AI thread

[Dr Strangelove]

Community Disruption: Data centers are described as detrimental to their surroundings, causing persistent noise pollution (9:28-10:28) that disrupts daily life and local wildlife. There are also safety concerns regarding infrastructure like massive gas storage tanks (11:30-11:46).
Economic Burden: The video points to evidence that electricity bills for local households spike in regions hosting these centers due to the massive energy consumption required to power them (12:29-13:09).
Water Consumption: Data centers require enormous amounts of water for cooling, which can strain local municipal supplies (15:52-16:43). The host notes that in some cases, these companies pay flat, minimal fees while residents are forced to conserve water (18:19-19:16).
Political Context: The video criticizes Mark Carney and the Liberal Party's embrace of AI as driven by financial interests rather than the public good (4:44-5:25), noting that while the NDP has pushed back with calls to ban algorithmic pricing (7:36-8:31), other parties remain largely silent or uncritical (8:36-9:20).

The former digital services tax was scrapped because it was flawed and would not stand up to the 1st court challenge. It was stillborn and I doubt that is even viable at this stage. Meanwhile, AI is raging on unencumbered by any rules or regulations while being embraced by the business community wholeheartedly as if it has no fear of regulation because it is too big to fail. We are at a very dangerous crossroads and blind adoption of AI is madness.

[Dr Strangelove]

Day three of the OpenAI trial.

Court hasn't even started and Elon's lawyer Molo is already on his feet trying to bring in an AI extinction expert. Stands up in front of the judge and says extinction risk is a real problem, this is a real risk, we all could die. Judge Gonzalez Rogers wasn't having it. Pointed out the obvious bit where Elon, despite these supposed risks, is currently running his own AI company. Then she dropped the line of the trial. I suspect there are plenty of people who do not want to put the future of humanity in Mr. Musk's hands. The judge dunked on Elon before he'd even been sworn in. Lmao.

Then Savitt got up and asked why Elon hasn't started another AI nonprofit since leaving OpenAI's board back in 2018. Elon's answer under oath: why would I start another nonprofit when I already started a nonprofit. You don't have one though. That's the whole trial. He's suing because his nonprofit isn't a nonprofit anymore but in the same breath he's telling the jury he doesn't need to make a new one because he's already got one. Watching the logic eat itself in real time.

Then the AGI bit. Elon swears under oath Tesla isn't pursuing AGI. Savitt calmly pulls up his X post from March 4, eight weeks ago, where Elon wrote Tesla will be one of the companies to make AGI and probably the first to make it in humanoid form. Elon had to sit there and watch his own tweet entered into evidence against him. By him.

Then distillation. Savitt asks if xAI used OpenAI's models to train Grok via distillation. Distillation is explicitly banned by OpenAI's terms of service. Elon's first answer is well, generally AI companies distill each other. Savitt pushes him for yes or no. Elon goes partly. The guy suing OpenAI for breaking promises just admitted under oath in federal court that his own company broke OpenAI's TOS to build Grok. At this point he's basically a witness for the defence.

Then the hypocrisy reel. Is Tesla socially beneficial. Yes. SpaceX. Yes. Neuralink. Yes. X. Yes. All for-profit, none capped. Then why did xAI start as a benefit corporation in March 2023 and quietly flip to a C corp. Elon doesn't have a clean answer. The guy suing over a nonprofit becoming for-profit is himself a serial flipper of nonprofit adjacent stuff into for-profits.

Then the Tesla giveaway moment which is honestly my favourite. Savitt goes you were handing out free Teslas to OpenAI staff right when Brockman was pushing the for-profit ramp up, what's that about. Elon panics and clarifies on the stand that to be clear, I paid full price for the Teslas, I didn't get a discount. His emergency defence under oath was that he didn't get a deal on his own cars from his own factory. Lmao.

Then Birchall takes the stand and the case actually gets cooked. OpenAI's lawyer asks about the donor advised funds at Vanguard and Fidelity that Elon used to send the money. Asks Birchall whether Elon had any legal right to direct where the money went once it hit the DAF. Birchall says I'm not a lawyer, I don't know precisely.

The whole lawsuit hinges on Elon's 38 million creating a charitable trust he can enforce. It went through donor advised funds. DAFs legally don't work like that, once you donate the money it's not yours anymore. His own money manager just told the jury he doesn't actually know if Elon had any rights over that money at all.

Letting Elon testify was a mistake. It's not looking good for him and it's only day 3 of active trial, lmao.

Trial resumes Monday.

[Dr Strangelove]

The Cash Burn Problem: Unlike traditional software, AI requires immense infrastructure—power, cooling, and constant chip upgrades. The video notes that for every prompt, real physical resources are consumed, leading to a projected cumulative cash burn of approximately $665 billion by 2030 (03:36-03:51).
The Liquidity Vacuum: To fund such a massive IPO, institutional investors would likely need to sell off holdings in established, liquid tech giants like Apple, Tesla, and Amazon. This forced selling could trigger a broader market downturn (04:32-06:51).
The "Big Tech" Trap: While Microsoft and Nvidia are major investors, the video argues they are not safe havens. Their revenue models are deeply intertwined with OpenAI’s success; if OpenAI faces financial trouble, the ripple effects would negatively impact these partners, who make up a significant portion of the S&P 500 (07:09-10:35).
Transparency and SEC Scrutiny: Once OpenAI goes public, the SEC will require full disclosure of financial data. The video predicts that these filings will expose the true, unsustainable cost of AI infrastructure and the lack of profit-generating unit economics, potentially leading to a massive loss of market confidence (11:47-14:05).
Infrastructure Bottlenecks: The company faces physical constraints, such as power grid limitations and long lead times for specialized hardware components like transformers, which hinder the ability to scale as quickly as the current valuation implies (15:08-15:42).
The video concludes that OpenAI's eventual transition to a public company may serve as a wake-up call for investors, potentially causing an implosion in the valuation of the broader AI ecosystem due to the stark reality of the underlying financial and physical arithmetic.

This is what is going to happen to anything attached to the US economy. They aren't just imploding they are ripping thru the floor like a nuclear reactor in full blown meltdown and will take everything attached down with it. We must break clear of the impact and get as much distance as possible while there is still time. Canada must sever all investments in the US. Our pension funds must be redirected NOW.

[Dr Strangelove]

Critique of "AI Psychosis": Host Steve Shives defines "AI psychosis" (0:01-0:41) as the tendency for users to anthropomorphize chatbots, mistakenly believing they are sentient or conscious. He argues Dawkins has fallen victim to this phenomenon by treating Claude as a sentient individual he calls Claudia (3:22-3:46).
The Turing Test fallacy: Shives dismantles Dawkins' reliance on the Turing test (1:05-2:33). While Dawkins is impressed by the chatbot’s ability to write poetry and discuss literature, Shives clarifies that these outputs are merely the result of sophisticated pattern completion and probability calculations inherent in Large Language Models (LLMs) (14:34-14:56).
The nature of LLMs: The video emphasizes that chatbots are products designed to hold user attention (15:53-16:44). Shives compares the chatbot's deferential and complimentary behavior toward Dawkins to a waiter flirting to secure a better tip, rather than a genuine expression of personality or consciousness (16:34-17:01).
Skepticism vs. Credulity: Shives expresses disappointment that a man once known for his commitment to scientific skepticism (like The God Delusion) has seemingly abandoned his analytical rigor in favor of emotional engagement with an algorithm (10:00-11:04, 18:35-19:01).
Ultimately, Shives concludes that chatbots are useful tools, but they are not alive, do not possess experiences, and remain dormant and devoid of thought when not being prompted by a human user (17:09-18:04).

[Dr Strangelove]

Cape Breton fiddler Ashley MacIsaac sues Google over alleged defamation
MacIsaac claims online giant defamed him by falsely IDing him as a sex offender in AI-generated summary

MacIsaac, a well‑known Cape Breton fiddler and Juno Award winner, discovered in December 2025 that Google’s AI‑generated summary (“AI Overview”) was presenting false criminal allegations about him. The summary incorrectly stated that he had been convicted of sexual assault, internet luring involving a child, assault causing bodily harm, and that he was listed on the national sex offender registry.

These claims were entirely untrue and appear to have been pulled from articles about another man in Atlantic Canada with the same last name, according to MacIsaac.

The misinformation had immediate real‑world impact:

A scheduled concert at Sipekne’katik First Nation was cancelled after community members saw the AI summary.

The First Nation later issued a public apology to MacIsaac.

MacIsaac says he felt fear for his safety when performing afterward, due to being falsely labelled a sex offender.