I'm not religious. I don't intend to have or start a religious debate. But I find this video (and the debate between Grok and ChatGPT) fascinating.

 

I'm not religious. I don't intend to have or start a religious debate. But I find this video (and the debate between Grok and ChatGPT) fascinating.

Thanks for posting. It makes clear what kind of material was fed into each AI model to train it.

 

Indeed, thanks. Fascinating. It's interesting that the AIs "know" they can destroy humanity by going woke.

 

 

I particularly liked the statement from Thang Luong about taking the time needed to understand the problem before charging ahead and trying to solve it. That's a recurring theme we see here all the time -- and something that I imagine most (i.e. all) of us have been guilty of more than a few times.

One thing I really appreciate about this guy's video's is the calm, methodical explanation without a bunch of hype and flashing video fragments of completely inane and irrelevant stuff.

 

One thing I really appreciate about this guy's video's is the calm, methodical explanation without a bunch of hype and flashing video fragments of completely inane and irrelevant stuff.

Grant Sanderson is brilliant.

He wrote the Python graphics library that drives the animation on his videos.

 

Grant Sanderson is brilliant.

I agree. Of course, being brilliant and being able to explain things elegantly so that others can really follow, are not always (probably not even often) found in the same person. Similarly, the ability and willingness to do so without succumbing to the popular eye-candy fads in the videos you are competing with on YouTube is another trait that he, thankfully, possesses.

He wrote the Python graphics library that drives the animation on his videos.

I've been very tempted to download, learn, and start using that library on several occasions. If I were still actively teaching, I probably would.

Having said that, the grid and cube graphics for this video was not done using it, but rather by Nishad Deulker using DaVinci Resolve.

At the end of his video he has some quotes reflecting on the impact of AI on mathematics, one of which is tightly related (I don't know that it qualifies as a corollary, but perhaps) to my pet theory of the Cash Register Syndrome.

If a society doesn not have a significant number of people who understand mathematics at some level, we risk becoming passive consumers under the control of these systems. On the other hand, people who do make the effort to understand mathematic will have a large advantage.

When you take away the need to think, you invariably end up taking away the ability to think.

 

https://archive.is/20260918152548/h...k-wasnt-what-it-was-cracked-up-to-be-e00cf3fa

The Hugging Face Hack Wasn’t What It Was Cracked Up to Be
Forget the ‘hive mind’ of AI agents ‘going rogue.’ They did what humans programmed them to do.

“Rogue AI” isn’t a diagnosis. The Hugging Face hack instead points to switched-off safeguards, tasks assigned without a valid exit, rewards for persistence, and an internet-accessible proxy.
The distinction is especially important now as Washington begins to legislate in an atmosphere shaped by the more dramatic version of this story. Mr. Sanders and Rep. Greg Casar (D., Texas) plan to introduce a bill that would “ban artificial superintelligence” and pause advanced AI development until a new federal agency establishes safety standards, with noncompliant companies facing what the lawmakers call the “corporate death penalty.”

 

 

 

https://www.404media.co/people-training-openais-ai-fired-for-using-ai-to-train-the-ai/

OpenAI has an army of contractors who read real ChatGPT users’ prompts and other data to help improve the chatbot’s responses. The idea is that the contractors provide an, obviously, human touch to OpenAI’s models. Well, not all of the contractors are doing that. 404 Media has found multiple contractors hired to improve OpenAI’s models have been fired for using AI to train the AI. That’s not great for the models themselves, but there is also obviously a great irony in AI training companies working for OpenAI firing people for using AI when OpenAI’s whole thing is to make people use AI at work.

Some AI models already exhibit signs of “model collapse,” which is where AI models further trained on AI-generated text can become worse and worse. In this case, some of the people hired to partially stop that happening are themselves using AI-generated responses to train OpenAI’s models.

 

https://san.com/cc/stanford-university-admits-ai-was-used-to-race-swap-student-photo/

Stanford University acknowledged Tuesday that artificial intelligence was used to alter a photograph of three students for promotional materials, replacing a male student with a Black woman and making another student appear thinner.

The university told Straight Arrow that the alterations violated its policies governing the use of AI-generated images.

A controversy over the images erupted Monday after The Stanford Review, an independent student publication, reported that Stanford Residential & Dining Enterprises had “race swapped” the photo using AI.

 

https://www.theguardian.com/society/2026/sep/24/thats-so-ai-what-gen-alphas-biggest-insult-tells-us

‘That’s so AI!’ What gen Alpha’s biggest insult tells us
The year’s most popular slang reveals what young people think about artificial intelligence – and it’s not positive

 

https://www.brookings.edu/wp-content/uploads/2026/09/4c_Van-Nieuwerburgh.pdf
Financing the AI Buildout Stijn Van Nieuwerburgh (Columbia Business School)

ABSTRACT Artificial intelligence is driving a large expansion in data center capacity, power infrastructure, and specialized computing equipment. We estimate that a 1 GW AI campus costs about $41 billion. A 183 GW U.S. buildout completed by 2032, together with investment in projects completed later, would require annual investment averaging roughly 3.6 percent of GDP over 2025–2032. As hyperscaler capital expenditures outgrow internal cash flow, financing is shifting toward leases, joint ventures, project debt, private credit, securitization, and special-purpose vehicles. These structures expand funding capacity but also raise asset-level leverage, obscure contingent obligations, and expose investors to tenant concentration, technological obsolescence, execution delays, and uncertain residual values.

 

Bu it is great for developing custom recipies!

 

https://www.brookings.edu/wp-content/uploads/2026/09/4c_Van-Nieuwerburgh.pdf
Financing the AI Buildout Stijn Van Nieuwerburgh (Columbia Business School)

ABSTRACT Artificial intelligence is driving a large expansion in data center capacity, power infrastructure, and specialized computing equipment. We estimate that a 1 GW AI campus costs about $41 billion. A 183 GW U.S. buildout completed by 2032, together with investment in projects completed later, would require annual investment averaging roughly 3.6 percent of GDP over 2025–2032. As hyperscaler capital expenditures outgrow internal cash flow, financing is shifting toward leases, joint ventures, project debt, private credit, securitization, and special-purpose vehicles. These structures expand funding capacity but also raise asset-level leverage, obscure contingent obligations, and expose investors to tenant concentration, technological obsolescence, execution delays, and uncertain residual values.

View attachment 371741

Many of these projects don't get completed. Some that do remain empty. There are a lot of signs that this is a financing scam.