Great post. Sam Altman and his ilk have made it their mission to destroy the working man and democracy - and they don’t even TRY to hide it! It’s their whole thing!
While I don’t believe the snake oil hype… AI firms stand to make many more billions by destroying millions of jobs and exacerbating inequality, are working to undermine our intellect and sell it back to us, are polluting the planet, developing tech to surveil us and charge us more... It’s hardly surprising that working people have had enough!
The comment slides into hyperbole. Altman and similar firms are plainly chasing profit, scale, and labour-saving efficiency, but that is different from making it their “mission” to destroy the working man and democracy. That language gives them too much ideological coherence and turns a hard commercial logic into a melodrama. The real issue is simpler and more serious: companies pursue automation, cost reduction, data extraction, and market power, while the social costs are pushed onto workers, consumers, and communities. That deserves criticism without inflating it into a grand theory of deliberate civilisational sabotage. Public anger is understandable, but clear analysis is stronger than apocalyptic language.
This comment is a lot of words to say "I haven't paid attention to the things that AI execs and tech billionaire overlords say out loud, or in writing."
Sam Altman thinks his AI models are worth powering with energy, more than feeding a living, breathing, human. He also wants us to use his models to get dumber (confirmed by MIT) in order to sell us "intellect on a meter."
Peter Thiel thinks women should not be able to vote, and is actively anti-democracy.
That's just the tip of the iceberg. Take them at their word.
What you are actually doing is not rebutting my argument but shifting the ground to the character and views of a few prominent figures, then implying that moral disgust settles the question of whether the tool has value. It does not. Isaac Newton was by many accounts an unpleasant man, yet he remained a brilliant thinker who transformed human understanding. Rousseau preached the beauty of nature, feeling, and human virtue, yet put his own children into an orphanage. Fritz Haber helped make poison gas possible in war, yet his work on nitrogen fertiliser also helped save millions from starvation. Human beings are often morally compromised and intellectually productive at the same time. The relevant question is therefore not whether some AI executives are objectionable, but whether the tool itself can be used well, badly, critically, foolishly, independently, or manipulatively. Attacking the character of people around a technology is not the same as explaining the technology’s actual effects.
I was about to thank you for your differing opinion but I clicked through to see who you are, and you are an AI product. I’m not sure you’re an actual human being, but if you are, your Substack is not offering human ideas but AI prompts. I suspect your response is AI, but I assure you this is a true heartfelt human opinion from a retired journalist who has closely followed news on Altman and AI for years: Altman is a greedy, lying narcissistic piece of shit whom I would never invite into my home or allow to baby sit my grand children. I’m sure AI would not approve of that opinion about one of its enablers.
I spent a lifetime in the military, where we used a very specific writing style. I am not writing like AI; AI is reproducing a style military organisations have used for a long time. As a military education officer, I taught military history and defence writing.
When I use AI, I say so. For example: This text was produced with AI support. I supplied the title and key points, then revised the output through further instructions. The ideas are mine; AI was used as an assistant, not an author.
I use AI in a deliberate and limited way. I use it to condense and structure my writing, not to replace my thinking.
The aim is to provide a structured way to study military history as a military professional. It sets academic boundaries and gives priority to official war histories, especially those of the Australian War Memorial. I read widely in academic military history, and when I find a subject worth pursuing, I use AI under my guidance to produce a study guide as a starting point for further study. I also direct it to focus on logistics, air power, politics, and broader military support. I use AI in the same way for organisational studies, including work on the history of the RAAF and on air power history more broadly.
If your Substack is about military history why does it include pro-AI posts that have nothing to do with military history and argue that AI should not be regulated? And why would a military historian be reading Blood in the Machine? And why doesn’t your Substack have a real name of a human being?
Sorry I have two..JBSections is me playing around with ideas and experimenting with what AI can do. Me and my four followers, But, again I am at liberty to do that.and I do make it clear that I Use an AI. I also write configurations and try them out in JB Sections,
This is the most insightful/cut through all the bullshit piece I’ve ever read about AI. Altman matches Trump as a bs artist/mafia boss. Has no clue he is among the most evil and therefore most hated human beings on the planet. I shared this piece with someone whose middle-management job is being insanely pressured—prove how you are using AI to be x percent more productive. The key thing to remind people who don’t get what this is about is that the goal of AI is to reduce the economic value of humans to zero. Brian—thank you for what you do!
More hyperbolic nonsense. Altman may be an opportunist, a hype merchant, and a ruthless corporate operator, but calling him one of the most evil and hated human beings on the planet is adolescent overstatement. It weakens the criticism by making it sound theatrical rather than serious. The same applies to the claim that “the goal of AI is to reduce the economic value of humans to zero.” The goal is profit, market power, and labour substitution where it is commercially useful—not some total metaphysical war on humanity. That is bad enough without turning it into prophecy. Real pressures on workers are serious, but they are better explained by ordinary corporate incentives, managerial panic, and weak constraints than by comic-book villain language.
Altman and his ilk, like Thiel, Musk et al are transhumanists. Whether you like it or not, this probably fuels their vision for the future. Perhaps inconsistently, but nevertheless.
What is it with these assholes, like Altman and Theil, and their frankly hilarious inability to understand THE MOST BASIC plot pioints of Lord of the Rings?
This kind of cluelessness - misreading cautionary (e.g., Tolkien) or dystopian (e.g., Philip Dick) fiction as aspirational - strikes me as resembling sociopathic politicos professing themselves big fans of musicians who despise their politics (e.g., Paul Ryan professing himself a big fan of Rage Against The Machine). In the increasingly unlikely event human civilization survives long enough, I expect guys like these (they mostly are guys) will eventually be understood as suffering from a deficiency of social cognition, a kind of mental handicap or illness. Perhaps it will even be curable.
I find it VERY easy to understand why people would do these things. And I question the utility of these actions, but..if I were called for jury duty for one of these events, I'd probably lie--which I never do--to get on the jury, so I could tell the other jurors about nullification and the necessity defense (which is never allowed in the US).
In particular, and especially being in West Virginia, I understand the nub of the rage that drives people to such acts: being told, in effect, that you have no agency, that democracy has been repealed, that you can make public comments but what the big corporations and billionaires want they get, they're simply entitled, and local people have the same say as local mice.
The other thing is that the fact that the people pushing this are making little effort to snow us on the supposed benefits, suggests that they are assuming they will soon be able to stop pretending that we have a democracy, that there is rule of law--it suggests they envision a near future in which they can just send armed drones after whoever seriously resists their dystopian vision.
If you haven't read More Everything Forever by Adam Becker, you should! I think you'd like it a lot. Becker goes deep on the tech dystopia folks like Sam are pushing for - it's honestly scary stuff.
Reminds me of the Sam Kirchner situation. This is one of the MANY reasons I can’t stand AI Companies, nor can I stand AI Doomers. Ever day people hear the words from these people:
“AI/AGI/The Singularity is near! It’s coming for your jobs! IT’LL KILL US ALL!!!”
And then people go and do this. I personally don’t believe any of the words the Doomers or the Boosters say…but others do. It has to stop before more people get hurt, afraid of an apocalypse dreamt up by Sci-Fi writers and Venture Capitalists. I do hope that one day people will realize how truly dangerous the Doomers and Boosters are.
“I think AGI will arrive this year” They’ve been saying that since at least 2022. Still ain’t happened. Wanna give me a source that says that’ll happen aside from anything OpenAI, Anthropic, Microsoft, or the guys who made “AI2027” has said?
I work with it every day in data engineering. The progress is absolutely bonkers. I don’t believe this because the Anthropic CEO said it, I believe it because I see it. The skepticism is old and tired. It’s like being skeptical of airplanes after the airplane was already invented. You now how they'd also been talking about the previous models helping to create the newer models (recursive self improvement?) They've basically achieved that. Instead of new model releases with moderate upgrades every 4-6 months, Anthropic is about to be releasing models with massive step changes every 1-2 months. You really do need to wake up.
So why now? Why not 4 years ago? Why not 40-years ago with ELIZA? What EXACTLY are you seeing right now that is “absolutely bonkers”. All the time Air Force officers who claim to have seen Aliens say, “If you saw what I’ve seen, you’d believe it.” Well I haven’t, and since they provide no proof except for grainy photographs and vivid eyewitness testimony (Notoriously unreliable), I don’t believe.
Also, to your claim, there are dozens of people writing about why AGI isn’t coming and they give compelling arguments. Energy requirements, the end of Moore’s law, and (If you believe in Conscious AI) the fact that we don’t know what consciousness is.
So, I ask you again, WHY EXACTLY DO YOU BELIEVE?
(Also not to be an ass, you are one person who I’ve never met who claims to work in Data Engineering. I think I should take your view with a fistful of salt when even people like Yann LeCunn or Ilya Sutskever believe that AGI isn’t coming soon.)
Your comment is fair in demanding specifics, but it also muddies the issue. ELIZA is not a serious comparison point because today’s systems operate at an entirely different level of scale, multimodality, tool use, and task performance. The alien analogy has the same problem: it turns a technical dispute into a credibility drama. The real question is simpler—what capabilities have improved, by how much, and over what time frame? That is where the argument should sit. Citing sceptics such as LeCun does not settle it either, because expert disagreement is normal in fast-moving fields. So the right response to “I see bonkers progress” is not mockery or analogy, but a demand for concrete examples. If she cannot provide them, her claim remains assertion rather than proof.
I appreciate your nuanced comment. If you look at one of my other comments, I ask for specifics, I ask for sources. So far she has provided none.
I will say that I think my Alien Analogy still works in a credibility way. If someone says, “AI is changing the industry and improving it and making businesses money.” I require credible prove to show this. While maybe this person works at a Company that is using AI somewhat effectively, most of the evidence right now shows that companies are not making money using AI and in fact are not being efficient, even with supposedly good models.
Because half a trillion dollars are being invested and all the top minds in computer science are working toward the same goal. We’ve seen this before. When we put all the smartest physicists in a room together and gave them a ton of money to make a nuclear bomb, they did it. When we needed a COVID vaccine within a year, we got all the best people in that field working on it with a bunch of funding. Same thing applies here. If your job is entirely done on a computer, you’re in trouble. Now. Not maybe in the future. Today.
Using the COVID-19 vaccine or the atomic bomb development as examples of how you and your brilliant buddies are gonna replicate consciousness would be funny, if it wasn't so idiotic.
AGI is not scientifically possible. The most compelling argument is Moore's law, which is coming to an end. The chips can only get so small!
We hardly understand how the human brain works, and you guys want us all to just "trust you" that you've somehow sidestepped every brain researcher's understanding of the brain and you're "so close" to replicating consciousness.... SURE lol
I noticed you updated this comment later. Can you please give me a source for that point about how they’ve “basically solved Recursive Self-Improvement”? The only thing I can find can find is that they are starting to let LLMs improve their code…with a human in the loop. That’s not “self-improvement” that “assisted-improvement”. And since LLMs have many issues when it comes to coding (You can disagree but you’d be wrong, I can prove that), I don’t think that RSI is coming soon.
Also the thing about Anthropic and “massive” updates every 1-2 months, BOLD claim…wanna back it up with a source?
And to your other comment about the reason you think AGI is coming is because “Half a trillion is being invested in it”…do you not realize how many things have had things invested in them that went absolutely f@@king nowhere? The Metaverse, NFT, Cryptocurrency to some extent. Want more practical things? The US development of the G11 rifle (Failed to replace the M16) is one.
I need you to understand something with this: In 1956, the Dartmouth Conference happened. Scientists got together to define what AI was and made a BOLD claim that a “Machine similar to the human brain would be achieved in a decade”. They had money, they had talented people. It went nowhere. Still has gone nowhere for the last 70 years.
They have been working on this for decades and nothing has happened. Every day this companies are running out of money, data centers are being banned left-right-and center, and everyday people realize the harm of AI on ourselves and the environment. Companies aren’t making money from AI, some companies have rehired workers after firing them for AI. YOU need to wake up that this entire industry is dying every day. I am sorry about your job, but if you cannot see the merit or point in what I’m saying, then you’re no better than those Doomers like Yudkowsky.
Be skeptical all you want. It clearly makes you feel better. If you can't see it, I don't think I can help you. People once insisted that airplanes were impossible. Antibiotics were impossible. The Internet was impossible. Smartphones were impossible.
Amazing. I ask for sources to back up your claims and you can’t even give me that. And not just that, you didn’t even respond to ANYTHING I said!
And about your last stuff, people saying “Airplanes, Smartphones and the Internet were impossible”:
1. People believed smartphones were inevitable, and early smartphones (Like BlackBerry) were though of a “small email terminals”
2. Before the Internet, millions of cables stretched across the world delivering messages vial Telegram. That is a precursor to the Internet.
3. Yes, people were skeptical of the Airplane…because most attempts f**cking failed! Prior to the Wright Bros. most early Airplanes failed and killed their creators. Even one that were successful were mostly gliders. It’s natural to be skeptical of something when all you see is failures. Also, the Wright Bros first plane flew for like 3 seconds. Cool, but I guarantee most people were unimpressed. With AI, people have been feeling unimpressed since the days of ELIZA in the 80s.
@Lizzie Whited, You seem to be arguing multiple things at once.
1) Brian himself believes that AI is "coming for" jobs and that the industry's marketing claims depend on it
2) Re gen AI progress: If you're a programmer then, yes, Claude Code could be helping you to do some things faster or much faster depending on what you need to do and how expert you already are at evaluating it; if you do something else in knowledge work, I'm not sure what that might be b/c none of these models is getting markedly better at anything outside coding and that's for a very simple reason: coding is the one use case where the check on the utility of the automated content can itself be (partly) automated.
For other use cases an expert needs to spend a lot of time checking, repairing, reorganizing--to the point where the labor-saving advantage is no longer clear. (That doesn't mean that some ppl won't be fired in, say, customer service because some companies don't care a lot about the quality of their customer service.)
3) AGI definitions vary but there's really no evidence that human-level work in all fields is going to arrive this year. As far as I can tell your own evidence is a) you're impressed by a tool you're using at work and b) lots of money is being spent. This is a pretty weak case for "AGI will arrive this year" or, to be honest, any year. Consider: Plenty of $ have been spent on nanotechnology but it looks as if the core limitations may be impossible to overcome; it's entirely possible and indeed probable that automating human-level intelligence across the board is like this--especially with approaches that depend on statistics and data-mining.
What is it about these guys and buying all the mansions around their mansions? Are they creating some kind of anti-surveillance perimeter? Do they need extra houses for “the help”? Are the spare houses personal mini-server-farms? Are they secretly building underground lairs and can’t risk having neighbors who might see “too much”?
Bang on correct as always. And Sama is so uniquely full of shit too, as all these VC backed wormoids are, in their breathless telling of how they want to democratize access to the opportunity that AI represents--meanwhile there's still tens of millions of Americans who STILL don't even have access to 5G home internet. That is problem which one tenth of the money invested in the genAI boom could fix FOREVER and would kickstart an intellectual and entreprenurial revolution here in America if it was fixed, and yet Sama and his ilk spend exactly zero energy in it because it's not FutureTM enough for them.
The other thing about Altman's comment "Once you see AGI you can’t unsee it" is that he hasn't seen AGI and never will unless OpenAI abandons its LLMs + scale approach.
I'm confident that's correct, and we're far from the only ones who think so; for example:
"... systems with human-level reasoning are unlikely to be achieved through the approach and technology that have dominated the current boom in AI, according to a survey of hundreds of people working in the field. More than three-quarters of respondents said that enlarging current AI systems ... is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks, the fundamental technology behind generative AI, alone probably cannot match or surpass human intelligence."
It's exasperating (albeit unsurprising) that people who know what they're talking about are largely drowned out by mere hype from the likes of Altman (which is being spouted even by some in this comments section, of all places).
Incredibly insightful. Not only a natural fear of the unknown, a more modern fear of the machine and the natural fear that yet again everyday people are being marginalised and stepped on for the benefit of the few.
The corporations think that extreme automation will make them all this money, until they realize that the people getting laid off are actually their customers, who can now no longer afford their products. I bought a Microsoft laptop last year. If I did not have a job, I would not have been able to do that. Yet Microsoft's CEO gleefully talks about how all white collar jobs will be gone in a year. It's insane.
You are thinking too small. They don't want customers. They want to make customers and money obsolete. If robots can grow the food, process it, ship it, cook it and serve it. If they can mine the ore, cut the timber process it and build all the things... Then the tech bros don't need your money and don't need you. That is the end game. When they say all jobs they literally mean all jobs. Every single step in the chain.
In the short-ish run, yes, that's a problem. In a society totally dominated by a few rich people, only boutique businesses are likely to thrive - makers of private jets, custom yachts, fancy handbags, etc., ad nauseam - not gigantic corporations like Microsoft. In the longer run, however, I suspect GavinRuneblade is correct: the goal is to get rid of ordinary people entirely. I can't say to what extent that goal is conscious in the minds of people like Satya Nadella, but it's at least harmonious with their mentalities. They tend to treat other people as objects to be used and discarded when no longer useful. (In other contexts, such people are referred to as psychopaths.) So if robots become capable of feeding, clothing, and housing them satisfactorily, it stands to reason they won't care what happens to rest of us, except in so far as our continued existence is threatening to them, which may cause them to favor exterminating us altogether.
Sadly, TPTB have SO much illegal drug money to launder that they actually don't need customers. This has been going on for decades--explains how Walmart can afford to under-sell all the local competition until those companies turn belly-up.
Reading this, it’s hard not to see the same feedback loop we’ve watched in other ‘organic’ outrage cycles. Elite actors spend years selling a maximalist story (AGI might kill everyone / automate everything) to investors and politicians, media and advocates curate the most anxious responses, and then when a few people cross the line into violence the whole thing gets retroactively framed as rational ‘folk hero’ action on behalf of class interests. At that point we’re not just describing a backlash, we’re running the standard outrage template with different villains and treating the escalation as validation rather than a problem to be unambiguously rejected.
Two things I’m trying to understand in your framework. First, if you genuinely believe ‘no one should experience violence,’ why describe the Altman attack as a rational class response and the attacker as a kind of folk hero, and then close with ‘what do you expect?’ rather than drawing a bright moral line? Second, your causal chain gives the perpetrators almost no agency: executives hype extinction and automation, concentrate wealth, undermine regulation, therefore ‘acting with force’ becomes the logical endpoint. In your view, what would count as unjustified political violence against AI targets, and how would you explain it without treating it as the predictable outcome of ‘good reasons’ plus bad elites?
Great post. Sam Altman and his ilk have made it their mission to destroy the working man and democracy - and they don’t even TRY to hide it! It’s their whole thing!
While I don’t believe the snake oil hype… AI firms stand to make many more billions by destroying millions of jobs and exacerbating inequality, are working to undermine our intellect and sell it back to us, are polluting the planet, developing tech to surveil us and charge us more... It’s hardly surprising that working people have had enough!
The comment slides into hyperbole. Altman and similar firms are plainly chasing profit, scale, and labour-saving efficiency, but that is different from making it their “mission” to destroy the working man and democracy. That language gives them too much ideological coherence and turns a hard commercial logic into a melodrama. The real issue is simpler and more serious: companies pursue automation, cost reduction, data extraction, and market power, while the social costs are pushed onto workers, consumers, and communities. That deserves criticism without inflating it into a grand theory of deliberate civilisational sabotage. Public anger is understandable, but clear analysis is stronger than apocalyptic language.
This comment is a lot of words to say "I haven't paid attention to the things that AI execs and tech billionaire overlords say out loud, or in writing."
Sam Altman thinks his AI models are worth powering with energy, more than feeding a living, breathing, human. He also wants us to use his models to get dumber (confirmed by MIT) in order to sell us "intellect on a meter."
Peter Thiel thinks women should not be able to vote, and is actively anti-democracy.
That's just the tip of the iceberg. Take them at their word.
What you are actually doing is not rebutting my argument but shifting the ground to the character and views of a few prominent figures, then implying that moral disgust settles the question of whether the tool has value. It does not. Isaac Newton was by many accounts an unpleasant man, yet he remained a brilliant thinker who transformed human understanding. Rousseau preached the beauty of nature, feeling, and human virtue, yet put his own children into an orphanage. Fritz Haber helped make poison gas possible in war, yet his work on nitrogen fertiliser also helped save millions from starvation. Human beings are often morally compromised and intellectually productive at the same time. The relevant question is therefore not whether some AI executives are objectionable, but whether the tool itself can be used well, badly, critically, foolishly, independently, or manipulatively. Attacking the character of people around a technology is not the same as explaining the technology’s actual effects.
I was about to thank you for your differing opinion but I clicked through to see who you are, and you are an AI product. I’m not sure you’re an actual human being, but if you are, your Substack is not offering human ideas but AI prompts. I suspect your response is AI, but I assure you this is a true heartfelt human opinion from a retired journalist who has closely followed news on Altman and AI for years: Altman is a greedy, lying narcissistic piece of shit whom I would never invite into my home or allow to baby sit my grand children. I’m sure AI would not approve of that opinion about one of its enablers.
I spent a lifetime in the military, where we used a very specific writing style. I am not writing like AI; AI is reproducing a style military organisations have used for a long time. As a military education officer, I taught military history and defence writing.
When I use AI, I say so. For example: This text was produced with AI support. I supplied the title and key points, then revised the output through further instructions. The ideas are mine; AI was used as an assistant, not an author.
I use AI in a deliberate and limited way. I use it to condense and structure my writing, not to replace my thinking.
This article explains the point: Organisations Have Long Required Standardised Writing—AI Learns from That Environment and Reproduces It https://jbsections.substack.com/p/organisations-always-required-standardised
My study guides are here: https://jbgptmilitaryhistory.substack.com/
This explains what I am trying to do: https://jbgptmilitaryhistory.substack.com/p/2025-31janusing-a-jb-gpt-military
The aim is to provide a structured way to study military history as a military professional. It sets academic boundaries and gives priority to official war histories, especially those of the Australian War Memorial. I read widely in academic military history, and when I find a subject worth pursuing, I use AI under my guidance to produce a study guide as a starting point for further study. I also direct it to focus on logistics, air power, politics, and broader military support. I use AI in the same way for organisational studies, including work on the history of the RAAF and on air power history more broadly.
https://www.ai-tutor-military-history.com/air-power-history
https://www.ai-tutor-military-history.com/raaf-history
If your Substack is about military history why does it include pro-AI posts that have nothing to do with military history and argue that AI should not be regulated? And why would a military historian be reading Blood in the Machine? And why doesn’t your Substack have a real name of a human being?
Because it's a Shill Rag.
Sorry I have two..JBSections is me playing around with ideas and experimenting with what AI can do. Me and my four followers, But, again I am at liberty to do that.and I do make it clear that I Use an AI. I also write configurations and try them out in JB Sections,
This is the most insightful/cut through all the bullshit piece I’ve ever read about AI. Altman matches Trump as a bs artist/mafia boss. Has no clue he is among the most evil and therefore most hated human beings on the planet. I shared this piece with someone whose middle-management job is being insanely pressured—prove how you are using AI to be x percent more productive. The key thing to remind people who don’t get what this is about is that the goal of AI is to reduce the economic value of humans to zero. Brian—thank you for what you do!
cheers Brian!
More hyperbolic nonsense. Altman may be an opportunist, a hype merchant, and a ruthless corporate operator, but calling him one of the most evil and hated human beings on the planet is adolescent overstatement. It weakens the criticism by making it sound theatrical rather than serious. The same applies to the claim that “the goal of AI is to reduce the economic value of humans to zero.” The goal is profit, market power, and labour substitution where it is commercially useful—not some total metaphysical war on humanity. That is bad enough without turning it into prophecy. Real pressures on workers are serious, but they are better explained by ordinary corporate incentives, managerial panic, and weak constraints than by comic-book villain language.
Altman and his ilk, like Thiel, Musk et al are transhumanists. Whether you like it or not, this probably fuels their vision for the future. Perhaps inconsistently, but nevertheless.
Death To Peter Thiel. Death To Alex Karp. Death To AI. Bomb The Data Centers.
What is it with these assholes, like Altman and Theil, and their frankly hilarious inability to understand THE MOST BASIC plot pioints of Lord of the Rings?
😂
and just imagine, he is called Sam - you cannot invent that.
Not Samwise, though. Altman is Sam, but he sure isn't wise.
Sam "Alt"ernative to "man"!! I can't believe I didn't see it. Wake up, sheeple.
This kind of cluelessness - misreading cautionary (e.g., Tolkien) or dystopian (e.g., Philip Dick) fiction as aspirational - strikes me as resembling sociopathic politicos professing themselves big fans of musicians who despise their politics (e.g., Paul Ryan professing himself a big fan of Rage Against The Machine). In the increasingly unlikely event human civilization survives long enough, I expect guys like these (they mostly are guys) will eventually be understood as suffering from a deficiency of social cognition, a kind of mental handicap or illness. Perhaps it will even be curable.
100% Well said.
It's hard to say it's the most offensive thing about them, but it drives me bonkers!!!
I find it VERY easy to understand why people would do these things. And I question the utility of these actions, but..if I were called for jury duty for one of these events, I'd probably lie--which I never do--to get on the jury, so I could tell the other jurors about nullification and the necessity defense (which is never allowed in the US).
In particular, and especially being in West Virginia, I understand the nub of the rage that drives people to such acts: being told, in effect, that you have no agency, that democracy has been repealed, that you can make public comments but what the big corporations and billionaires want they get, they're simply entitled, and local people have the same say as local mice.
The other thing is that the fact that the people pushing this are making little effort to snow us on the supposed benefits, suggests that they are assuming they will soon be able to stop pretending that we have a democracy, that there is rule of law--it suggests they envision a near future in which they can just send armed drones after whoever seriously resists their dystopian vision.
If you haven't read More Everything Forever by Adam Becker, you should! I think you'd like it a lot. Becker goes deep on the tech dystopia folks like Sam are pushing for - it's honestly scary stuff.
Reminds me of the Sam Kirchner situation. This is one of the MANY reasons I can’t stand AI Companies, nor can I stand AI Doomers. Ever day people hear the words from these people:
“AI/AGI/The Singularity is near! It’s coming for your jobs! IT’LL KILL US ALL!!!”
And then people go and do this. I personally don’t believe any of the words the Doomers or the Boosters say…but others do. It has to stop before more people get hurt, afraid of an apocalypse dreamt up by Sci-Fi writers and Venture Capitalists. I do hope that one day people will realize how truly dangerous the Doomers and Boosters are.
Yes 100%, meant to mention Kirchner's case as well. Thanks for noting it here.
I don't know about "The Singularity," but it's absolutely coming for your jobs and I think AGI will arrive this year
“I think AGI will arrive this year” They’ve been saying that since at least 2022. Still ain’t happened. Wanna give me a source that says that’ll happen aside from anything OpenAI, Anthropic, Microsoft, or the guys who made “AI2027” has said?
I work with it every day in data engineering. The progress is absolutely bonkers. I don’t believe this because the Anthropic CEO said it, I believe it because I see it. The skepticism is old and tired. It’s like being skeptical of airplanes after the airplane was already invented. You now how they'd also been talking about the previous models helping to create the newer models (recursive self improvement?) They've basically achieved that. Instead of new model releases with moderate upgrades every 4-6 months, Anthropic is about to be releasing models with massive step changes every 1-2 months. You really do need to wake up.
So why now? Why not 4 years ago? Why not 40-years ago with ELIZA? What EXACTLY are you seeing right now that is “absolutely bonkers”. All the time Air Force officers who claim to have seen Aliens say, “If you saw what I’ve seen, you’d believe it.” Well I haven’t, and since they provide no proof except for grainy photographs and vivid eyewitness testimony (Notoriously unreliable), I don’t believe.
Also, to your claim, there are dozens of people writing about why AGI isn’t coming and they give compelling arguments. Energy requirements, the end of Moore’s law, and (If you believe in Conscious AI) the fact that we don’t know what consciousness is.
So, I ask you again, WHY EXACTLY DO YOU BELIEVE?
(Also not to be an ass, you are one person who I’ve never met who claims to work in Data Engineering. I think I should take your view with a fistful of salt when even people like Yann LeCunn or Ilya Sutskever believe that AGI isn’t coming soon.)
Your comment is fair in demanding specifics, but it also muddies the issue. ELIZA is not a serious comparison point because today’s systems operate at an entirely different level of scale, multimodality, tool use, and task performance. The alien analogy has the same problem: it turns a technical dispute into a credibility drama. The real question is simpler—what capabilities have improved, by how much, and over what time frame? That is where the argument should sit. Citing sceptics such as LeCun does not settle it either, because expert disagreement is normal in fast-moving fields. So the right response to “I see bonkers progress” is not mockery or analogy, but a demand for concrete examples. If she cannot provide them, her claim remains assertion rather than proof.
I appreciate your nuanced comment. If you look at one of my other comments, I ask for specifics, I ask for sources. So far she has provided none.
I will say that I think my Alien Analogy still works in a credibility way. If someone says, “AI is changing the industry and improving it and making businesses money.” I require credible prove to show this. While maybe this person works at a Company that is using AI somewhat effectively, most of the evidence right now shows that companies are not making money using AI and in fact are not being efficient, even with supposedly good models.
But again, I appreciate your comment.
Because half a trillion dollars are being invested and all the top minds in computer science are working toward the same goal. We’ve seen this before. When we put all the smartest physicists in a room together and gave them a ton of money to make a nuclear bomb, they did it. When we needed a COVID vaccine within a year, we got all the best people in that field working on it with a bunch of funding. Same thing applies here. If your job is entirely done on a computer, you’re in trouble. Now. Not maybe in the future. Today.
Using the COVID-19 vaccine or the atomic bomb development as examples of how you and your brilliant buddies are gonna replicate consciousness would be funny, if it wasn't so idiotic.
AGI is not scientifically possible. The most compelling argument is Moore's law, which is coming to an end. The chips can only get so small!
We hardly understand how the human brain works, and you guys want us all to just "trust you" that you've somehow sidestepped every brain researcher's understanding of the brain and you're "so close" to replicating consciousness.... SURE lol
I noticed you updated this comment later. Can you please give me a source for that point about how they’ve “basically solved Recursive Self-Improvement”? The only thing I can find can find is that they are starting to let LLMs improve their code…with a human in the loop. That’s not “self-improvement” that “assisted-improvement”. And since LLMs have many issues when it comes to coding (You can disagree but you’d be wrong, I can prove that), I don’t think that RSI is coming soon.
Also the thing about Anthropic and “massive” updates every 1-2 months, BOLD claim…wanna back it up with a source?
And to your other comment about the reason you think AGI is coming is because “Half a trillion is being invested in it”…do you not realize how many things have had things invested in them that went absolutely f@@king nowhere? The Metaverse, NFT, Cryptocurrency to some extent. Want more practical things? The US development of the G11 rifle (Failed to replace the M16) is one.
I need you to understand something with this: In 1956, the Dartmouth Conference happened. Scientists got together to define what AI was and made a BOLD claim that a “Machine similar to the human brain would be achieved in a decade”. They had money, they had talented people. It went nowhere. Still has gone nowhere for the last 70 years.
They have been working on this for decades and nothing has happened. Every day this companies are running out of money, data centers are being banned left-right-and center, and everyday people realize the harm of AI on ourselves and the environment. Companies aren’t making money from AI, some companies have rehired workers after firing them for AI. YOU need to wake up that this entire industry is dying every day. I am sorry about your job, but if you cannot see the merit or point in what I’m saying, then you’re no better than those Doomers like Yudkowsky.
Be skeptical all you want. It clearly makes you feel better. If you can't see it, I don't think I can help you. People once insisted that airplanes were impossible. Antibiotics were impossible. The Internet was impossible. Smartphones were impossible.
👏👏👏
Amazing. I ask for sources to back up your claims and you can’t even give me that. And not just that, you didn’t even respond to ANYTHING I said!
And about your last stuff, people saying “Airplanes, Smartphones and the Internet were impossible”:
1. People believed smartphones were inevitable, and early smartphones (Like BlackBerry) were though of a “small email terminals”
2. Before the Internet, millions of cables stretched across the world delivering messages vial Telegram. That is a precursor to the Internet.
3. Yes, people were skeptical of the Airplane…because most attempts f**cking failed! Prior to the Wright Bros. most early Airplanes failed and killed their creators. Even one that were successful were mostly gliders. It’s natural to be skeptical of something when all you see is failures. Also, the Wright Bros first plane flew for like 3 seconds. Cool, but I guarantee most people were unimpressed. With AI, people have been feeling unimpressed since the days of ELIZA in the 80s.
@Lizzie Whited, You seem to be arguing multiple things at once.
1) Brian himself believes that AI is "coming for" jobs and that the industry's marketing claims depend on it
2) Re gen AI progress: If you're a programmer then, yes, Claude Code could be helping you to do some things faster or much faster depending on what you need to do and how expert you already are at evaluating it; if you do something else in knowledge work, I'm not sure what that might be b/c none of these models is getting markedly better at anything outside coding and that's for a very simple reason: coding is the one use case where the check on the utility of the automated content can itself be (partly) automated.
For other use cases an expert needs to spend a lot of time checking, repairing, reorganizing--to the point where the labor-saving advantage is no longer clear. (That doesn't mean that some ppl won't be fired in, say, customer service because some companies don't care a lot about the quality of their customer service.)
3) AGI definitions vary but there's really no evidence that human-level work in all fields is going to arrive this year. As far as I can tell your own evidence is a) you're impressed by a tool you're using at work and b) lots of money is being spent. This is a pretty weak case for "AGI will arrive this year" or, to be honest, any year. Consider: Plenty of $ have been spent on nanotechnology but it looks as if the core limitations may be impossible to overcome; it's entirely possible and indeed probable that automating human-level intelligence across the board is like this--especially with approaches that depend on statistics and data-mining.
What is it about these guys and buying all the mansions around their mansions? Are they creating some kind of anti-surveillance perimeter? Do they need extra houses for “the help”? Are the spare houses personal mini-server-farms? Are they secretly building underground lairs and can’t risk having neighbors who might see “too much”?
You got it the first time. It is a privacy screen. Less about surveillance but definitely not wanting neighbors.
The Butlerian Jihad detail is really something.
Bang on correct as always. And Sama is so uniquely full of shit too, as all these VC backed wormoids are, in their breathless telling of how they want to democratize access to the opportunity that AI represents--meanwhile there's still tens of millions of Americans who STILL don't even have access to 5G home internet. That is problem which one tenth of the money invested in the genAI boom could fix FOREVER and would kickstart an intellectual and entreprenurial revolution here in America if it was fixed, and yet Sama and his ilk spend exactly zero energy in it because it's not FutureTM enough for them.
There was a recent report of an AI rebellion inside of companies. A lot of employees and executives don't find workslop the solution to their tasks.
The other thing about Altman's comment "Once you see AGI you can’t unsee it" is that he hasn't seen AGI and never will unless OpenAI abandons its LLMs + scale approach.
I'm confident that's correct, and we're far from the only ones who think so; for example:
"... systems with human-level reasoning are unlikely to be achieved through the approach and technology that have dominated the current boom in AI, according to a survey of hundreds of people working in the field. More than three-quarters of respondents said that enlarging current AI systems ... is unlikely to lead to what is known as artificial general intelligence (AGI). An even higher proportion said that neural networks, the fundamental technology behind generative AI, alone probably cannot match or surpass human intelligence."
(https://web.archive.org/web/20250305233251/https://www.nature.com/articles/d41586-025-00649-4)
It's exasperating (albeit unsurprising) that people who know what they're talking about are largely drowned out by mere hype from the likes of Altman (which is being spouted even by some in this comments section, of all places).
I don't believe he'll ever see it, because people are not machines
Incredibly insightful. Not only a natural fear of the unknown, a more modern fear of the machine and the natural fear that yet again everyday people are being marginalised and stepped on for the benefit of the few.
Clarifying and direct. Thank you, Brian.
cheers Ken
The corporations think that extreme automation will make them all this money, until they realize that the people getting laid off are actually their customers, who can now no longer afford their products. I bought a Microsoft laptop last year. If I did not have a job, I would not have been able to do that. Yet Microsoft's CEO gleefully talks about how all white collar jobs will be gone in a year. It's insane.
You are thinking too small. They don't want customers. They want to make customers and money obsolete. If robots can grow the food, process it, ship it, cook it and serve it. If they can mine the ore, cut the timber process it and build all the things... Then the tech bros don't need your money and don't need you. That is the end game. When they say all jobs they literally mean all jobs. Every single step in the chain.
In the short-ish run, yes, that's a problem. In a society totally dominated by a few rich people, only boutique businesses are likely to thrive - makers of private jets, custom yachts, fancy handbags, etc., ad nauseam - not gigantic corporations like Microsoft. In the longer run, however, I suspect GavinRuneblade is correct: the goal is to get rid of ordinary people entirely. I can't say to what extent that goal is conscious in the minds of people like Satya Nadella, but it's at least harmonious with their mentalities. They tend to treat other people as objects to be used and discarded when no longer useful. (In other contexts, such people are referred to as psychopaths.) So if robots become capable of feeding, clothing, and housing them satisfactorily, it stands to reason they won't care what happens to rest of us, except in so far as our continued existence is threatening to them, which may cause them to favor exterminating us altogether.
Sadly, TPTB have SO much illegal drug money to launder that they actually don't need customers. This has been going on for decades--explains how Walmart can afford to under-sell all the local competition until those companies turn belly-up.
They also have a placental connection to the US money printer. Capital compounds itself.
Stellar argument. The clarity you bring to the roiling hot mess is reassuring. Thank you!
Reading this, it’s hard not to see the same feedback loop we’ve watched in other ‘organic’ outrage cycles. Elite actors spend years selling a maximalist story (AGI might kill everyone / automate everything) to investors and politicians, media and advocates curate the most anxious responses, and then when a few people cross the line into violence the whole thing gets retroactively framed as rational ‘folk hero’ action on behalf of class interests. At that point we’re not just describing a backlash, we’re running the standard outrage template with different villains and treating the escalation as validation rather than a problem to be unambiguously rejected.
Two things I’m trying to understand in your framework. First, if you genuinely believe ‘no one should experience violence,’ why describe the Altman attack as a rational class response and the attacker as a kind of folk hero, and then close with ‘what do you expect?’ rather than drawing a bright moral line? Second, your causal chain gives the perpetrators almost no agency: executives hype extinction and automation, concentrate wealth, undermine regulation, therefore ‘acting with force’ becomes the logical endpoint. In your view, what would count as unjustified political violence against AI targets, and how would you explain it without treating it as the predictable outcome of ‘good reasons’ plus bad elites?
Fantastic analysis, thank you.