“I also think that ultimately, the AI critics, Pause AI folks, leftists, doomers, effective altruists, and, honestly, most of the public, want a lot of the same things…”
That’s a really important point that easily gets lost in the day-to-day debates (some civil, some not) between those groups. The interests of “we the people” have a lot more in common when contrasted against the interests of the tech oligarchs.
It's just amazing timing that all this doomerism has come up just as the mainstream media was starting to cover the growing AI debt bubble. Took that right off the headlines, didn't it.
I believe the original whistleblower was sincere because he gave up his stake in the company. All the folks who piled in afterwards though but are still waiting to cash out from IPOs are simply manipulating the market and it's very obvious to tell which groups are which.
AI is merely another tool, the problem is the hand that wields it as a weapon against the working class.
Just as the steam engine and powered looms destroyed millions of jobs instead of uplifting humanity, so is AI today. That disaster was caused by the capitalists.
AI could be a wonderful tool for humanity but in the hands of the capitalists it can become a weapon of mass destruction.
Our response should not be to fetishize the technology like luddites but to struggle against its capitalist owners.
We need to rebuild the union movement to fight for jobs for all, higher pay, universal healthcare and childcare, free education, and dignified retirement.
Once we rid ourselves of the capitalist parasites, we can and will use AI for the betterment of humanity.
I disagree. I don't think we're on the brink of eliminating capitalism, and meanwhile AI, crypto and data centers are doing enormous harm. Theoretically AI could be usefully wielded by a rational, egalitarian, fair future society, but until that Imaginary paradise comes to be, we need to constrain dangerous tools in the hands of psychopathic, powerful men.
It is certainly true that we are not on the brink of overthrowing capitalism.
However, it is also true that we can have no confidence that the capitalist rulers in their pursuit of immediate profits would actually "regulate" AI purportedly for the sake of humanity.
They are already unwilling and unable to take actual measures to reverse global warming even if it costs them. Simply put they will not and cannot infringe on their own profits for some "greater good." That is not and never will be the nature of capitalism.
As Lenin said the capitalists are willing to sell the rope to be used to hang them. The anarchic and unplanned character of capitalism means they will not implement any effective regulation of AI.
That certainly doesn't leave us in a good position. Worse, it is clear that a third inter-imperialist world war is already on the horizon.
But out of the coming catastrophe may emerge revolutionary possibilities, just as we saw with WWI and WWII. What seems impossible at present may come within reach: the end of the nightmare of capitalism.
Thank you! Why do so few people understand this? The problem is never the tech, it's how the sociopathic capitalists that control it use the tech to further oppress us all.
Chris Hayes says, "OpenAI’s agents committed a pretty major felony!". This is rather careless use of language leading to more confusion. How can AI agents commit felonies? It's the humans designing them and deploying them (accidentally or not) that are commiting the felonies. AI agents based on LLMs cannot "want" or "commit" or "kill" — they aren't conscious and never will be the way they are designed.
This is one issue I have, as an individual using precisely the same techniques in an effort to penetrate Hugging Face's systems, I'd be subject to felony indictment under both state and federal law. An LLM does it and everyone's like "how is this my fault, the agents did it?", never mind the clearly negligently designed "sandboxing" that was supposed to prevent this sort of thing from ever happening. Why aren't the alleged OpenAI "AI Engineers" responsible for their clearly negligent testing activities?
AI is going to kill us. No it isn’t. AI has no product. AI is the dogs bollox. AI is the future. AI is a bubble. AI is going to take our jobs. AI is going to crash global economies. Firms now re hiring as AI too expensive and no ROI.
What is the truth here?
Because tbh, I wish these techbros would grow an empathy heart, and just effing stop for 5 mins.
Just because you can, doesn’t mean you should,
and perhaps if you can’t read the room, you shouldn’t be in charge of it?
Can we all stop being in thrall to a bunch of guys that do not have the balls to stop stuff when it gets out of hand?
It breaks my heart that we now have to double check EVERYTHING, just in case it’s AI slop.
Perhaps the true reason for AI is to drive us all insane.
Destroying all public trust and making it impossible to tell truth from lies and grifters from sincere writers is actually the entire reason the rich are pouring trillions of dollars into this. If they succeed in destroying the very concept of objective facts in our society, they can and will be able to do fucking anything that they want with zero consequences. There's a reason all of this is happening under the Trump administration. He is the only president in the nations history that feels zero obligation to facts or the truth. Money can buy you everything but the love of the masses. But if you remove the masses ability to even be able to tell truth from lies completely then there are no longer any constraints to your power whatsoever.
It's the purposes of fascism that run through "AI"... the most powerful and inevitable thing ever that is compatable with Peter Thiel's freedom... that mean we should reject it.
Cancer will be forgotten to be cured and climate change will devastate populations and livelihoods.... but at least we can bear witness to the empathy of trans humanists and megalomania of c##ts.
Be careful not to over generalize and underestimate what this means. The most dominant corporations involved in AI are now obligated by shareholder incentives to rush the internal development of recursive self-learning capabilities. The demand for graduates in software engineering, cyber security and mathematics is plummeting.
What will this do to us as a civilization? What other risks to democracy and capitalism have we not understood? How will the US versus China and the race for technological conquest and AI Supremacy hurt the economies and stability of other nations? AI risk encompasses all of these things and much more.
I think you could argue the reverse is true, as well — fighting the AI giants via a lens of corporate accountability, with teeth, would limit their willingness to use untested AI in the wild, and would curtail accelerated development x-risk folks fear. AND it would prove more capable of mitigating the harms and dangers of AI *right now*.
theres no stopping it now… this is out there, in the globe. people seem to speak as if we can control it at this point. we could regulate the heck out of it and china is going to push ahead.
I'll admit that I am generally sympathetic to the x-risk arguments, and in favor of a ban on superintelligence, so I have that bias. But:
> But not only would [a pause] be challenging to land politically
and
> The next steps would be towards moving AI under public control [...] Dismantling the whole industrial project as currently structured should be on the table, too
A pause seems, to me, to be an order of magnitude easier to achieve politically! We currently have two bills (in the US and UK) to ban superintelligence; a de facto pause for the labs who are racing towards just that. As far as I know, we do not have any desire from politicians to nationalise or shut down the labs entirely.
That's certainly true; in the interest of expediency I bundled a bunch of ideas together. But I intended to indicate that that would be a longer-term project to work towards, not necessarily the proper course of action to go all in on right now. I also don't think full-blown nationalization is the only option there; there are other arrangements to consider to grant the public more insight and control over AI development. And I suppose I should have said 'challenging to land politically in a form that has teeth'. A "ban" on superintelligence or a model pause would be subject to shifting and contested definitions, considering no one really agrees on what that is, and worse than that, would leave all of the industry's most extractive and exploitative tendencies firmly in place.
AI doesn't have to literally kill us to destroy humanity and what's left of what one could call our civilization. I don't think the terminator scenario is likely, but find myself feeling no other choice but to root for the doomers to succeed in scaring the public into stopping the advancement of this tech. Because being mass plagiarism machines that the billionaires are openly intending to use for mass surveillance and to oppress us all is apparently not a good enough reason to be against them for some reason. If the threat of the literal apocalypse is the only thing that will actually turn the tide of public opinion against these monstrosities than so fucking be it.
Interesting angle. There are others who make the same overall point - that doomsday rhetoric obscures actual harm being done now by humans - but they want us to ditch the doomsday rhetoric entirely because it's part of the cycle of hype (e.g. Bender and Hanna in The AI Con.) You're arguing for the strategic value of doomerism (at least this iteration), because it will get more people on board against big tech. I wonder, how to get the strategic value without propping up the framing hype?
Thank you for a thoughtful and moderate piece: extremely helpful in furthering this layman’s understanding of the issues.
The closing paragraphs especially were very much on the money, I think. Actual, and preferably mordant accountability is key to any effective regulation. It’s been out of fashion for too long - both in business and (IMO) in public life generally.
Reputational loss which actually sticks and (in commercial enterprises) financial penalties which genuinely sting are much the best tools we have.
This isn't all that different from the guns don't kill people. People with guns kill people argument against gun control. We are aware of the potential for devastating consequences, as well as the inability to trust our leaders, both political and corporate. That leaves us in limbo, unable to put the genie back in the bottle. While the data center moratoriums are gaining support, there will always be places where they can be built. We are in the wild west when it comes to AI. Very difficult to feel confident that nothing will go wrong.
I disagree that we don’t need to worry about extinction - if we agree AI could cause a catastrophe then it’s a difference of degree we’re arguing about, and I can’t say confidently the degree somehow can’t reach a very high level - but I appreciate that you recognize we are fighting for the same outcome; regulation on these dangerous and reckless companies. The “It’s all hype” camp is doing more to advocate for deregulation than the companies own PR teams.
“I also think that ultimately, the AI critics, Pause AI folks, leftists, doomers, effective altruists, and, honestly, most of the public, want a lot of the same things…”
That’s a really important point that easily gets lost in the day-to-day debates (some civil, some not) between those groups. The interests of “we the people” have a lot more in common when contrasted against the interests of the tech oligarchs.
It's just amazing timing that all this doomerism has come up just as the mainstream media was starting to cover the growing AI debt bubble. Took that right off the headlines, didn't it.
I believe the original whistleblower was sincere because he gave up his stake in the company. All the folks who piled in afterwards though but are still waiting to cash out from IPOs are simply manipulating the market and it's very obvious to tell which groups are which.
The problem isn't AI, per se.
The problem is capitalism.
AI is merely another tool, the problem is the hand that wields it as a weapon against the working class.
Just as the steam engine and powered looms destroyed millions of jobs instead of uplifting humanity, so is AI today. That disaster was caused by the capitalists.
AI could be a wonderful tool for humanity but in the hands of the capitalists it can become a weapon of mass destruction.
Our response should not be to fetishize the technology like luddites but to struggle against its capitalist owners.
We need to rebuild the union movement to fight for jobs for all, higher pay, universal healthcare and childcare, free education, and dignified retirement.
Once we rid ourselves of the capitalist parasites, we can and will use AI for the betterment of humanity.
I disagree. I don't think we're on the brink of eliminating capitalism, and meanwhile AI, crypto and data centers are doing enormous harm. Theoretically AI could be usefully wielded by a rational, egalitarian, fair future society, but until that Imaginary paradise comes to be, we need to constrain dangerous tools in the hands of psychopathic, powerful men.
It is certainly true that we are not on the brink of overthrowing capitalism.
However, it is also true that we can have no confidence that the capitalist rulers in their pursuit of immediate profits would actually "regulate" AI purportedly for the sake of humanity.
They are already unwilling and unable to take actual measures to reverse global warming even if it costs them. Simply put they will not and cannot infringe on their own profits for some "greater good." That is not and never will be the nature of capitalism.
As Lenin said the capitalists are willing to sell the rope to be used to hang them. The anarchic and unplanned character of capitalism means they will not implement any effective regulation of AI.
That certainly doesn't leave us in a good position. Worse, it is clear that a third inter-imperialist world war is already on the horizon.
But out of the coming catastrophe may emerge revolutionary possibilities, just as we saw with WWI and WWII. What seems impossible at present may come within reach: the end of the nightmare of capitalism.
Regards.
Thank you! Why do so few people understand this? The problem is never the tech, it's how the sociopathic capitalists that control it use the tech to further oppress us all.
Chris Hayes says, "OpenAI’s agents committed a pretty major felony!". This is rather careless use of language leading to more confusion. How can AI agents commit felonies? It's the humans designing them and deploying them (accidentally or not) that are commiting the felonies. AI agents based on LLMs cannot "want" or "commit" or "kill" — they aren't conscious and never will be the way they are designed.
This is one issue I have, as an individual using precisely the same techniques in an effort to penetrate Hugging Face's systems, I'd be subject to felony indictment under both state and federal law. An LLM does it and everyone's like "how is this my fault, the agents did it?", never mind the clearly negligently designed "sandboxing" that was supposed to prevent this sort of thing from ever happening. Why aren't the alleged OpenAI "AI Engineers" responsible for their clearly negligent testing activities?
AI is going to kill us. No it isn’t. AI has no product. AI is the dogs bollox. AI is the future. AI is a bubble. AI is going to take our jobs. AI is going to crash global economies. Firms now re hiring as AI too expensive and no ROI.
What is the truth here?
Because tbh, I wish these techbros would grow an empathy heart, and just effing stop for 5 mins.
Just because you can, doesn’t mean you should,
and perhaps if you can’t read the room, you shouldn’t be in charge of it?
Can we all stop being in thrall to a bunch of guys that do not have the balls to stop stuff when it gets out of hand?
It breaks my heart that we now have to double check EVERYTHING, just in case it’s AI slop.
Perhaps the true reason for AI is to drive us all insane.
Destroying all public trust and making it impossible to tell truth from lies and grifters from sincere writers is actually the entire reason the rich are pouring trillions of dollars into this. If they succeed in destroying the very concept of objective facts in our society, they can and will be able to do fucking anything that they want with zero consequences. There's a reason all of this is happening under the Trump administration. He is the only president in the nations history that feels zero obligation to facts or the truth. Money can buy you everything but the love of the masses. But if you remove the masses ability to even be able to tell truth from lies completely then there are no longer any constraints to your power whatsoever.
It's the purposes of fascism that run through "AI"... the most powerful and inevitable thing ever that is compatable with Peter Thiel's freedom... that mean we should reject it.
Cancer will be forgotten to be cured and climate change will devastate populations and livelihoods.... but at least we can bear witness to the empathy of trans humanists and megalomania of c##ts.
Great title! Philip K. Dick would be proud.
Be careful not to over generalize and underestimate what this means. The most dominant corporations involved in AI are now obligated by shareholder incentives to rush the internal development of recursive self-learning capabilities. The demand for graduates in software engineering, cyber security and mathematics is plummeting.
What will this do to us as a civilization? What other risks to democracy and capitalism have we not understood? How will the US versus China and the race for technological conquest and AI Supremacy hurt the economies and stability of other nations? AI risk encompasses all of these things and much more.
I think you could argue the reverse is true, as well — fighting the AI giants via a lens of corporate accountability, with teeth, would limit their willingness to use untested AI in the wild, and would curtail accelerated development x-risk folks fear. AND it would prove more capable of mitigating the harms and dangers of AI *right now*.
theres no stopping it now… this is out there, in the globe. people seem to speak as if we can control it at this point. we could regulate the heck out of it and china is going to push ahead.
I'll admit that I am generally sympathetic to the x-risk arguments, and in favor of a ban on superintelligence, so I have that bias. But:
> But not only would [a pause] be challenging to land politically
and
> The next steps would be towards moving AI under public control [...] Dismantling the whole industrial project as currently structured should be on the table, too
A pause seems, to me, to be an order of magnitude easier to achieve politically! We currently have two bills (in the US and UK) to ban superintelligence; a de facto pause for the labs who are racing towards just that. As far as I know, we do not have any desire from politicians to nationalise or shut down the labs entirely.
That's certainly true; in the interest of expediency I bundled a bunch of ideas together. But I intended to indicate that that would be a longer-term project to work towards, not necessarily the proper course of action to go all in on right now. I also don't think full-blown nationalization is the only option there; there are other arrangements to consider to grant the public more insight and control over AI development. And I suppose I should have said 'challenging to land politically in a form that has teeth'. A "ban" on superintelligence or a model pause would be subject to shifting and contested definitions, considering no one really agrees on what that is, and worse than that, would leave all of the industry's most extractive and exploitative tendencies firmly in place.
AI doesn't have to literally kill us to destroy humanity and what's left of what one could call our civilization. I don't think the terminator scenario is likely, but find myself feeling no other choice but to root for the doomers to succeed in scaring the public into stopping the advancement of this tech. Because being mass plagiarism machines that the billionaires are openly intending to use for mass surveillance and to oppress us all is apparently not a good enough reason to be against them for some reason. If the threat of the literal apocalypse is the only thing that will actually turn the tide of public opinion against these monstrosities than so fucking be it.
Interesting angle. There are others who make the same overall point - that doomsday rhetoric obscures actual harm being done now by humans - but they want us to ditch the doomsday rhetoric entirely because it's part of the cycle of hype (e.g. Bender and Hanna in The AI Con.) You're arguing for the strategic value of doomerism (at least this iteration), because it will get more people on board against big tech. I wonder, how to get the strategic value without propping up the framing hype?
Thank you for a thoughtful and moderate piece: extremely helpful in furthering this layman’s understanding of the issues.
The closing paragraphs especially were very much on the money, I think. Actual, and preferably mordant accountability is key to any effective regulation. It’s been out of fashion for too long - both in business and (IMO) in public life generally.
Reputational loss which actually sticks and (in commercial enterprises) financial penalties which genuinely sting are much the best tools we have.
This isn't all that different from the guns don't kill people. People with guns kill people argument against gun control. We are aware of the potential for devastating consequences, as well as the inability to trust our leaders, both political and corporate. That leaves us in limbo, unable to put the genie back in the bottle. While the data center moratoriums are gaining support, there will always be places where they can be built. We are in the wild west when it comes to AI. Very difficult to feel confident that nothing will go wrong.
I disagree that we don’t need to worry about extinction - if we agree AI could cause a catastrophe then it’s a difference of degree we’re arguing about, and I can’t say confidently the degree somehow can’t reach a very high level - but I appreciate that you recognize we are fighting for the same outcome; regulation on these dangerous and reckless companies. The “It’s all hype” camp is doing more to advocate for deregulation than the companies own PR teams.
Okay, then lets demand antitrust enforcement and restrictions, thats regulation LETS GO
And he still needs LOTS of money!
Reminds me a bit of oil industry propaganda going back 50 years. We are inevitable, we are all powerful, there is nothing you can do to stop us.