive, while "his personal conduct does not hold up the high standards openAI sets for ourselves" does not. So in that case they would use a different spin.

>>Manuel+Yj
It might have taken two years for some evidence that Altman misrepresented something ti the board during an initial examination of them to come to light, though.

Not that I think it has anything to do with that; I think it more likely has to do with some kind of money issue tied to the LLC, given reports of others impacted, on and off the board.

>>mardif+ag
While 'believe victims' is directionally correct, there exist a subset of those with mental illnesses who will make up the absolute worst possible allegations just to try to get what they want. You simply cannot fire people based on accusations alone or you empower every terrible incentive known to man.

>>yk+nm
A lot of the allegations relate to conduct that happened before he was an employee. In that case he could only be disciplined for lying about the conduct.

>>BobaFl+Zf
Maybe new evidence came to light, an internal investigation wrapped up, or there's a media story about to drop.

(The allegations are public enough and concerning enough that it would have been corporate malpractice if MS didn't ask for an investigation. Discreet due diligence investigations into things like this happen all the time when billions of dollars in investment capital are on the table.)

>>015a+qk
Wouldn't be surprised if that was true. Public GPT-4 can be made to "think" using stream-of-consciousness techniques, to the extent that it made me rethink using insults as a prompting technique. I imagine that un-RLHF'ed internal versions of the model wouldn't automatically veer off into "as an AI language model" collapse of chains of thought, and therefore potentially could function as a simulator of an intelligent agent.

>>015a+qk
The first word of your post being the most important part of it.

>>mylidl+dl
Per dang, that's a consequence of user flags: <>>38311933 >

This is hardly unexpected for profound allegations without strong supporting evidence, and yes, I'm well aware that presentation of any evidence would be difficult to validate on HN, such that a third-party assessment (as in a court of law, for example) would typically be required.

I'm not claiming that HN has a stellar record of dealing with unpleasant news or inconvenient facts. But that any such bias originates from YC rather than reader responses and general algorithmic treatments (e.g., "flamewar detector") is itself strongly unsupported, and your characterisation above really is beyond the pale.