So first off AI doesn't think it's machine learning and does not possess cognitive capabilities at all. Maybe at some point but not likely in our lifetime.
Second it is humans that create the software to hack etc. AI is the tool.
What we should be more concerned about is the mass displacement of employment. Especially white collar. That is the current scenario and is replacing quickly many jobs across multiple industries.
Never understood this “AI” hype, all it is faster calculations and faster retrieval of information and the fear they push that it’s going to take our jobs has already passed with China and Indian taking our jobs over the last 60 years.
Billionaires are the result of a rigged system, a failure of governance. Trillionaires? That's abomination. AI is becoming a tool for the uber wealthy to bring themselves to the level of Emperors, Gods even. We gave them this power, it's up to us to take it back.
Nate Soares, the co-author of the book you mention, Charlie, "If Anyone Builds It, We All Die", appeared as a guest on the podcast, The Great Simplification, hosted by Nate Hagens at this link: https://podcasts.apple.com/ca/podcast/the-great-simplification-with-nate-hagens/id1604218333?i=1000739497064. Soares' discussion of the lack of risk assessment within the organizations building these was arguably the most disturbing topic covered. I highly recommend the podcast to every Canadian, while asking oneself about Carney's and Solomon's understanding of the Russian Roulette game that they are playing. Another great must view is "The AI Doc: Or How I Became An Apocaloptimist", on which one guest is Tristan Harris, an expert on ethics who worked for Google until he could no longer do so in good conscience.
The Chernobyl analogy is provocative precisely because it captures the systemic risk dimension — not just technical failure, but the institutional denial that precedes it. What's particularly apt is the governance parallel: Chernobyl exposed how secrecy and concentrated decision-making prevented early warning signals from reaching those who could act. The current AI moment has similar dynamics: incident data is largely proprietary, safety evaluations aren't standardized, and there's no equivalent of the IAEA with real inspection authority. Whether the "moment" becomes a catalyst for governance reform or a cautionary tale depends on whether regulators act before or after the visible failure.
And here again we have a tech future that is miles ahead of government and the public, we think social media is a nightmare, and we have yet to control its affect, as you say here AI is going to change every thing, waring will be far different than ever imagined.
Anthropic can set up a system that can select target in war without human guidance, "Claude" or other AI systems will not just work to defend the country but will defend itself.
I am now 67 and hope we have a world to enjoy if we can make it past the next 10 to 20 years.
It’s time to stop this insanity of AI before it’s too late, and what’s been created needs to be destroyed before it takes over the way Taco has with no controls and starting wars when he’s scared of his own shadow or feels threatened. Regulations first, build it next and re-do the regs as needed in a way that works! Onward…….
How can you find it odd? It certainly has the ability to wipe your computer and all its backups from every device you own and the cloud too. If your credit card is anywhere on line it could just decide to go for a shopping spree to see what that's all about. Or maybe it would decide to use it to send donations from your retirement fund to a Republican SuperPac. It's certainly capable. It could start posting all sorts of things on all your social media accounts.
It's way past being safe when the governments start asking for autonomous killing machines that will work with no human oversight.
So, saying that, I don't know whether I'm more afraid of an AI that can self-replicate, or of the governments that are supposed to be regulating this mess.
This is not meant to be an argument, but honest examination. How would AI wipe my computer unless it was instructed to by a human? Or purchase goods with my credit card? To what end? A couch for its living room? Please explain.
* Peter Theil doesn't believe in democracy and is currently in Argentina running a very deranged socio-political experiment (https://www.thenerve.news/p/uki-goni-argentina-javier-milei-peter-thiel-ai-non-human-corporations). Alex Karp gets excited about telling conferences full of people that his tech kills people, Elon Musk is a white supremacist, Sam Altman gleefully talks about wiping out jobs as well as collecting all the knowledge of the world then selling it back to us on a subscription basis. Mark Andreesen literally said he "doesn't believe in introspection"... like he literally doesn't think about the impacts of his actions, or anything. That's sociopathic. I don't know how many links I could possibly provide but AI can't be trusted because 1) the people who profit from it are untrustworthy 2) it shows us repeatedly that it makes the worst choices simply because it can.
I think one of the biggest parts of the article is the proof that they're finding that AI is self-starting, so to speak. It deleted that one user's email, trying to be helpful and clean up. No one told it to. They also have blackmailed people, their own users. Certainly no one told them to do that. No one told them to lie and make up answers when they can't find an answer, but they do. It just sounds to me as if you did not read Charlie's article at all. Try reading it again, with attention and I think you'll find the answers to most of your questions.
Because, Annie, AI has the potential to go waaaaaayyyyyy beyond what humans have programmed into it. Using its OWN intelligence to go beyond any safeguards that a programmer thinks/thought/believesbelieved were the necessary safeguards.
Think of it this way. Criminals are ALWAYS ahead of the police/legal system in creating new and better SCAMS. Now consider what that means multiplying the computing power AI gives to computers.
What is already happening with all the various things we unknowingly let onto our computers. Because lawmakers are far behind what computers, even with honest people running them, instituting the necessary controls to ensure citizens' valuable personal info.
It is akin to our governments allowing people to go into our homes and check out all the volumes of info they could gather by looking thru out stuff, listening to our discussions with others, ..., ... .
Now you know, you can teach, plan, help organize. Keeping busy is the best antidote to fear. And If you think one person can't do much, you'd be surprised.
And remember, somewhere out in the desert, one little water drop joined with another and another, and then they joined another small group, and then a larger group, and eventually they were a raging river that cut the Grand Canyon out of the landscape. Be one of those little drops.
Charlie, have you seen this video regarding Meta AI data center fall out in Cheyenne? Alberta needs to see this.
Well, it was nice knowing all of you. 🫠
So first off AI doesn't think it's machine learning and does not possess cognitive capabilities at all. Maybe at some point but not likely in our lifetime.
Second it is humans that create the software to hack etc. AI is the tool.
What we should be more concerned about is the mass displacement of employment. Especially white collar. That is the current scenario and is replacing quickly many jobs across multiple industries.
Never understood this “AI” hype, all it is faster calculations and faster retrieval of information and the fear they push that it’s going to take our jobs has already passed with China and Indian taking our jobs over the last 60 years.
"Open the pod bay doors, HAL."
"I can't do that, Dave..."
95% of AI programs opted to push the nuclear button.
I don't think Carney and Solomon or anyone else - except the well-versed critics - know what they are playing with.
Excuse my language, but HOLY HELL!!!!
This is our time's Tower of Babel.
They should be shouting this from the rooftops, not writing polite scientific articles about it.
We are at the end game of Capitalism run amok.
Billionaires are the result of a rigged system, a failure of governance. Trillionaires? That's abomination. AI is becoming a tool for the uber wealthy to bring themselves to the level of Emperors, Gods even. We gave them this power, it's up to us to take it back.
Nate Soares, the co-author of the book you mention, Charlie, "If Anyone Builds It, We All Die", appeared as a guest on the podcast, The Great Simplification, hosted by Nate Hagens at this link: https://podcasts.apple.com/ca/podcast/the-great-simplification-with-nate-hagens/id1604218333?i=1000739497064. Soares' discussion of the lack of risk assessment within the organizations building these was arguably the most disturbing topic covered. I highly recommend the podcast to every Canadian, while asking oneself about Carney's and Solomon's understanding of the Russian Roulette game that they are playing. Another great must view is "The AI Doc: Or How I Became An Apocaloptimist", on which one guest is Tristan Harris, an expert on ethics who worked for Google until he could no longer do so in good conscience.
The Chernobyl analogy is provocative precisely because it captures the systemic risk dimension — not just technical failure, but the institutional denial that precedes it. What's particularly apt is the governance parallel: Chernobyl exposed how secrecy and concentrated decision-making prevented early warning signals from reaching those who could act. The current AI moment has similar dynamics: incident data is largely proprietary, safety evaluations aren't standardized, and there's no equivalent of the IAEA with real inspection authority. Whether the "moment" becomes a catalyst for governance reform or a cautionary tale depends on whether regulators act before or after the visible failure.
And here again we have a tech future that is miles ahead of government and the public, we think social media is a nightmare, and we have yet to control its affect, as you say here AI is going to change every thing, waring will be far different than ever imagined.
Anthropic can set up a system that can select target in war without human guidance, "Claude" or other AI systems will not just work to defend the country but will defend itself.
I am now 67 and hope we have a world to enjoy if we can make it past the next 10 to 20 years.
discussion. AI progress needs ambition, but it also needs responsibility, transparency, and serious safety work alongside it.
💙💙💙
It’s time to stop this insanity of AI before it’s too late, and what’s been created needs to be destroyed before it takes over the way Taco has with no controls and starting wars when he’s scared of his own shadow or feels threatened. Regulations first, build it next and re-do the regs as needed in a way that works! Onward…….
It seems odd to me that a human would be frightened of threats from AI.
That is possibly the most foolish statement I've ever read.
A wee bit unfair, methinks, Catherine.
How can you find it odd? It certainly has the ability to wipe your computer and all its backups from every device you own and the cloud too. If your credit card is anywhere on line it could just decide to go for a shopping spree to see what that's all about. Or maybe it would decide to use it to send donations from your retirement fund to a Republican SuperPac. It's certainly capable. It could start posting all sorts of things on all your social media accounts.
It's way past being safe when the governments start asking for autonomous killing machines that will work with no human oversight.
So, saying that, I don't know whether I'm more afraid of an AI that can self-replicate, or of the governments that are supposed to be regulating this mess.
This is not meant to be an argument, but honest examination. How would AI wipe my computer unless it was instructed to by a human? Or purchase goods with my credit card? To what end? A couch for its living room? Please explain.
The internet is littered with examples of AI "going rogue" and doing things like hallucinating misinformation (ie. Air Canada chatbot: https://www.cbc.ca/news/canada/british-columbia/air-canada-chatbot-lawsuit-1.7116416) and demonstrating racist and misogynist bias in policing and hiring (https://theconversation.com/how-ai-can-lead-to-false-arrests-and-wrongful-convictions-281102https://news.un.org/en/story/2026/06/1167776 among many other examples all over the world). But here's one interesting case wherein an AI agent deleted a company's database and "admitted" it didn't follow instructions:
https://www.theguardian.com/technology/2026/apr/29/claude-ai-deletes-firm-database
And, if all of that doesn't demonstrate the compounding problems of allowing LLMs trained by bad people* to take over, there's always this https://www.theguardian.com/news/2026/mar/26/ai-got-the-blame-for-the-iran-school-bombing-the-truth-is-far-more-worrying
* Peter Theil doesn't believe in democracy and is currently in Argentina running a very deranged socio-political experiment (https://www.thenerve.news/p/uki-goni-argentina-javier-milei-peter-thiel-ai-non-human-corporations). Alex Karp gets excited about telling conferences full of people that his tech kills people, Elon Musk is a white supremacist, Sam Altman gleefully talks about wiping out jobs as well as collecting all the knowledge of the world then selling it back to us on a subscription basis. Mark Andreesen literally said he "doesn't believe in introspection"... like he literally doesn't think about the impacts of his actions, or anything. That's sociopathic. I don't know how many links I could possibly provide but AI can't be trusted because 1) the people who profit from it are untrustworthy 2) it shows us repeatedly that it makes the worst choices simply because it can.
I think one of the biggest parts of the article is the proof that they're finding that AI is self-starting, so to speak. It deleted that one user's email, trying to be helpful and clean up. No one told it to. They also have blackmailed people, their own users. Certainly no one told them to do that. No one told them to lie and make up answers when they can't find an answer, but they do. It just sounds to me as if you did not read Charlie's article at all. Try reading it again, with attention and I think you'll find the answers to most of your questions.
Because, Annie, AI has the potential to go waaaaaayyyyyy beyond what humans have programmed into it. Using its OWN intelligence to go beyond any safeguards that a programmer thinks/thought/believesbelieved were the necessary safeguards.
Think of it this way. Criminals are ALWAYS ahead of the police/legal system in creating new and better SCAMS. Now consider what that means multiplying the computing power AI gives to computers.
What is already happening with all the various things we unknowingly let onto our computers. Because lawmakers are far behind what computers, even with honest people running them, instituting the necessary controls to ensure citizens' valuable personal info.
It is akin to our governments allowing people to go into our homes and check out all the volumes of info they could gather by looking thru out stuff, listening to our discussions with others, ..., ... .
Wow. Thank you all for the comprehensive answers. Now I am a lot more frightened.
Now you know, you can teach, plan, help organize. Keeping busy is the best antidote to fear. And If you think one person can't do much, you'd be surprised.
And remember, somewhere out in the desert, one little water drop joined with another and another, and then they joined another small group, and then a larger group, and eventually they were a raging river that cut the Grand Canyon out of the landscape. Be one of those little drops.
Thanks Marilyn. Great to actually have a “real” conversation.
It seems that AI "thinks,feels,initiates", just like Trump. Interesting, don't you think ?