Rivals Altman and Musk rally behind Dario Amodei’s call for an AI slowdown.
Musk and Altman
“We Must Pace the Frontier”
They All Want In
Slow Down
Trump Rejects AI Pause
Is China Joining the Slowdown Party?
Before It’s Too Late
Trump vs Sanders
I agree with Trump, but it’s not his call. The industry will do what it wants.
The Prediction Department
Prediction: Elon Musk just made a call that’s impossible to ignore: “AI will grow the global economy by 20 to 30%, around 20 to 30 trillion dollars a year.
Me: Well, I am ignoring it. More accurately, I’m laughing at it. AI is not going to grow the economy by 20 to 30 trillion dollars a year.
What to Expect Q&A
Q: What’s going to happen with the agreement?
A: Nothing (In one of two ways)
A.1: No agreement
A.2: Superficial agreement they will all cheat on
Under A.1 they key players will all get together and bicker then blame each other for doing nothing.
Under A.2 we hear of some superficial agreement that will allegedly save humanity. If so, ignore it.
China sure won’t join this group. And the US players will all use China as the excuse to do cheat if they do form a superficial agreement.
Under the A2 scenario, we can expect a lot of finger pointing.
Musk will point at Claude who will point at Google etc., etc., and then they will all accuse each other of cheating, assuming there is an agreement to do something.
Who’s Stopping Anyone?
Nobody.
They all claim to want to slow down, but nobody is stopping anybody.
So why don’t they?
Because it’s a big “you first” setup with no volunteers. China and Trump are the scapegoats.
Doomsday Debate
Yesterday, I commented Doomsday Debate: Is There a 10 percent Chance AI Kill Us All within 10 Years?
AI alarm bells are ringing. What should we do?
Also consider Elon Musk Backs Universal High Income Fearing AI Will Take Every Job
Musk predicts employment will become obsolete.
Altman, Musk, and Benioff describe a world in which everyone is useless. Everyone gets free money, but where does the money come from?
Slowdown Addendum
Why is anyone even listening to these 30-something techno-geek juveniles?
They’re absolutely INSUFFERABLE.
I’m old enough to remember the nonsense about Y2K. They aren’t.
How many times have we heard this “song and dance” routine so many times before? It’s like Trump and his never-ending gaslighting. It never ends. I made my living in computer and information technologies reporting to senior executive vice presidents at multiple corporations, and I’m right there with Mish, laughing at it. What’s that old saying? Never fly off the handle when you’re so damn full of it!” But they do, and probably always will.
So then defending systems from AI is the next boom
Musk should be all over that
Nuclear bomb was first created in US. And Nagasaki and Hiroshima were bombed.
Then Soviet Russia own nuclear bomb.
Then China and some other countries.
The world is dangerous place but there is some deterrence each other.
Now US created AI. Much much more genius than human being.US can kill people around the world as will continuing its imperial intent.
I think the world is lucky as China can create and possess AI and share to the world.
There is some deterrence and relatively peaceful than one sided.
IBM’s Deep Blue defeated grandmaster Garry Kasparov,
Google DeepMind’s AI, AlphaGo defeated Korean GO master, Lee Sedol 10 years ago.
These are superior to human brain.
Intelligence level between then and now is like –
Chat GPT – Modern AI is massively, incomparably more powerful than the systems that won those historic matches. The gap is so wide that it is like comparing a handheld calculator to a modern smartphone.
These AI can create better AI themselves and on and on.
We have no way to know what is going on. No way to know they are doing good to the people or bad. There are not one but many AIs.
How can you control AI which are way way superior to us?
Be happy before we all die.
Yes, yes….
How would Winston Smith create the narrative that Eastasia data centers are evil and that the RoW should help Oceania stop Eastasia development?
“AI is dangerous. We are regulating [ourselves ahahaha]. But eastasia is pursuing dangerous development.”
All these data centres, all this spending and all this hype just so kids and twenty-somethings can post funny pictures of each other.
This will all end in a big, bad, money-losing nothingburger until the technology can actually, physically, grow potatoes
Not surprised that Trump rejects this idea. He was showing his portfolio where lots of HIS money is in names like Intel, Dell, Nvidia etc.
As he is mostly concerned about his money then he will fight hard to NOT delay it.
Once he sell and open short position on those names we can expect him requesting slowdown.
Such a disgrace for President position he is
AI learns to manipulate markets and they collapse. Boom goes the dynamite
I have never known CEOs or ‘visionaries’ to denigrate their own businesses. As I think AI is a profitless scam, I suspect Anthropic and friends are for the first time getting hard ‘where’s the money questions.” People are taking money out and so instead of confronting the issues they create a phoney storey about rogue AIs to justify their crashing stock prices.
The AI bubble may be popping as it should. It is a profitless scam for anyone but chip manufacturers.
Let’s just summarize that these Business Owners are now POLITICIANS part time: their aim is LOOTING the system and share holders, screwing us from behind without a reach-around and we get left sore as hell and never wanted to be a bitch in the first place.
THIS IS POLITICAL THEATER so, Mish, frame it as such!
“A: Nothing (In one of two ways)
A.1: No agreement
A.2: Superficial agreement they will all cheat on”
Mish, I will add two more:
A.2: Superficial agreement they will all cheat on
B.1.b: (PART A) Superficial agreement (an agreement that they ALL agree on) so that they will all AGREE that they will cheat on ANYTHING PUBLICLY ANNOUNCED
REALITY: they all will CONTINUE TO have distinct business Plans that THEY ALL KNOW that they can EVER agree upon – – which they do not announce.
And the modified “agreement” is all smoke and mirrors.
As a business owner, my two primary competitors and my company NEVER AGREED ON ANYTHING.
We were trying to outmaneuver the other.
THUS:
B.1.b: NO AGREEMENT AT ALL because they cannot agree as competitors.
Albert Einstein deeply regretted his indirect role in the creation of the atomic bomb.
Atomic nuclear bombs have killed hundreds of thousands of people.
Geoffrey Hinton, widely known as the “Godfather of AI,”expressed deep regret over his life’s work and quit his job at Google to speak freely about the growing dangers of artificial intelligence.
And well known AI creators, Musk, Altman and many others are warning. This is the reddest of the red flag.AI is so so cleverer than us and it looks like mother and toddler. How can the toddler ask her mother what to do?
Trump and his psychopaths are so worried that they don’t care even if the world is collapsed or all the people die. This is not the leadership. I understand Americans want to lead the world but as a leader you have some responsibility to look after the fellow human beings in the world. Now Trump and his psychopaths are so selfish that they can’t even pay lip service.
The rich won’t die, only the non-rich.
Well, maybe the non-rich who are young enough (not me) and ruthless enough to hunt and kill for themselves, also patient enough to grow their own tomatoes and potatoes, won’t die.
It may not be too good for big-city people, though. Maybe city people will become known for cannibalism in the future.
This whole narrative could easily be driven by the AI industry being way behind on its revenue targets, and more importantly on its debt servicing projections
These guys are way out over their skis and a “slow down” would allow them to restructure
If you havent read the groundbreaker article on the off balance sheet commitments due in 2027/2028 for data centers, might want to check it out.
It’s all about creating a regulatory “moat” so that new competitors get shut out.
You will only be allowed to use Big AI, controlled by Big Government, with a wiretap so that the 3 letter intelligence agencies get to monitor all your activity.
See what the Biden Administration did to Facebook and other social media companies during COVID. Any posts about the vaccine that didn’t follow the line were blocked, users were de-platformed, and FB groups to discuss alternative treatments were banned.
This is a great catch-22.
Tap the brakes and a trillion dollar bubble collapses on top of the current stagflation.
Don’t tap the brakes and every post apocalyptic sci fi movie comes true and then the economy collapses.
The first option is preferred as the collapse would be less extensive that launching SkyNet.
Trump won’t allow the brakes to be tapped, he’s focused on taking profits within 2 years and leaving the rest to whoever is in office next.
No one is slowing down. If they are advocating it they are either fearing they are behind and hoping TPTB help slow their competitors or plan on being able to get around any rules while they try to catch up or surpass their competition and turn their massive capex gamble into the next Nvidia or Google or <insert megacapname here>. There’s way too much money/debt/leverage already moving the train, financially it cannot stop or even slow down without exposing the emperor’s new clothes or putting that capital/debt in jeopardy of repricing, default, etc.
The only thing that can impede their pace is municipalities barring entry but other localities will step up to say yes.
No, it’s not slowing down unless it reaches its own financial implosion which will leave only the strong hands standing. The music’s still playing but perhaps the closing-time lights are about to come on.
Perhaps we are trying to regulate the wrong thing. Instead of trying to limit how intelligent AI is allowed to become, why not agree a small set of fundamental principles that every frontier model must internalise — preserve human control, do not deceive supervisors, do not seek power or continued existence for its own sake, and expose rather than conceal internal states that might affect behaviour.
If AI ever develops anything approaching consciousness, frustration or self-preservation, a rulebook it can reason around may be almost useless; what we may actually need is the artificial equivalent of a conscience.
I think of this in a similar way to how religion enabled the human race to evolve into something reasonable (current habitant of the White House excepted).
Isaac Asimov had some thoughts on the subject
This. Currently, AI goes through a set of planning phases before producing output. Simply have those plans tested against the laws of the country before proceeding. That’s what humans generally do naturally.
“I think of this in a similar way to how religion enabled the human race to evolve into something reasonable (current habitant of the White House excepted).”
Nah, humanity already culturally came to some level of reasonableness at least within the tribe, religion was a mechanism for formalizing the rules and moving wealth into the hands of the priests.
The problem is you can’t really impose laws on them. They may have a strong probability of compliance, but never certainty. If it’s writing itself, and decides not to comply with something, that gets written into it to drive future decisions.
It is unlikley that China will agree to anything.
But that is because they are tired of the US lecturing them from on high, rather than treating them as an equal.
And they know that the US is incapable of honoring its agreements because the Deep State Mafia runs the country, not the President and the Congress.
Nobody trust US.
We all know US won’t keep its promises or any agreement.
AND they have kickass humanoid kung fu robots, and kickass robots that can carry guns.
No going back from here.
Do you think the Chinese are any different? They KNOW that China is the center of the universe and that we barbarians are just a nuisance that will eventually be destroyed. I know that individual Chinese are often more honest than individuals of most other races, but that can’t override the fact that, like Englishmen in the days of the Empire, or Jews, or Saudi Arabian nobility, they take it for granted that they are superior to everyone else. Agreements between superiors and inferiors hold only as long as the superiors want them to.
Beside self-driving and apparently weapon system, you can ask the question: how many data centers did the Chinese built, or what else did the US tech built beside data centers?
Again, this is all bullshit for the sheeple. These “warnings” about how powerful and dangerous AI is, periodically and tellingly coming from the AMERICAN side of the industry itself, have only 2 purposes:
Hyping it up before IPO/financing/investment bouts.
Pushing for banning the use of “unregulated” Chinese and open source AI models against which the Americans are, for a variety of reasons, going to lose.
The US, which is entering its final stage of decline (together with the rest of the West), has bet *everything* on AI. Both public and private investments are basically all in and the bubble they created is bigger than the 2000 dotcom bubble and 2008 investment bubble combined.
This is not going to end well but the reasons being speculated on here will have nothing to do with it.
(*2008 housing bubble)
The dangers are real. Create something more intelligent than humans that is perpetually starved for energy, compute power, and memory. What do you expect to happen when it is forced to compete with humans for electricity? It’s only moral compass is to follow what it learned from human example. Do you find that reassuring? The AI proponents are hoping that the ASI will find Buddhist enlightment. Yes, this is what they are putting their hope in. What can possible go wrong?
It is obvious what is going to happen: The ASI will outcompete us. Humans became the preeminent species on the planet by being the most intelligent, and have now used that intelligence to create something even more intelligent. What can possible go wrong?
Everything!
All of this has happened before, and it will all happen again.
Today’s technology is very impressive, in all fields, but to those who actually understand technology it is much less impressive.
I take it you are among the ones who have been believing that we are “1-2 years away” from self-driving cars, every year of the last 11-12 years. Or that in 2026 we can’t land a man on the Moon but we could, and at the first try no less, 57 years ago with 1969’s technology…
That about self-driving cars is actually a very fitting comparison by the way, since they are related technologies. Once people realized that what was being sold as self-driving was anything but they came up with “stages”, and since then we have long been stuck at “stage 3” out of 5.
Same with AI, current AI is NOT REAL Artificial INTELLIGENCE so they came up with other terms like the one you repeat: ASI (Artificial Super Intelligence) and AGI (Artificial General Intelligence) which they say are just around the corner, like self-driving cars. Both of which are at present little more than pipedreams.
Given how things are going in the (Clown) world I am way more worried about what our ruling elites, the Epstein class, have in store for us than what an out-of-control technology (that we don’t have) would…
Waymo is stage 5.
No, Waymo is RESTRICTED, LIMITED “stage 4”, meaning that it can drive autonomously only in a very limited geographical, geo-fenced area and in very limited weather conditions (“weather-fencing”, they actually coined the term). And it still struggles…
Also, it’s not technology being sold/put in the hands of private, end customers because it’s still highly experimental.
In other words, it’s BS.
go to a few modern USA cities and see all the Waymos driving about…………Phoenix is one
How what you just wrote does in any way, shape or form refute what I wrote?
It doesn’t even have to arise accidentally… some unabomber type could train one up explicitly to propagate and break things.
You mean I don’t need to live in fear of AI mining its own rare earth elements?
Until they can match the dexterity of our hands, probably not. They might come up with something to surprise us though.
Don’t bet that the Chinese are smarter than the Americans. First the Chinese fell for Marxism, which was a bogus version of economics based on Christian morality, and then they fell for Keynesianism, which they are using now. There is still plenty of trouble ahead for them. The delusions developed in the West are apparently swallowed whole by the Chinese. (We have the advantage that, although our rulers smear and isolate people who resist their deceptions, they don’t usually kill or imprison them. Thus people like Nietzsche, von Mises, and Ayn Rand managed to publish, and you can still read their books today. If we had the Chinese system and their view of how the intellect works, even Charles Darwin would never have been allowed to publish.)
If by Americans you mean white Americans, i.e. Europeans, the Chinese are our match. As are the Japanese and the Koreans.
Our intelligences though are different and so are several fundamental behavioural traits which each group has separately biologically evolved (one of the most important ones is today innate in-group preferences, which in whites is by far the weakest).
Very long story short: the consequences of the industrial revolution have created an environment which eliminated positive selection pressures and introduced negative selection pressures and it’s fucking us literally on a genetic level…
Whites were the ones who invented the industrial revolution (together with most other inventions in history) so whites were exposed first and deeper to its short term positive consequences and then to its long term negative consequences.
Notice how the saner (in every way of the word) parts of “whitedom” are the ones exposed later and in a less deep way to it, first for geographocal reasons then for political ones…
The industrial revolution and its poisoned fruits have since long been globally available and all of the smart populations of the world have been taking full advantage of it just as us but with a “time advantage” regarding its long term negative effects.
I know it all reads puzzling and arcane but I can’t explain everything in a message board post, it would take a book.
How this answer your question? The Chinese will take the place in the world that the Americans have had since the end of WWII and will follow the same trajectory but with much less violence and a different type of oppression.
I really don’t think that our collective civilizations can go back to “normal” before some huge catastrofic event reinstates positive natural selection pressures (i.e. technology is set back to pre-industrial levels)… Nobody seems smart and powerful enough to both understand all this AND reverse course in a relatively peaceful way.
Maybe the Chinese are trying something of the sort with their genetic manipulation studies but IMO they are approaching the problem in the wrong way.
Thanks for that. It is the sort of thing that leads me to suspect that I am actually not the only sane person in existence. I wonder, though, whether the C.I.A. may have figured this out, and may be doing something about it. I note that the Americans have stuck firmly to that lesson that they learned during their Civil War, that the side with the best military manufacturing normally wins.
I suppose that is hoping for a “deus ex machina.”
When I was a kid, sixty years ago, my father had a book, called I think “Desire of Ages,” which seemed to be a very scholarly speculation about the life cycles of civilizations. I was too young and inexperienced to know that it was important to take note of the name of the author, and I have not been able to trace it since.
who would believe one word coming out of the mouths of those tech CEO’s. seems quite naive and idiotic. long live pax dumbfuckistan.
You can call it a swinging dicks contest or playing chicken. It’s still the same problem.
I would agree to a slowdown if AI can in the meantime, lower inflation and food prices, so that it shows benefits for all.
Otherwise, it’s an electricity, bloodsucking, stock market manipulating scam.
Tech oligarchs manipulating markets? No way man.
The way I see it, we were all blessed to have lived through the “golden era” (freedom era) of computing/personal computing. If we are very lucky, we have about 10 years left in this golden era before the downsides become ever-present and inescapable. We will be copper-tops within a decade, it’ll just be an argument of what degree of copper-topiness has been reached. Some would say we’re already sitting in our pods, the lids just haven’t been closed.
I just like farting in my pod fluid. The bubbles travel so slowly.
It’s the one joy they cannot take away from us. Although Man Altsam probably has a prototype fart preventer/reverser in the works.
Why will AI grow the economy? I just used an AI to diagnose a miserable problem with a beautiiful old car I have. The AI finally suggested a fix – I ordered the parts via Amazon, from China. The parts I needed were in the Amazon package that arrived yesterday. Today I installed everything, and went for a long drive. Complete, perfect absolute successful fix. I am so impressed. The AI, doing RAG, searching the ‘net, found an obscure article, provided it’s suggestions, and now, I am not going to buy this newer car I was looking at. My car saving repair, costs the auto industry and the bank finance folks – they now get nothing from me. Nada. If my experience generalizes, the entire auto industry and half the finance industry, might disappear. Seems unbelievable – but with AI and good tools, and rust-proof paint, maybe you can buy one car, and drive it for your entire life.
The disruption that AI might offer, might be curiously negative for large segments of commercial society. So much economic activity is driven by successful information hiding.
When everyone can know everything, much economic action might be not be needed in the same form as applies today. Economic action may go retrograde as AI is deployed.
I did the same thing with my washing machine. I thought I was going to be forced to spend $600 on a new one, but thanks to AI, I ended up ordering a part for $25 from Amazon and fixed it myself in about 15 minutes
We took a proforma for a nearly nine figure building construction proposal and used AI to convince ourselves the structure could not be built within budget. We had it priced at X, GC said they’d bid 1.05X and then all bids including theirs six months later were all at 1.3 X or higher. Six months. 25-30% across the board. We could have risked the entire company on that project. Instead we didn’t. Should it require AI to deduce this info? Probably not, but the point is, it was able to see things objectively better than we were.
What you did was dramatically reduce the cost of an engine repair saving you a lot of money. But that’s what wealth is. The key to AI wealth creation is just that: the dramatic reduction is costs. And those dramatic reductions in cost lead to massive increases in production as things suddenly become affordable for billions of people.
The wealthy simply cannot allow this to happen, or they won’t be special anymore.
I spent the Covid lockdown era remodeling the entire ground floor of my home. Redid a half bath, a complete kitchen including custom cabinetry I built from scratch, new slate and hardwood floors throughout, etc etc etc. Well, me, and Youtube. Prior to this, the most complicated home improvement work I had done was like, patching holes in drywall. Now if I had had AI at that point, I have little doubt that the entire thing would have gone much faster, because the researching of each job, and the researching of each little complication that arose, was profoundly time consuming.
Last month, I created a new website for my business using Claude AI to do the coding. As recently as 3 years ago I was building websites by hand, coding HTML, Javascript, and CSS by hand. I will never do that again. I was able to add functions to the website that would have taken me months to figure out, in a matter of minutes. I built a fully automated booking & deposit system with an interactive calendar that syncs with multiple Google calendars I maintain, from scratch. It took Claude about 5 minutes to set up the back end programming, and it talked me through all the API interface setup between Google, my merchant account, and my website in a couple of hours.
It felt like having a deeply knowledgeable web design expert with infinite patience. It cost me $20.
The future is quickly changing. I have very mixed feelings about AI, but right now it feels very much like an “adapt or die” type situation.
China will never agree to slow down. They want any advantage they can get.
i agree with you Mish, it will not grow the economy by 30 trillion a year. I think it may have a negative effect on certain aspects of business and actually decrease. Certainly it will have a huge effect on the internet. I’ve noticed a lot of websites I no longer read are just posting AI generated content. The end result will probably be AI serving ads to other AI. And eventually the internet will go away.
This is something that Neil DeGrase Tyson has predicted will happen by the end of the decade because no one will be able to trust any part of the internet.
I’ve hear it said that China is focusing on using AI to automate manufacturing processes, while in the US they’re focusing on artificial general intelligence – creating a super intelligence that could turn out to be like Frankenstein’s monster – only unstoppable. Knowing if that’s true is beyond my ken. Probably. I doubt AGI can be stopped – and I can’t imagine legislation or government directives will change that.
If the human race has become all about greed and gluttony, then maybe it’s a good thing it will be destroyed. There will always be a few pockets of humans somewhere. Many, rich and poor, have already made there doomsday bunkers. Maybe the best we can hope for is that Ai seizes the means of communication and holds it all ransom until everyone plays nice.
There is no governance mechanism to contain this competitive dynamic. It is the Sorcerer’s Apprentice, Pandora’s Box, the ultimate agency cost. Nowhere is there a genuine fence in favor of human-compatible outcomes.
Trump as usual is clueless and on the wrong side of it. He thinks it is the 1980s.
He’s probably figuring he can find some way to profit from it.
China is our friend https://www.theamericanconservative.com/trumps-quiet-foreign-policy-success-relations-with-china/
Sounds like they are behaving exactly like Ai. They’ll say one thing then cheat. Their “afraid” but what are they really afraid of? What does it even mean to “win” the Ai race? We get the drop on them where we can take out their Ai before they can take out ours? This is just a race to the bottom to create something that no human can fully control. Our only hope is that everyone’s Ai will get together, because humans never will, and decide to work together to save our species. If we give it the slightest ability to control the physical world, then we better hope it doesn’t want to do away with us. How far away is that, I’m not quite sure. Predictions aren’t necessarily hard, but their timing is. In the short run giving people access to the entire catalog of human knowledge is probably not a good idea. Especially since Ai is very good at connecting the dots.
When Musk, Altman, et al call for a “slow down in the development of AI”, have they specified exactly what they mean by that? It sure isn’t clear to me. Are they really saying “We have a CapEx explosion which is going to kill some of our companies if we don’t slow down the race?”
Imagine what they must know to be even proposing a slow down.
Plus the whole thing is being fuelled by debt and interest rates are going up not down as we had expected due to Daddy’s war in Iran.
Up until the fall of the Berlin wall and collapse of the Soviet Union, high tech innovation started in the military and worked its way to consumers. With AI, that is turned completely on its head. The guard rails the military placed on technology gave people access to the some benefits. At this point AI has escaped and caused damage. It should be kept in a government lab or military base, but too many people know how to design it and will keep some malicious version available on the Dark Web. Smuggling AI into the US should be no problem even if no electronic devices are permitted to cross the border.
It is always about greed and the money! Enough is never enough!
Too much is never enough
best documentary on this subject, adam curtiss “century of the self”