11 mo. ago • 95%

New report illuminates why OpenAI board said Altman “was not consistently candid”

arstechnica.com New report illuminates why OpenAI board said Altman “was not consistently candid”

Insider report details clash over one board member's criticism in an academic paper.

53 comments

So a Board member wrote a paper about focusing on safety above profit in AI development. Sam Altman did not take kindly to this concept and started pushing to fire her (to which end he may or may not have lied to other Board members to split them up). Sam gets fired for trying to fire someone for putting safety over profit. Everything exploded and now profit is firmly at the head of the table.

I like nothing about this version of events either.

337
- Wasn’t that evident from the very first few days, when we learned the board stood for the non profit, safety first mother org while the booted ceo stands for reckless monetization?
  
  Now he’s back, the safety concerns got silenced, money can be made, people can get fucked. A good day for capitalists
  
  106
  
  That's why I was so confused that all the workers stood behind the CEO and threatened to go to Microsoft.
  
  71
- I feel like this isn't surprising knowing about all the other stuff altman has done. Seems like yet another loss for the greater good in the name of profit.
  
  59
  
  What other stuff has he done? Genuinely curious.
  
  13
  
  Now what would the company do if the AI model started putting safety above profit (i.e. refusing to lie to profit the user (aka reducing market value))? How fucked are we if they create an AGI that puts profit above safety?
  
  13
- So basically it's exactly what I expected and I'm not surprised in the slightest. Amazing how that works.
  
  It's not too surprising considering they don't even have basic essential security features in 2023 like two-factor authentication. Absolutely pitiful.
  
  1
The thing that shits me about this is google appear to the public to be late to the party but the reality is they DID put safety before profit when it came to AI. The sheer amount of research and papers put out by them on AI should have proven to people they know what they're doing.

And then openAI throw caution into the wind and essentially make google and others panic knee jerk because there's really money to be made, and now everyone seems to be throwing caution into the wind and pushing it into the mainstream before society is ready.

All in the name of shareholders.

119
- 10k%! A friend works in brand marketing at Google. They'd been using internally for months before market pressure forced them to start onboarding public end users. I've been in the earliest of the external betas (bc I give a lot of product feedback over the years?) and from the beginning the user experiences have been the most locked down of all the consumer LLMs
  
  11
- I think it's not enough, disable all the safe guards and let people decide if the output is what they want, hate being treated like a child trying to buy a M rated game.
  
  -65
  
  But this isn't an M rated game, it's a transformative new technology with potentially horrifying consequences to misuse
  
  66
  
  And while you’re at it, remove safety on guns. And seatbelts. And might as well get rid of those pesky boom gates. I can hear the trains just fine, I don’t like being treated like a child. /s
  
  18
Yet another jackass CEO. Anybody surprised?

38
This is the best summary I could come up with:

Toner, who serves as director of strategy and foundational research grants at Georgetown University’s Center for Security and Emerging Technology, allegedly drew Altman's negative attention by co-writing a paper on different ways AI companies can "signal" their commitment to safety through "costly" words and actions.

In the paper, Toner contrasts OpenAI's public launch of ChatGPT last year with Anthropic's "deliberate deci[sion] not to productize its technology in order to avoid stoking the flames of AI hype."

She also wrote that, "by delaying the release of [Anthropic chatbot] Claude until another company put out a similarly capable product, Anthropic was showing its willingness to avoid exactly the kind of frantic corner-cutting that the release of ChatGPT appeared to spur."

At the same time, Duhigg's piece also gives some credence to the idea that the OpenAI board felt it needed to be able to hold Altman "accountable" in order to fulfill its mission to "make sure AI benefits all of humanity," as one unnamed source put it.

"It's hard to say if the board members were more terrified of sentient computers or of Altman going rogue," Duhigg writes.

The piece also offers a behind-the-scenes view into Microsoft's three-pronged response to the OpenAI drama and the ways the Redmond-based tech giant reportedly found the board's moves "mind-bogglingly stupid."

The original article contains 414 words, the summary contains 215 words. Saved 48%. I'm a bot and I'm open source!

17

You've viewed 53 comments.