Ben Bernanke Joins Anthropic Oversight Trust(anthropic.com)
anthropic.com
Ben Bernanke Joins Anthropic Oversight Trust
https://www.anthropic.com/news/ben-bernanke
90 comments
Imagine Rupert Murdoch (95, worth $21.7 billion) arguing the pros and cons of job displacement and widening economic inequality caused by generative AI.
Big fan of Larry Summers here. He seems to be part of everything, Harvard, Epstein, etc.
so the architect of government bailout gets a cushy gig. probably one of the most harmful precedents set and now companies expect bailouts. to bailout the company instead of people and small shareholders was always poor decision, emboldened the worst of the business class. don’t love this hire lol
So does this appointment signpost a likely industry-wide token squeeze requiring Benanke’s specific domain knowledge to navigate?
He's 72 years old, I'm sure he has the best brains and everyone's best interests in mind. This is exactly what I want to see. I'm sure he will have very good opinions on technology and it's implications.
I had conversations with Ben a few years ago about his economics research (incidentally related to some research I was doing at the time). I wouldn't have guessed he was close to 70. He's a sharp guy.
(I also think his economics research is excellent, and the comments elsewhere about his track record aren't particularly fair.)
(I also think his economics research is excellent, and the comments elsewhere about his track record aren't particularly fair.)
Hires like these are made for their brand rather than expertise. His job will be to attend galas, schmooze with investors and give legitimacy to the operation. The company will hire others to do the real work.
> The company will hire others to do the real work.
For 1/10th the pay, 10x the effort, and 0% of the credit!
Reminds me of Walter Isaacson, the Musk biographer. Supposedly he was hired by an investment bank after he made his fortune, and is a member of the American Philosophical Society [1]. Truly a one of a kind person: a biographer of famous people, an expert on financial markets, and a leading philosopher all at the same time!
[1] https://en.wikipedia.org/wiki/Walter_Isaacson
For 1/10th the pay, 10x the effort, and 0% of the credit!
Reminds me of Walter Isaacson, the Musk biographer. Supposedly he was hired by an investment bank after he made his fortune, and is a member of the American Philosophical Society [1]. Truly a one of a kind person: a biographer of famous people, an expert on financial markets, and a leading philosopher all at the same time!
[1] https://en.wikipedia.org/wiki/Walter_Isaacson
Bjarne Stroustrup is 75 years old. I don't think that age per se tells us much.
What's Stroustrup up to these days though?
The C++ documentary released last month was great and featured him a lot: https://www.youtube.com/watch?v=lI7tMxzSJ7w
He's just old enough to be part of congress. Give him another decade and he'll be ready to run for president.
This sounds like age discrimination.
Yeah, this seems like reputation renting similar to Larry Summers being put on the board of OpenAI, which turned out, um...very well I guess. For different reasons of course.
I guess it's better than him joining OpenAI, where I could see Altman being more "creative" with the books than Amodei. But I just don't Bernanke pushing back or blowing any whistles when things start smelling funny.
I guess it's better than him joining OpenAI, where I could see Altman being more "creative" with the books than Amodei. But I just don't Bernanke pushing back or blowing any whistles when things start smelling funny.
I assume this is sarcasm?
Good instinct — this is exactly the kind of issue we should keep an eye on.
Fair callout.
Ben Bernanke would have figured that out already.
That's how I read it. Do you have trouble detecting sarcasm?
This feels like Theranos loading up their board with big names.
Except Anthropic has delivered a truly world changing product…
So did Los Alamos?
Edit: don't get me wrong, I'm a happy user. But I'd also be a happy consumer of refined sugar in the early 20th century. I'm still not sure if these tools won't destabilize society to the point of collapse. I don't think we understand the complexity of what's going on nearly enough, and am certainly not optimistic about AI being net good for us
Edit: don't get me wrong, I'm a happy user. But I'd also be a happy consumer of refined sugar in the early 20th century. I'm still not sure if these tools won't destabilize society to the point of collapse. I don't think we understand the complexity of what's going on nearly enough, and am certainly not optimistic about AI being net good for us
Pretending there is any equivalence between the atomic bomb and LLMs is playing right into the hands of the companies who want to pretend it’s so dangerous that (a) only they can have it (b) the government should give them money to “win” a “race” against China. Playing that game is the only part that has any atomic bomb level geopolitical consequences (because those involved have real bombs).
The analogy to sugar or cigarettes I get, ten years from now our brains will have atrophied so much it will probably show on a MRI. But the acute danger is BS and just some people looking for power plus hangers on that want to imagine they’re in a sci-fi novel.
The analogy to sugar or cigarettes I get, ten years from now our brains will have atrophied so much it will probably show on a MRI. But the acute danger is BS and just some people looking for power plus hangers on that want to imagine they’re in a sci-fi novel.
This came up on the Lunch Money livestream yesterday. The entire episode is worth a listen but here's the relevant sections:
Krugman: "I've been writing some about downsides of technological change and I realized afterwards that if I really wanted a really stellar example of a productive important innovation that had terrible effects on society would be the cotton gin."
https://www.youtube.com/watch?v=uJxQbfbpf7M&t=25m38s
Richardson: "I always have a hard time articulating this, but the number of large plantations in which enslavers owned in air quotes, you know, more than 25 or more than 50 other human beings was a very very small proportion of the American South, less than 1%. The majority of people who again owned their black neighbors had one or two enslaved people on their farms. They weren't necessarily called plantations. And they would be working alongside those black Americans. And the cotton gin could have made small farms viable and could have ended human enslavement. And instead what they gave us was, you know, the the Trail of Tears in the 1830s that cleans indigenous Americans out of the southwestern land. You get an extraordinary land rush into the American South in the 1830s and the 1840s. And you get the establishment of these gigantic essentially factory farms. And that's a place where, you know, the majority of southerners, obviously the indigenous southerners and the black southerners, wanted no part of this system, but it actually didn't serve the white farmers either. It served a really small, less than 1% group of American enslavers in the American South. And you look at that and you think that wasn't the technologies fault. That was the fault of the people who um who set up the political system that enabled it to work that way."
https://www.youtube.com/watch?v=uJxQbfbpf7M&t=28m18s
Krugman: "I've been writing some about downsides of technological change and I realized afterwards that if I really wanted a really stellar example of a productive important innovation that had terrible effects on society would be the cotton gin."
https://www.youtube.com/watch?v=uJxQbfbpf7M&t=25m38s
Richardson: "I always have a hard time articulating this, but the number of large plantations in which enslavers owned in air quotes, you know, more than 25 or more than 50 other human beings was a very very small proportion of the American South, less than 1%. The majority of people who again owned their black neighbors had one or two enslaved people on their farms. They weren't necessarily called plantations. And they would be working alongside those black Americans. And the cotton gin could have made small farms viable and could have ended human enslavement. And instead what they gave us was, you know, the the Trail of Tears in the 1830s that cleans indigenous Americans out of the southwestern land. You get an extraordinary land rush into the American South in the 1830s and the 1840s. And you get the establishment of these gigantic essentially factory farms. And that's a place where, you know, the majority of southerners, obviously the indigenous southerners and the black southerners, wanted no part of this system, but it actually didn't serve the white farmers either. It served a really small, less than 1% group of American enslavers in the American South. And you look at that and you think that wasn't the technologies fault. That was the fault of the people who um who set up the political system that enabled it to work that way."
https://www.youtube.com/watch?v=uJxQbfbpf7M&t=28m18s
World changing in a good way?
Think in terms of salaries, health care expenses, etc, that was saved in the noble pursuit of freeing capital to optimize capital increase.
Yes, absolutely in a good way
The benefits of modernity (electricity, cars, iPhone, Claude) are good, but they come bundled with potentially terminal ecological costs which is bad.
Have people (smarter than me) come up with a good equation - or at least heuristic - to determine what inventions are morally good?
I suppose it'd be from a utilitarian perspective?
Ex: My gut feeling is nitrogen fixing would rank "low" on "terminal ecological impact" against "positive benefits to humanity"; the Vinyl resurgence would be around the middle; private jets for the Epstein class would the highest etc.
I suppose it'd be from a utilitarian perspective?
Ex: My gut feeling is nitrogen fixing would rank "low" on "terminal ecological impact" against "positive benefits to humanity"; the Vinyl resurgence would be around the middle; private jets for the Epstein class would the highest etc.
I was being a bit cheeky, but I’m not really arguing that individual inventions can be determined as good or bad. My point is it comes from the same underlying mode of production. "Claude is useful" and "the way we have organised society that led to its creation may be ecologically catastrophic” can both be true.
It doesn't really matter if you can justify the technological developments because they are inevitable. What matters is what society does with those advancements. So a utilitarian perspective might work, but we have to apply it to the social response to technology, not the development of technology.
The turbojet or maybe some material science that made the turbojet feasible is a technological advancement. It can provide benefits to socity in the form of clean(er than coal) power or affordable transportation. Using that technology to build private jets while maintaining a social class that uses them for harmful purposes is a social problem, not a technological one.
It's a similar argument for the machine gun or nuclear weapons. Arguably if you make a terrifying enough weapon people will refuse to fight wars. So from that perspective inventing the Maxim gun (technological) was moral but using it to fight WWI (social) wasn't.
The turbojet or maybe some material science that made the turbojet feasible is a technological advancement. It can provide benefits to socity in the form of clean(er than coal) power or affordable transportation. Using that technology to build private jets while maintaining a social class that uses them for harmful purposes is a social problem, not a technological one.
It's a similar argument for the machine gun or nuclear weapons. Arguably if you make a terrifying enough weapon people will refuse to fight wars. So from that perspective inventing the Maxim gun (technological) was moral but using it to fight WWI (social) wasn't.
A friend of mine that works in habitat restoration had an interesting take. Technological developments in logging outpaced conservation efforts and by the time we figured out what was happening we had done irreversible damage to ecosystems. The trees that were clearcut to make neighborhoods like Ballard in Seattle had truly enormous old growth trees that would take centuries to replace. Like so big you could live in the stumps.
Here's [1] an example from farther north but this was apparently common in what is modern day Seattle as well.
For a more modern example napkin math says you can go 1.2 miles in a modern Prius for 1 hour of H100 time.
The Prius gets 57mpg and gallon of gasoline is 33.7kWh for 57/33.7=0.597kWh per mile. An nVidia H100 consumes 0.7 kW. 0.7/0.597=1.18 miles in a modern Prius for an hour of H100 time. I'm not sure how much H100 time is consumed per token or kilotoken or whatever so I can't comment on the scale of ecological damage caused by AI and data centers in general. I suspect it's at least several orders of magnitude off from cars but it's definitely not nothing. This also ignores a ton of other factors like water use and vehicle emissions.
[1]: https://mohai.org/collections-and-research/search/item/1983....
Here's [1] an example from farther north but this was apparently common in what is modern day Seattle as well.
For a more modern example napkin math says you can go 1.2 miles in a modern Prius for 1 hour of H100 time.
The Prius gets 57mpg and gallon of gasoline is 33.7kWh for 57/33.7=0.597kWh per mile. An nVidia H100 consumes 0.7 kW. 0.7/0.597=1.18 miles in a modern Prius for an hour of H100 time. I'm not sure how much H100 time is consumed per token or kilotoken or whatever so I can't comment on the scale of ecological damage caused by AI and data centers in general. I suspect it's at least several orders of magnitude off from cars but it's definitely not nothing. This also ignores a ton of other factors like water use and vehicle emissions.
[1]: https://mohai.org/collections-and-research/search/item/1983....
There are certainly ecological costs, but in the long run, Earth's life will only survive if an advanced species like ours is able to transport it off the planet before the sun expands and boils away the ocean and atmosphere, in approximately 800 million years.
We need rockets for that not vibecoding.
Fortunately, the rockets for that will be helped along with the GPU capacity to run rocket simulations on. GPUs not being used to run LLMs can be used instead for physics simulations to help make those rockets work.
I'm looking at the second and third order effects of technologies. LLMs massively increase the surplus capacity of human civilization, and it is this surplus capacity enables resources (including human capital) to be expended on developing frontier technologies like rockets that can accelerate the development of space-faring capabilities.
That remains to be seen.
I for one am doubtful that AI as a whole has meaningfully improved the lives of just about anyone besides the few who have gotten rich. Meanwhile many have already lost their jobs as a result, even if AI is just a convenient scapegoat.
I for one am doubtful that AI as a whole has meaningfully improved the lives of just about anyone besides the few who have gotten rich. Meanwhile many have already lost their jobs as a result, even if AI is just a convenient scapegoat.
I disagree. People get dumber because they outsource thinking to fancy autocomplete and I haven't really yet seen work by LLMs that significantly improve the world.
You really think your life is better than 2 years ago because of AI chat bots?
If AI went away tomorrow idk if my life would meaningfully change
If AI went away tomorrow idk if my life would meaningfully change
An AI chatbot diagnosed my rolled shoulders. I had assumed I had bad workout form. I would google "discomfort in upper back" or whatever and it would say "back pain after exercise means take a rest" and I would rest and it wouldn't get better until I lost interest in working out. It never occurred to me that this could be a medical problem, so I never asked a doctor (not that I had one to ask, as I was working retail at the time). Last time I started working out, I went to ChatGPT instead of google and it said "here's your problem, here's how to fix it" and it was right and I continue to work out and my quality of life has risen tremendously. YMMV
I would see less fake news, fake profiles and fake content in general. I would be happier, even if I would miss some of it.
It’s kind of sad how much we accept the idea of ”trust absolutely nothing” nowadays, even movie trailers for fictional movies are made to drive clicks for ads… obviously there has always been a large trust issue online, but with gen ai we entered a new era of it.
Your question implies a belief that things are 'good' or 'bad', but the reality of the world is a lot more nuanced than that. Pretty much everything that doesn't lead directly to human suffering can be seen as both good and bad.
No, that would be OpenAI (or Google if you want to talk technicals). Anthropic's strategy was just let RL on coding and jack up the price. I can only assume their real strategy is to speedrun getting the whole industry turned into a utility.
8 months ago I asked this question, I will ask again:
Where are your browsers? Where are your compilers? Where are your databases? Where are your operating systems?
Can you point me to literally anything useful that works and was created by this world changing technology? All I see is dead project after dead project.
Where are your browsers? Where are your compilers? Where are your databases? Where are your operating systems?
Can you point me to literally anything useful that works and was created by this world changing technology? All I see is dead project after dead project.
If you take them at their word, the people working on the current browsers, compilers and databases are all using it
Google self reports 70% ai usage in code, bun was fully ai rewritten to be rust
Google self reports 70% ai usage in code, bun was fully ai rewritten to be rust
Let’s not forget we are talking about a world changing technology here. So when you tell me “some people working on some projects are using it”, I’ll pretend you didn’t say “all” because that’s untrue, you haven’t asked all of them, and the company that got bought out by Anthropic did their “rewrite-in-rust” meme, do you think it’s unreasonable to incredulous?
Now the goalpost is “all” of code?
I don’t think anything I say is likely to change your mind. You might need to get through this valley of depression by yourself in a few years and join the rest of devs using ai
I don’t think anything I say is likely to change your mind. You might need to get through this valley of depression by yourself in a few years and join the rest of devs using ai
Your reading comprehension is awful.
You said “all” as in “all” developers on those projects. You haven’t spoken to all of them to make that claim. You’re just making stuff up. That’s what I was pointing out.
If you are suggesting that we can call the Chromium project a work of LLMs because some developers may or may not be using LLMs on the project it let’s just stop the conversation right here.
This is becoming a classic case of LLM brain rot. I hope you come back to your senses.
You said “all” as in “all” developers on those projects. You haven’t spoken to all of them to make that claim. You’re just making stuff up. That’s what I was pointing out.
If you are suggesting that we can call the Chromium project a work of LLMs because some developers may or may not be using LLMs on the project it let’s just stop the conversation right here.
This is becoming a classic case of LLM brain rot. I hope you come back to your senses.
Why do many AI advocates sound so, so much like the cryptocurrency zealots of the mid-2010's?
Is it something you aim for, or is it just the natural way of communication for people excited by a new technology, patricianly one they feel gives them some kind of insider advantage or hidden knowledge?
Is it something you aim for, or is it just the natural way of communication for people excited by a new technology, patricianly one they feel gives them some kind of insider advantage or hidden knowledge?
How long did it take from the first DBMS to get to Postgres? The first OS to get to Linux? The first compiler to get to LLVM? For Postgres and Linux and LLVM to become mature enough to hold the revered reputation they have now?
The jury's still out on AI, but coding agents have only really worked for about 6 months now. It's not exactly a fair statement to make. Obviously good things take time and thought. And understanding the full implications of technological advancement also takes time and thought.
The jury's still out on AI, but coding agents have only really worked for about 6 months now. It's not exactly a fair statement to make. Obviously good things take time and thought. And understanding the full implications of technological advancement also takes time and thought.
AI isn't re-creating the world, it's augmenting it.
The death of Stack Overflow being a visible indiction: https://data.stackexchange.com/stackoverflow/query/1926661#g...
The death of Stack Overflow being a visible indiction: https://data.stackexchange.com/stackoverflow/query/1926661#g...
You can complain that the economics don't work out for you, but Theranos was a fraud, meaning they didn't have a product. Fable is very much a real thing that I can interact with over the Internet.
Has it ever been proven that the board knew that Theranos was fraudulent? I’m speaking broadly - of course they probably had their own suspicions.
Wouldn't that be exactly the board's fiduciary responsibility? To know?
The government response to the ‘08 crisis seems to have worked out better for most big banks (low taxing of negative externalities, growing larger and more profitable), than for regional banks (consolidated) and the bulk of Americans (low median wealth, rising costs of housing/living)
Given the data on this[1], this is a confusing choice of hire to ensure AI gains are distributed equitably
[1] https://economicprinciples.org/Why-and-How-Capitalism-Needs-...
Given the data on this[1], this is a confusing choice of hire to ensure AI gains are distributed equitably
[1] https://economicprinciples.org/Why-and-How-Capitalism-Needs-...
What gives you the idea that they want to distribute the gains from AI equitably?
The having so many tiny regional banks is a holdover from the bad old days when branch banking was largely outlawed.
Someone here recently said, “Dishonesty is a core value of Anthropic,” and that aligns with my experience of the company as a user. All their talk about AI safety since the company’s inception now feels like pure theater, given their conduct in everyday operations. It’s a shame how quickly their image has deteriorated.
That's actually exactly how I feel about Anthropic.
They play such a PR game, trying really really hard to be seen as the good guys. It feels as another satirical episode of Silicon Valley. It's very clear they are all money and power motivated while also pretending to do all of this for the good of humanity. I have rarely seen that level of hypocrisy and cultish behavior from leadership and employees there.
I would honestly just prefer if they were honest about being power and money hungry instead of playing that game of AI Safety.
They play such a PR game, trying really really hard to be seen as the good guys. It feels as another satirical episode of Silicon Valley. It's very clear they are all money and power motivated while also pretending to do all of this for the good of humanity. I have rarely seen that level of hypocrisy and cultish behavior from leadership and employees there.
I would honestly just prefer if they were honest about being power and money hungry instead of playing that game of AI Safety.
Palantir is a good guide. Their ad copy is frightening. "we will create the ultimate killers for total domination to make your enemies SUFFER"
The funny thing is it’s so transparent. Like… is that the point? They want us to know how dodgy they are, kind of as a “** you”?
Often the point of propaganda is not to convince, it’s to demoralise.
On the other hand I could also believe that they live in such a bubble they genuinely don’t understand how it comes across. Add in a non-negligible amount of neurodivergence and maybe that’s the simplest explanation
Often the point of propaganda is not to convince, it’s to demoralise.
On the other hand I could also believe that they live in such a bubble they genuinely don’t understand how it comes across. Add in a non-negligible amount of neurodivergence and maybe that’s the simplest explanation
> On the other hand I could also believe that they live in such a bubble they genuinely don’t understand how it comes across
Yes, at least until ~6 months ago that was my reading too. I felt they were part of the EA/less wrong crowd. Earnest, convinced they were smarter than everyone else, paternalistic, massively lacking in real wisdom.
Now, I think maybe they're still struggling with that but they've had a real taste of power and like everyone smart but lacking in wisdom, faced with the real world, all their idealism has become lip service and pandering to their previous in group while their real target (which they probably haven't even admitted to themselves) slips further and further towards gathering more power.
As another example of a tech company where something similar happened, see Google. Although at least they never started out with this condescending "we're here to save you from... us" vibe.
Yes, at least until ~6 months ago that was my reading too. I felt they were part of the EA/less wrong crowd. Earnest, convinced they were smarter than everyone else, paternalistic, massively lacking in real wisdom.
Now, I think maybe they're still struggling with that but they've had a real taste of power and like everyone smart but lacking in wisdom, faced with the real world, all their idealism has become lip service and pandering to their previous in group while their real target (which they probably haven't even admitted to themselves) slips further and further towards gathering more power.
As another example of a tech company where something similar happened, see Google. Although at least they never started out with this condescending "we're here to save you from... us" vibe.
Yes I fear you’re right, power is addictive to most people, once they have a taste they want more.
Likewise loss of power is traumatic, picture the barely conscious 90 year old senator getting wheeled into congress to vote or something or other. Few ever give up power willingly even if they’re just a small node in a bigger system and not even really in control
Likewise loss of power is traumatic, picture the barely conscious 90 year old senator getting wheeled into congress to vote or something or other. Few ever give up power willingly even if they’re just a small node in a bigger system and not even really in control
> On the other hand I could also believe that they live in such a bubble they genuinely don’t understand how it comes across. Add in a non-negligible amount of neurodivergence and maybe that’s the simplest explanation
I mean... if you put a bunch of anti-social people in charge of a dehumanizing technology what do you really expect as the outcome?
I mean... if you put a bunch of anti-social people in charge of a dehumanizing technology what do you really expect as the outcome?
Oh come on, like any other AI company is better? They all remind me of a comical villain - Anthropic the least.
At least Sam Altman appeared to drop all pretenses of his pure sociopathy some time ago
Hm, they are still the most transparent lab when it comes to publishing system cards and safety research. For example the system card for Fable 5 runs 319 pages.
The stuff with Fable falling back to Opus was a bad business move but seems consistent with their position on safety and was published in the system card. Is Ben Bernanke joining the board a dishonest move?
The stuff with Fable falling back to Opus was a bad business move but seems consistent with their position on safety and was published in the system card. Is Ben Bernanke joining the board a dishonest move?
Dishonesty is at the core of Effective Altruism, which strangles a lot of the sensible choices Anthropic should be making. Although this feels more like, "anyone with socio-political edge worming their way in to suckle on the feed of imaginary printed money" more than anything.
Can you provide some context for this? I’m aware of SBF’s EA links, and how empty those sentiments appear to have been, but I’m just some guy, and it isn’t clear to me that the whole idea is dishonest, even if I don’t think it’s terribly realistic.
Any idea how one can join a board? Eating donuts and bagels with coffee once a quarter for 200k is my dream job.
Already be rich. Sorry buddy, as George Carlin said it's a big club, and you ain't in it.
Is he for loosening or tightening AI safety policy?
If anything he would be for tightening it, but I suspect his role is less about being a vote one way or the other.
The value he brings is in his data, knowledge & analyses - which he surely has from the Fed - on the scope and extent of AI's potential rrisks in capital sustainability, market stability and wage/job displacement
The value he brings is in his data, knowledge & analyses - which he surely has from the Fed - on the scope and extent of AI's potential rrisks in capital sustainability, market stability and wage/job displacement
Does the Ben Bernanke have a lot of policy experience?
These guys have to produce a hit piece everyday...everyone by now knows that "we are doing this for humanity" is bullshit.
> everyone by now knows that "we are doing this for humanity" is bullshit.
There are people here in this thread that act like LLMs are the best invention since sliced bread
There are people here in this thread that act like LLMs are the best invention since sliced bread
Whoop de doo. I'm sure there'll be huge earth-shaking changes in their activities now, right?
Prepairing another bailout, I see
Illuminati
> Genie Energy's Strategic advisory board is composed of: Dick Cheney since 2009 (former vice president of the United States),[3] Rupert Murdoch (media mogul and chairman of News Corp), James Woolsey (former CIA director), Larry Summers (former head of the US Treasury), Michael Steinhardt, Jacob Rothschild,[4][5] and Mary Landrieu, former United States Senator from Louisiana.