Have you been reading Matt Levine's bit on Sam Altman? I love it. As usual, he doesn't explicitly express disapprobation; on the contrary, he couches his remarks in a tone of admiration. I hope you will forgive me for quoting the Altman passage from his September 30th newsletter in its entirety:
"The place you want to reach in your career is where you work for a company and you are like “you know what, I am just so rich, I don’t want you to pay me anymore, it’s fine, I’ll work for free,” and your bosses are like “nope, sorry, we insist, we cannot allow you to work here for less than $10 billion.” And then you’re like “ohhhhhhhh fine, fine, fine, I do love working here, and if there’s really no other way, I guess I will take the $10 billion.” Nothing remotely like this has ever happened to me in my life but here’s Sam Altman:
'On Thursday, Altman told some employees that there were “good reasons” he shouldn’t take equity, though he didn’t elaborate. And he said investors were pushing for an equity grant to align his financial interests with those of OpenAI, said someone who heard the comments.
Altman also said a Wednesday news report that he might get a 7% stake in the new OpenAI was “ludicrous.”'
We talked about this last week: There have been reports that, as OpenAI becomes a for-profit company, it might give Altman (its co-founder and chief executive officer, who currently owns no equity) a 7% stake worth about $10 billion. I surmised that this was not something he wanted, but something the investors wanted: “He is the founder and CEO of a hot startup, and the founder and CEO of a hot startup is supposed to own equity. Not just for his sake — not just so that he can be rich — but to align incentives.” And here is Altman saying that: He doesn’t want the $10 billion, but the investors are insisting.
Nobody in history has ever been better at, like, business negging than Sam Altman. He got OpenAI to a $150 billion valuation in part by going around saying “oh no, nobody should allow us to build our product, we’re going to destroy humanity,” and now he is allegedly going to get handed a $10 billion stake in OpenAI because he’s going around saying “oh no, nobody should give me equity, that’s ludicrous.”"
Yes, it follows up with: instead it makes art and writes creatively, but very poorly/cheaply and I still have to do the chores. It is solving the problem that it should not solve and not doing a good job.
Ideas about geo-engineering are frightening, mostly because they have to act at such a scale that the risk is enormous. People who plan to remove 5% sunlight form reaching the earth by seeding space with reflecting confetti scare me to no end. What kind of disaster are we going to create if we suddenly remove 5% of the basic driver of everything on earth? Are you crazy?
But pushing things true without knowing what we're doing *is* kind of a human trademark. I think it would be fitting to no longer label us an 'intelligent species' until we actually start to behave intelligently (wars, fouling our ecosystem, the list is endless).
You’re on the Titanic. The engineers are shouting: “The bulkheads are too low! The rudder is too small! There aren’t enough life boats!”. The sailors mumble: “It has been cold, there will be many more icebergs than usual and further south”. The owners are pressing the captain: “You should be in New York in six days, we desperately need a record!”. And the captain thinks: “I have execution power. I can break through. I will be successful.” and orders: “Northerly course and full steam ahead!”.
Geoengineering is also stupendously silly or, more cynically, a money-making scam. It takes little account of the sheer size of our planet and the scale of pollution we have inflicted on it. In a way geoengineering is a tribute to its proponents' wilful ignorance of atmospheric chemistry and of the earth sciences in general. Their arrogance is breath-taking. We are currently injecting 50 gigatonnes of carbon gases—that's 50,000 million tonnes— into the skies each yea. Since the start of the Industrial Revolution we have released approximately 1.77 million, million tonnes. In turn, the Earth's atmosphere contains roughly 5x10¹⁵tonnes of gases. That's 5,000 million, million tonnes of gas. Step outside and look upwards. The atmosphere reaches to a height of at least 700km (438 miles-ish). All of it is gas, and it extends over the entire planet's surface (about 500 million square kilometres). How on earth do these surface-bound idiots think their tiny schemes could possibly work? Whatever they release (or sequester) will be so insignificant that it will have almost zero impact on climate change. Even their balloons need hot air generated by, guess what? By burning fossil fuels (or by using helium—an incredibly scarce resource needed for vital medical equipment). In 'Termination Shock' the plan is to mimic the effect volcanic eruptions have had in shielding the planet from solar radiation. Unfortunately for these dimbos, even a mega-eruption like Mount Pinatubo in 1991 'only' released just under twenty million tonnes of sulphur dioxide. It did have a mild effect on the climate for a few years but that was all. These geoengineers have no sensible explanation about how they are going to generate a even similar amounts of SO2 and get it to the upper atmosphere without, in that process, making the greenhouse effect even worse.
Like I said, it's simply a testament to the ignorance and arrogance of these people, nothing more. We can solve the problem, but only by everyone taking responsibility with collective action to reduce emissions. Their approach is very similar to the scientific absurdity and selfishness of longtermism.
I'll just note that the NY Times ran another article on geoengineering the day after the one you mentioned that lumps concerns about it with the anti-vaccine and conspiracy cranks -- https://www.nytimes.com/2024/09/26/climate/geoengineering-conspiracy-theorists-skeptics.html. That's certainly accurate to a degree, given M.T. Greene's recent tweet accusing the Dems of controlling the weather and causing the Helene damage, but it isn't leaving much room for those of us who worry about the tech bro types who miss the point of science fiction and can get their hands on some money.
As always, I appreciate your take on these things. Off to read your piece in the Atlantic.
"Science fiction" was a brilliant marketing misnomer for what it actually is. The vast majority of it save for a few exemptions, like Greg Egan, is just fantasy with science-talk in place of magic spells. I've honestly had a lot of trouble enjoying the entire genre outside of Greg Egan, Ada Palmer, and Kim Stanley Robinson out of just awareness of how sordid the history is.
Thanks for these quick hits. All very helpful. Altman's assertion that AGI would solve "all of physics" also caught my eye. One challenge there is that AI systems have been trained only on existing physics knowledge, which is constrained by the physicalist worldview in which it has existed for the past 300 years. So it's handicapped in that sense. However, I do wonder if an AI system could provide some kind of meta-insight that would break the physicalist logjam, as I suggested in a recent post https://open.substack.com/pub/cosmicwit/p/zombies-transhumanists-and-the-meaning?r=a1he9&utm_campaign=post&utm_medium=web
…people reading other people’s fiction as the blueprint for their/our construction for reality is a wildly dystopian idea also…yet here we are…i just wish the altman’s of the world would obsess over fixing problems instead of creating new ones…so much energy and intellect on this globe focused on their future when our present offers just as much opportunity…
You absolutely have to believe to run a con at the scale he's using, and to convince supposedly hard-nosed investors to,hand over that much money for literal promises and nothing else.
Thanks for the reminder about AI Snake Oil. It’s been on my radar for a while. I just bought it.
Have you been reading Matt Levine's bit on Sam Altman? I love it. As usual, he doesn't explicitly express disapprobation; on the contrary, he couches his remarks in a tone of admiration. I hope you will forgive me for quoting the Altman passage from his September 30th newsletter in its entirety:
"The place you want to reach in your career is where you work for a company and you are like “you know what, I am just so rich, I don’t want you to pay me anymore, it’s fine, I’ll work for free,” and your bosses are like “nope, sorry, we insist, we cannot allow you to work here for less than $10 billion.” And then you’re like “ohhhhhhhh fine, fine, fine, I do love working here, and if there’s really no other way, I guess I will take the $10 billion.” Nothing remotely like this has ever happened to me in my life but here’s Sam Altman:
'On Thursday, Altman told some employees that there were “good reasons” he shouldn’t take equity, though he didn’t elaborate. And he said investors were pushing for an equity grant to align his financial interests with those of OpenAI, said someone who heard the comments.
Altman also said a Wednesday news report that he might get a 7% stake in the new OpenAI was “ludicrous.”'
We talked about this last week: There have been reports that, as OpenAI becomes a for-profit company, it might give Altman (its co-founder and chief executive officer, who currently owns no equity) a 7% stake worth about $10 billion. I surmised that this was not something he wanted, but something the investors wanted: “He is the founder and CEO of a hot startup, and the founder and CEO of a hot startup is supposed to own equity. Not just for his sake — not just so that he can be rich — but to align incentives.” And here is Altman saying that: He doesn’t want the $10 billion, but the investors are insisting.
Nobody in history has ever been better at, like, business negging than Sam Altman. He got OpenAI to a $150 billion valuation in part by going around saying “oh no, nobody should allow us to build our product, we’re going to destroy humanity,” and now he is allegedly going to get handed a $10 billion stake in OpenAI because he’s going around saying “oh no, nobody should give me equity, that’s ludicrous.”"
It makes me wonder about the people giving him all that money, if they are this susceptible to such things.
Great read, as usual.
I remember a recent meme that said:
"I want AI to do my laundry, dishes, and clean the house, so that I can make art, write creatively, and express myself"
I think we are far away from that, and no, Musk's "Optimus" robots ain't gonna cut it.
Yes, it follows up with: instead it makes art and writes creatively, but very poorly/cheaply and I still have to do the chores. It is solving the problem that it should not solve and not doing a good job.
That’s what I was missing. I claim Covid brain mush
"Solve physics"! LOL. I guess AGI is Laplace's Demon. Why knew?
These people really need to read some philosophy of science. They are constantly conflating epistemology with ontology.
In addition: the way the techbros are reacting to the EU actually trying to reign in the worst excesses and risks is illustrative.
Ideas about geo-engineering are frightening, mostly because they have to act at such a scale that the risk is enormous. People who plan to remove 5% sunlight form reaching the earth by seeding space with reflecting confetti scare me to no end. What kind of disaster are we going to create if we suddenly remove 5% of the basic driver of everything on earth? Are you crazy?
But pushing things true without knowing what we're doing *is* kind of a human trademark. I think it would be fitting to no longer label us an 'intelligent species' until we actually start to behave intelligently (wars, fouling our ecosystem, the list is endless).
You’re on the Titanic. The engineers are shouting: “The bulkheads are too low! The rudder is too small! There aren’t enough life boats!”. The sailors mumble: “It has been cold, there will be many more icebergs than usual and further south”. The owners are pressing the captain: “You should be in New York in six days, we desperately need a record!”. And the captain thinks: “I have execution power. I can break through. I will be successful.” and orders: “Northerly course and full steam ahead!”.
Move fast and break things…
https://ea.rna.nl/2024/10/01/can-we-break-through-the-inertia-that-plagues-it-change/
Geoengineering is also stupendously silly or, more cynically, a money-making scam. It takes little account of the sheer size of our planet and the scale of pollution we have inflicted on it. In a way geoengineering is a tribute to its proponents' wilful ignorance of atmospheric chemistry and of the earth sciences in general. Their arrogance is breath-taking. We are currently injecting 50 gigatonnes of carbon gases—that's 50,000 million tonnes— into the skies each yea. Since the start of the Industrial Revolution we have released approximately 1.77 million, million tonnes. In turn, the Earth's atmosphere contains roughly 5x10¹⁵tonnes of gases. That's 5,000 million, million tonnes of gas. Step outside and look upwards. The atmosphere reaches to a height of at least 700km (438 miles-ish). All of it is gas, and it extends over the entire planet's surface (about 500 million square kilometres). How on earth do these surface-bound idiots think their tiny schemes could possibly work? Whatever they release (or sequester) will be so insignificant that it will have almost zero impact on climate change. Even their balloons need hot air generated by, guess what? By burning fossil fuels (or by using helium—an incredibly scarce resource needed for vital medical equipment). In 'Termination Shock' the plan is to mimic the effect volcanic eruptions have had in shielding the planet from solar radiation. Unfortunately for these dimbos, even a mega-eruption like Mount Pinatubo in 1991 'only' released just under twenty million tonnes of sulphur dioxide. It did have a mild effect on the climate for a few years but that was all. These geoengineers have no sensible explanation about how they are going to generate a even similar amounts of SO2 and get it to the upper atmosphere without, in that process, making the greenhouse effect even worse.
Like I said, it's simply a testament to the ignorance and arrogance of these people, nothing more. We can solve the problem, but only by everyone taking responsibility with collective action to reduce emissions. Their approach is very similar to the scientific absurdity and selfishness of longtermism.
I'll just note that the NY Times ran another article on geoengineering the day after the one you mentioned that lumps concerns about it with the anti-vaccine and conspiracy cranks -- https://www.nytimes.com/2024/09/26/climate/geoengineering-conspiracy-theorists-skeptics.html. That's certainly accurate to a degree, given M.T. Greene's recent tweet accusing the Dems of controlling the weather and causing the Helene damage, but it isn't leaving much room for those of us who worry about the tech bro types who miss the point of science fiction and can get their hands on some money.
As always, I appreciate your take on these things. Off to read your piece in the Atlantic.
"Science fiction" was a brilliant marketing misnomer for what it actually is. The vast majority of it save for a few exemptions, like Greg Egan, is just fantasy with science-talk in place of magic spells. I've honestly had a lot of trouble enjoying the entire genre outside of Greg Egan, Ada Palmer, and Kim Stanley Robinson out of just awareness of how sordid the history is.
Thanks for these quick hits. All very helpful. Altman's assertion that AGI would solve "all of physics" also caught my eye. One challenge there is that AI systems have been trained only on existing physics knowledge, which is constrained by the physicalist worldview in which it has existed for the past 300 years. So it's handicapped in that sense. However, I do wonder if an AI system could provide some kind of meta-insight that would break the physicalist logjam, as I suggested in a recent post https://open.substack.com/pub/cosmicwit/p/zombies-transhumanists-and-the-meaning?r=a1he9&utm_campaign=post&utm_medium=web
Regarding Altman:
"OpenAI's product is the promise that your boss can just fire everyone."
(Amy Castor and David Gerard, https://pivot-to-ai.com/2024/09/14/openai-releases-new-sora-demo-videos-with-a-few-less-glitches/)
Altman is an obvious charlatan, but that promise is exceedingly seductive to the myopically selfish people who dominate corporate capitalism.
Regarding hubristic tech bros:
I got bored and gave up on *Cryptonomicon* after three chapters. Perhaps that's indicative of why I didn't become a hubristic tech bro.
Regarding reading recommendations:
Thanks for recommending Narayanan and Kapoor's book, which I was already considering.
…people reading other people’s fiction as the blueprint for their/our construction for reality is a wildly dystopian idea also…yet here we are…i just wish the altman’s of the world would obsess over fixing problems instead of creating new ones…so much energy and intellect on this globe focused on their future when our present offers just as much opportunity…
The sad thing is that the Altmans of the world may very well believe they *are* fixing the world's problems.
You absolutely have to believe to run a con at the scale he's using, and to convince supposedly hard-nosed investors to,hand over that much money for literal promises and nothing else.
Hanlon's Razor might hold even here. Certainly Ilya and Geoff Hinton are in that camp.
Convictions are essential automation of human intelligence. https://ea.rna.nl/2022/10/24/on-the-psychology-of-architecture-and-the-architecture-of-psychology/
…i’m sure he will tell his therapist anything to fall asleep at night…
Probably doesn't need to.