52 comments

[ 0.19 ms ] story [ 52.1 ms ] thread
As much as the headline leads to satirization, I think this is a good thing. I’d much rather have someone who’s leading the forefront of a technical innovation on the board than not. At least they can participate and inform the board what is being done and what isn’t being done. What can be changed and what we don’t yet have the tools to do.

A board that has Sam Altman on it is going to be much better able to craft safety guidance for the use and misuse of AI than a board that doesn’t have access to such a domain expert.

Just, you know, don’t give him veto power.

> A board that has Sam Altman on it is going to be much better able to craft safety guidance for the use and misuse of AI than a board that doesn’t have access to such a domain expert.

Incentives. You can't trust a man who makes billions on a this tech to make decisions in the best interest of the nation.

It's inevitable that commercial motives become a factor here. It's well known that OpenAI wants a "moat" and holding back competition with red tape for things they themselves have already managed to achieve before said red tape was introduced is a great way to make that happen.

Domain expert? Link please?

I’ve never heard Altman describe so much as a ReLU, I’ve never seen him publish code, I’ve never seen him do a company’s thats not at least two of: fired for shady shit and self dealing (YC, OpenAI), a massive loss to investors and/or acquirers (Loopt, Socialcam, countless), paid for by a massive network of people who just back who the other guy backed.

The guy is a dark-triad sociopath, this is well documented [0], and yeah, let’s get fucking Larry Summers in there to play the cool head.

Guy’s a super, super creepy scam artist who fails up because people say difficult to define stuff like “domain expert” to keep the myth going.

[0] https://www.newyorker.com/magazine/2016/10/10/sam-altmans-ma...

Exactly. Why would it be bad for Altman to be on the board? Aren’t these things more about understanding the dangers? Why would OpenAI benefit from hiding the dangers of this technology?

I mean, they’d just as much get screwed as human victims right?

OpenAI’s public advocacy centers around selling a particular, highly controversial image of what the salient dangers are and are not, which (purely coincidentally) aligns with the policy advocacy of erect barriers to new and emergent competition in favor of a narrow set of large incumbents.

OpenAI has a lot to gain by selling a particular view of the dangers of this technology, orthogonal to the accuracy of that view.

But he’s just one member. So I understand if we don’t want him to be an AI czar or something. That would be problematic.

But a member of a board of experts seems reasonable.

> But a member of a board of experts

What is he an expert in?

Conniving and duplicitous self promotion
I would wager that Altman is one of probably, say, 50 or 100 people on the planet that have an extremely intimate preview of the next couple of years of AI models (the other 50-100 people being his contemporaries in the various companies working on such things, whether they're senior engineers or "AI safety" people, or all the people that are working with this technology at its most raw and bleeding edge forms).

So I'd definitely count him as someone that has a good view of what's going on, and what's coming over the horizon.

Most of the board members are company leaders that also benefit from regulatory capture and limited competition. They probably will act identically or even just explicitly coordinate.
"You will find that one billion dollars will not deter the silicon the way it does the carbon."
Someone on the forefront of crypto like Sam Bankman Fried should pick the regulatory body for crypto. Ahahaha. It really is hilarious how some baseless principles get people to go to insane ends.
From the membership I can't interpret it as anything other than a pure industry council.

An AI safety board would be dominated by representatives of grassroots civil society organisations, maybe some labour unions. Almost all of these are CEOs of large for-profit corporations. Altman fits right in with the rest.

If you want a board that actually deals with AI safety, these are the precise wrong people to put on it.

I doubt it's a safety focused council. It seems rather like a council for economically useful applications. If there's any interest in safety it's in the limited sense of making AI robust enough for these applications.

(comment deleted)
Safe for who/what?

From the official post (linked elsewhere in discussion):

> DHS is responsible for the overall security and resilience of the nation’s critical infrastructure

Not people, apparently.

Yes, and it's not even bad, as a council. It's perfectly legitimate thing to have an 'industry council for robust and secure AI applications in critical infrastructure'. I'd probably have at least some people whoa aren't software firm CEOs-- probably some experts in managing electrical grids and the like, but you could have it like this.

The bad thing is really only calling it an AI safety board.

> An AI safety board would be dominated by representatives of grassroots civil society organisations, maybe some labour unions.

the “safety” in “AI Safety” isn’t what you think it is. it’s homeland security, not labour/societal security.

It could mean "safe space for corporations trying to monopolize technologies"
Yes, but that isn't what AI safety actually means.

AI safety is AI safety. I only really object to the name of the board. Otherwise it seems quite reasonable to have one, although if it's for what I think it is I think some grid people should also be on it.

(comment deleted)

  Ed Bastian, CEO, Delta Air Lines; 
  Vicki Hollub, President and CEO, Occidental Petroleum;
  Wes Moore, Governor of Maryland; 
Most of the other names make some lick of sense, but why are these people on an AI safety board?
Delta is the world's biggest airline and a part of America’s transportation infrastructure. The DHS probably has an interest in ensuring that airline infrastructure is reliable and secure, particularly from cyberattacks.

Fossil fuel infrastructure is critical for self explanatory reasons.

I’m not sure why Wes Moore or Bruce Harrell (Mayor of Seattle) are on there - I view them as people looking to use these boards as a path to future federal positions. And the DHS probably wants them to be able to claim they have representation from political leaders of various levels.

The public statement to Delta goes something like this...

""" Dear Ed. Bastian,

Just as Delta has decided to move away from Boeing for the majority of its fleet, so should it also steer clear of AI as it may relate to the rest of your business model.

Sincerely, The Internet """

(comment deleted)
Having political activists on this board is an embarrassment. For example the statement from one of the board members, Maya Wiley, President and CEO, The Leadership Conference on Civil and Human Rights, talks about “threats, including the spread of bias and hate speech online, stoking fear, distrust, and hate in our communities of color”.

These types of highly biased political organizations and leaders are rarely interested in actually upholding civil and human rights like free speech and instead want to impose their personal moral code on everyone else under labels like “ethics” or “safety”.

But I suppose that’s not as bizarre a pick as Li Fei Fei, who allegedly has various links to Chinese leadership and China in general.

And Altman is interested in upholding human rights? And would never push his own personal ethics on the world, right?
I am skeptical of them as well but mainly that they’ll use this as a way to get richer.
And the consequences of that are more acceptable to you?
Nothing surprises me in a country where a regulatory body (FAA) has it's client (Boeing) do it's own certification.

This reads the same way.

Altman is basically as close to the platonic ideal of “never give this psycho power” of anyone I can think of.

Dude makes Zuckerberg from the movie look flat fucking humanitarian.

Eyeball scanning poor countries.

Fuck Altman.

Im currently in one of those countries and I saw homeless people that don’t even know when if ever they will get access to a computer giving away their info for something they’ll never be able to cash in anywah
It’s dystopian, Minority Report dark.

Why this guy is becoming an unappointed regulator rather than sharing a bunk with SBF is like, a lagging indicator.

Don't you work at FB for Mark?

I hate them both.

You and I used to share an office in SF, btw. Your emacs Java hot-reload setup was cool. Cheers.

I left FB in 2018.

I try to be fair to Meta because I’m in the position to criticism them accurately, this raises the bar on my responsibility to acknowledge good decisions.

My memory is that of a guy my age and your username isn’t jumping out, what did we collaborate on?

Can you give some examples?

I only know about the worldcoin idea, his post apocalyptic bunker (not sure if true) and the fact ppl tried to oust him because he dismissed ethical AI considerations?

He’s a business man, wtf does he know about AI safety on a technical level?

Altman advice: Yeah bro, just instruct tune harder and it’ll be just fine, I believe in you.

Absolutely disgusting, no for-profit CEOs should be on boards such as these especially when it comes to something so publicly important.
> Altman handpicked for Homeland Security's AI safety board

No conflict of interest whatsoever. /s

So they put the people in the forefront of creating AI to steer AI Policy when it comes to safety? There are 4 Civil Rights people on the board. The rest are autocrats of the software industry.
AI should not be applied to "critical infrastructure".
This reminds me that the only place I dislike more than Silicon Valley is Washington.