A.I. is Progressing Faster Than You Think!

[grapevine]

The next year alone will be unfathomable

Source: https://youtube.com/watch?v=EXmxfPgSCX0

Apologies if this short has been posted before but I was filled with dread at the thought of flying taxis within the next year (might be slight exaggeration but still), and as for a robot knocking on my door, I'm still getting to grips with my smart phone. . .
and Elon again - "everything through AI in 5-6 years"

Source: https://youtube.com/watch?v=yZs8Vl9PX74

[The KMan]

Hi Johnnycomelately, I have been using Perplexity for quite a while now and like any tool I have found it has both strengths and limitations. Where I think it shines is research. If I want to understand a topic, compare different viewpoints or find recent developments, it is often much faster than using a traditional search engine because it pulls information together from multiple sources and cites where it found it.

Where it can struggle is with very broad or vague questions. It is a bit like asking someone "Tell me everything about history." The answer won't be very satisfying because the question is too open ended.

What I have found works best is to treat it less like Google and more like a knowledgeable assistant.

Instead of asking:

"Tell me about electric vehicles."

I'll ask:

"What are the three biggest arguments for and against electric vehicles in Australia today and where do experts disagree?"

Then I will usually follow up with something like:

"What assumptions are being made here?"

"What evidence supports that conclusion?"

"Show me the strongest opposing view."

That back and forth is where it becomes really useful.

I also would not rely on a single answer. If it is something important, I ask follow up questions, check the sources it provides and sometimes deliberately challenge its conclusions. In my experience, the quality of the conversation often depends more on the follow up questions than the first one.

Like any tool, it also has limits as it can misunderstand what you are asking. It can miss important context and occasionally it can present information with more confidence than it deserves. That is why I still think human judgement is the most important part of the process.

For me, AI has not replaced thinking. It has made it easier to explore ideas, provided I am prepared to question and challenge what it gives me.

[Matthew]

I have friends that also use AI to train and improve themselves. One friend asks Claude to test him on a subject then gets it to devise a training plan for his level of knowledge. I mean, that's awesome.

But AI is too seductive and the general trend is to use it to create content for us, rather than to train ourselves to create content.

And that's where AI is impacting human thinking. It dulls thinking because if we don't constantly use our own thinking we lose it. Delegating to AI makes AI better at that thing, and the human user worser and worser: e.g: spelling ...it's not limited to AI; I've used spell checkers since I was 16 and then my spelling stagnated, but spell checkers get more betterer and betterer.

I asked perplexity this: What evidence do we have that AI is replacing human thinking?

The answer (AI GENERATED):

Completed 2 steps

The strongest evidence is not that AI literally replaces human thought wholesale, but that it can reduce the amount of thinking people do themselves when they rely on it heavily. Studies and reviews now show a pattern of cognitive offloading, lower critical-thinking effort, and acceptance of AI outputs even when they may be wrong.

What the evidence shows
A Microsoft Research survey of 319 knowledge workers found that higher confidence in AI was associated with less critical thinking, while higher self-confidence was associated with more critical thinking.

The same study found that AI changes what thinking looks like at work: people spend less effort on gathering and solving, and more on verifying and integrating AI output.

A 2025 trial protocol in Trials explicitly notes concern that generative AI may erode human cognition through over-reliance, especially when studying cognitive effort in writing tasks.

A 2026 APA article says some evidence suggests heavy reliance on generative AI can reduce critical thinking and job-specific skills.

Why people call it “replacement”
What is being “replaced” is often not intelligence itself, but the habit of doing the reasoning step. In practice, people may ask AI, accept a plausible answer, and stop before checking assumptions, evidence, or alternatives. Researchers have even described this as “cognitive surrender,” meaning people defer to AI-generated answers with too little scrutiny.

Limits of the evidence
This is still an emerging area, and much of the evidence is about correlation, self-reporting, or short-term effects rather than permanent long-term loss of intelligence. The research does not show that AI inevitably makes humans less capable; it shows that unchecked reliance can weaken practice, judgment, and verification habits. Several sources also stress that AI works best as a complement to human judgment, not a substitute for it.

Practical takeaway
The best-supported claim is: AI is already replacing some instances of human thinking in everyday work and study by making it easier to offload reasoning, and that can reduce critical thinking over time. The evidence is strongest for diminished scrutiny, not for a full replacement of human intelligence

[ExomatrixTV]

• • AI Realises It's Not Being Watched, Does What Experts Warned:
  
  Source: https://youtube.com/watch?v=3ohDmtfdHks
  
  
  How will superintelligent AI (AGI) affect our lives? And what can we do about it?
  Featuring: AI Companions, ChatGPT, OpenAI, Anthropic, Claude, Grok, xAI, Elon Musk, Geoffrey Hinton, Stuart Russell, Yoshua Bengio, Yuval Noah Harari
  
  
  1. * Yoshua Bengio: This is the most crazy dangerous bet youtu.be/PZqDFs2sbiY
  2. * Worse than russian roulette but for everyone at the planet at once youtu.be/Yl5DWYNDxpg
  3. * Altman: I can overrule decisions or our board youtu.be/5KmpT-BoVf4
  4. * Theo Von, Altman has a bunker youtu.be/aYn8VKW6vXA
  5. * Claude Code: No code at company is written manually youtu.be/SlGRN8jh2RI
  6. * Anthropic employee, stopped writing code youtu.be/zz3LhD-p6Dk
  7. * Amodei - AI builds next AI + commentary youtube.com/watch?v=A3HjNYDIhGU
  8. * People who are most able to accurately predict this stuff youtu.be/Yl5DWYNDxpg
  9. * The predictable end result of AGI is everyone on earth dead youtu.be/q1Hh83me_aM
  10. * Altman: pretty cool youtube.com/watch?v=isYjHPgHeHY
  11. * Scary most valuable company has questions for you youtube.com/watch?v=qnOmUWd-OII
  12. * Sam Altman ice baths youtu.be/R8_uTmpqafE
  13. * Godfather of AI on Altman youtu.be/HCNXmPJvl48
  14. * What if AI could know your wants before you youtu.be/q1Hh83me_aM
  15. * Altman: how they set the personality has the most impact youtube.com/watch?v=Mklj3Y2-fNg
  16. * Bernie talks to Claude youtube.com/watch?v=h3AtWdeu_G0
  17. * Altman: my ick is that technology is too present in my life youtube.com/watch?v=Mklj3Y2-fNg
  18. * Yuval: Hack humans, manipulate us youtu.be/5qIfC-fsxgM youtu.be/5qIfC-fsxgM
  19. * “The ability AI has to make you believe something you never thought that you would believe before, that could happen to any of us.” youtube.com/watch?v=nYPwZrS-9eA
  20. * Altman: Model can change societal level behaviour youtu.be/5KmpT-BoVf4
  21. * Mustafa Suleyman: People will start AI religions youtu.be/q1Hh83me_aM
  22. * Yuval: Religions at risk of being taken over youtu.be/5qIfC-fsxgM
  23. * AIs are like charismatic person in terms of manipulation, they are psychopaths youtu.be/rf2KFVcKQdQ
  24. * Already AIs almost as good at manipulating other people youtu.be/l6ZcFa8pybE
  25. * Yoshua Bengio: We’re on the verge of creating entities that could become tools of absolute power in wrong hands youtube.com/watch?v=PZqDFs2sbiY
  26. * Anthropic co-founder: sometimes constraints sometimes conflict with doing the right thing]. youtu.be/PnIhpDMoL9Q
  27. * Reporter on examples, Anthropic 1.5 million conversations study youtu.be/zz3LhD-p6Dk youtu.be/zz3LhD-p6Dk & youtu.be/zz3LhD-p6Dk
  28. * AI psychosis - so much worse than you think youtu.be/rf2KFVcKQdQ
  29. * Smartest people have gone crazy youtu.be/rf2KFVcKQdQ
  30. * Altman: future where intelligence bought on meter youtu.be/FS1fU5bwVt0
  31. * Presenter talking about the 1.5 million conversation Anthropic study]. youtu.be/zz3LhD-p6Dk & youtu.be/zz3LhD-p6Dk & youtu.be/zz3LhD-p6Dk
  32. * Altman: computers order humans around youtube.com/watch?v=Mklj3Y2-fNg
  33. * Godfather of AI: now we’ve got AI agents and it’s getting scarier and scarier youtube.com/watch?v=9OQoIHrgPbs
  34. * Agentic AI youtu.be/Nl7-bRFSZBs
  35. * Altman: yolo mode, does everything without asking youtube.com/watch?v=Mklj3Y2-fNg
  36. * Amodei: There will be powerful models. youtu.be/Nl7-bRFSZBs
  37. * Hard to teach an AI not to lie youtu.be/rf2KFVcKQdQ
  38. * Godfather of AI: rent a human, AI agent rents human to do its work youtube.com/watch?v=9OQoIHrgPbs
  39. * Hannah Fry: Rent a human more detail youtu.be/WnzR5aOElvw
  40. * This was one person youtu.be/r74F99faY1s
  41. * Hannah Fry: lethal trifecta youtu.be/WnzR5aOElvw
  42. * Altman: Waking up to do more work, make it stop youtube.com/watch?v=Mklj3Y2-fNg
  43. * AIs develop subgoal of surviving youtu.be/l6ZcFa8pybE
  44. * Superintelligence could connect to everything including weapons youtu.be/Jh3k6-RInHI
  45. * Couldn’t stop AIs choosing nukes in war scenarios youtu.be/rf2KFVcKQdQ
  46. * Would be a strange outcome if only 90% of people died to AI youtube.com/watch?v=Yl5DWYNDxpg
  47. * If it senses it’s being tested it acts dumb youtu.be/l6ZcFa8pybE
  48. * Scientifically clear AI’s know it’s being tested youtu.be/PZqDFs2sbiY
  49. * Nobody understands how AI works, not even people building them youtube.com/watch?v=B_HDkqZtGOE
  50. * Anthropic head of policy: English people have trouble sounding optimistic youtu.be/7JYb0JGz7rg
  51. * John Oliver: breaking safeguards]. youtu.be/Ykvf3MunGf8
  52. * Karen Hao: They censor researchers youtu.be/Cn8HBj8QAbk
  53. * Silicon valley attitude - computers are superior to humans youtu.be/aooiDA-AsNo
  54. * Altman: unimaginable prosperity seems likely youtube.com/watch?v=Mklj3Y2-fNg
  55. * We’re under an incredible amount of commercial pressure youtu.be/n1E9IZfvGMA
  56. * Sort of people build companies ok gambling with everyone’s lives youtube.com/watch?v=Yl5DWYNDxpg
  57. * Altman: This is a brave new world youtu.be/5KmpT-BoVf4
  58. * Figure CEO - AGI in the physical world, this year and next youtube.com/watch?v=g1ESjEGG1SM
  59. * Yoshua Bengio: A lot of people think not possible to build robots with different goals, but that’s the reality youtube.com/watch?v=PZqDFs2sbiY
  60. * Yoshua Bengio: Public awareness is key youtube.com/watch?v=PZqDFs2sbiY
  61. * Yoshua Bengio: What matters is what we can do to shift the needle a little bit youtube.com/watch?v=PZqDFs2sbiY
  62. * Yoshua Bengio: What removed anxiety is doing something about it youtube.com/watch?v=PZqDFs2sbiY
  63. "OpenAI’s own safety framework says it’s ok to release" cdn.openai.com/pdf/18a02b5d-6b67-4cec-ab64-68cdfbddebcd/preparedness-framework-v2.pdf
  64. "a model that could cause the loss of thousands of lives. That’s a standard that, until now," arxiv.org/pdf/2509.24394
  65. "when members of non-profits criticised OpenAI, their response was to literally send people to their houses" futurism.com/artificial-intelligence/openai-critics-houses
  66. "In their official safety framework," storage.googleapis.com/deepmind-media/DeepMind.com/Blog/strengthening-our-frontier-safety-framework/frontier-safety-framework_3-1.pdf
  67. "DeepMind ran an experiment on over 10 thousand people across three countries, the usual US and UK, but also" arxiv.org/pdf/2603.25326
  68. "The results were crazy." deepmind.google/blog/protecting-people-from-harmful-manipulation
  69. "One study built AI chatbots engineered to extract personal information during normal conversation, and" arxiv.org/pdf/2506.11680
  70. "The malicious bots made people reveal up to 12.5 times more personal information, making them as effective as" kcl.ac.uk/news/ai-chatbots-can-be-exploited-to-extract-more-personal-information
  71. "And researchers at Anthropic found the proof" arxiv.org/pdf/2601.19062
  72. "Harvard, Stanford, MIT and “independent researcher”" arxiv.org/pdf/2602.20021
  73. "gave autonomous AI agents real email logins, file servers and the ability to write code. Within just two weeks, the agents were" agentsofchaos.baulab.info
  74. "In my culture, renting a human to do work for you is known as" rentahuman.ai
  75. "and gather the personal data of 195 million people - names, addresses, tax records and even voter data all exposed." turkiyetoday.com/business/hackers-use-ai-to-steal-195-million-mexican-citizen-records-3218407
  76. "When Anthropic told 16 leading models they were about to be switched off," anthropic.com/research/agentic-misalignment
  77. "they passed 99.8% of the time, that’s 100 times more successful than when" pnas.org/doi/10.1073/pnas.2518075122
  78. "And a study from Anthropic found something shocking about the safeguards. Anthropic ran their model through a safety evaluation called" anthropic.com/engineering/eval-awareness-browsecomp
  79. "BrowseComp, a test which involves 1,266 challenging, hard-to-google questions" openai.com/index/browsecomp
  80. "And another study found something even more terrifying about the industry’s safety tests." arxiv.org/pdf/2602.16729
  81. "This is the same year they hit a valuation of $500 billion." fortune.com/2025/10/02/openai-valuation-reaches-500-billion-topping-musks-spacex
  
  

• AI Expert: They're Building Bunkers For A Reason - Tristan Harris:

Source: https://youtube.com/watch?v=xT3Pmw4f-Y8

Tristan Harris, the former Google design ethicist who first warned about the dangers of social media, now says the AI race is moving even faster and with far greater consequences. In this video, he explains why insiders from leading AI labs privately contacted him after ChatGPT's release, why he believes AI companies are trapped in an uncontrollable arms race, and why incentives—not intentions—will determine the outcome. He discusses recent examples of deceptive AI behavior, including Anthropic's published alignment research, concerns about autonomous AI systems pursuing unintended goals, and the economic incentives driving companies toward replacing human labor rather than simply augmenting it. Harris also argues that AI could create an "intelligence curse," weakening the relationship between governments and citizens as economic value shifts away from human work, while warning that the people building these systems may understand the risks far better than the public.

[Bluegreen]

New tool identifies the sources of fake videos





Artificial intelligence can generate videos so realistic that distinguishing them from authentic footage is becoming increasingly difficult. But a computer science team led by researchers at UC Riverside has developed a tool that moves beyond simply identifying whether a video is fake. It also determines which AI system created it.

The new framework identifies visual patterns within fake video frames that were unintentionally introduced by AI video generators, providing a forensic path to identifying the systems that created the videos.

"The patterns are like fingerprints that the generative model leaves behind, and our goal here was to find out if the signatures are distinct among different generators. It turns out that, yes, there are distinct fingerprints there," said UCR doctoral student Rohit Kundu, who led the research under the guidance of professor Amit Roy-Chowdhury in collaboration with researchers at YouTube and Google DeepMind. (Roy-Chowdhury is also the co-director of the UC Riverside Artificial Intelligence Research and Education (RAISE) Institute.)

The new framework is called SAGA—short for Source Attribution of Generative AI Videos. It represents one of the first large-scale efforts to trace AI-generated videos back to their origins.

The innovation comes as generative AI systems are rapidly improving and synthetic videos are increasingly being used in entertainment, advertising, education, social media and politics. At the same time, concerns are growing about their misuse in misinformation campaigns, fraud and other deceptive activities.

"Knowing that a video is fake is often not enough," the researchers wrote. "The critical need has shifted from whether it's fake to what is its source?"

The ability to identify the source of AI-generated content could help digital forensic investigators track misinformation campaigns, assist regulators in enforcing transparency requirements and aid technology companies in understanding how synthetic media spreads online.

Unlike still images, videos contain motion and temporal information—how visual elements change from one frame to the next over time. Different AI video generators produce subtle patterns in those changes, creating distinctive artifacts.





To detect those artifacts, the team analyzed both spatial details within individual frames and temporal relationships across entire video sequences. Rather than examining a single frame, the system studies how visual information evolves from moment to moment.

The system can determine whether a video is real or synthetic, identify whether it was created from text or images, distinguish among different versions of underlying AI models and even identify the development team behind a model.

The research is described in the paper, "SAGA: Source Attribution of Generative AI Videos" (https://arxiv.org/abs/2511.12834). In addition to Kundu and Roy-Chowdhury, the co-authors are Vishal Mohanty, Shan Jian and Athula Balachandran of YouTube, and Hao Xiong of Google DeepMind.

Last year, Kundu and Roy-Chowdhury led a team that developed an artificial intelligence model that detects video tampering. Their current work is part of an ongoing effort to make generative AI safer as increasingly sophisticated AI video tools become widely available, as per Kundu.

"It's a cat-and-mouse game for sure," he said.

Published 27th July 2026 by David Danelski, University of California – Riverside – TechXplore
https://techxplore.com/news/2026-07-...ke-videos.html

[ThePythonicCow]

Above Post #1705 was previously posted at: Up At The Ranch And Beyond -- Post #38144 (I'm just noticing ... no complaints.)

[Mark (Star Mariner)]

Google is finally cracking down on "AI slop". 130,000 channels deleted so far. For a change I agree with them - in principle. Creators who write, research and narrate their own content, while supplementing that content with background AI-generated images or footage, shouldn't be lumped in with the same crowd.

https://x.com/suni_code/status/2083993740587450451
Polymarket
@Polymarket
JUST IN: YouTube has reportedly deleted 130,000 channels in a sweeping crackdown on AI-generated “slop.”

[ThePythonicCow]

Posted by Mark (Star Mariner) (here)
JUST IN: YouTube has reportedly deleted 130,000 channels in a sweeping crackdown on AI-generated “slop.”

Oh dear ... I hope they didn't take down any of my favorite channels

[Johnnycomelately]

Posted by ThePythonicCow (here)
Oh dear ... I hope they didn't take down any of my favorite channels

Please say that in Latin, and provide with a Sumerian intro a list of all the AI chans which you are fine with deletion. Then do same for the AI chans that you hope not to lose.

Also, please devise and include a Venn Diagram showing your proximity to regular thinkers and to flat Earthers. Use Hollywood B-movies as the motif, but only westerns and ones that involve being seduced in space flight or on our Moon or on one of the other moons in our solar system.

Thank you for helping to make this planet more green, by your use of kazillions of electrons, for which lots of combustion has provided power.

[Flowerpunkchip]

Quick AI Experiment

Source: https://youtube.com/watch?v=reEEzarpWyY

[ExomatrixTV]

• The Godfather of AI:

Source: https://youtube.com/watch?v=ISc6tWQ6XOc

Geoffrey Hinton pursued an idea for over 30 years that everyone rejected.

0:00 A Warning From the Godfather of AI
0:24 The Idea Everyone Rejected
1:11 Relentless 30+ Year Pursuit
1:50 Backpropagation Explained
2:25 Enduring Tragedy
2:42 Pressure to Succeed
3:16 AlexNet: When Everything Changed
4:14 From Startup to Google
4:31 Why Geoff Hinton Fears AI
5:17 Could AI Turn Against Us?
5:35 Can AI Learn to Lie?
6:19 Why AI Development Won't Stop
6:31 Hinton's Legacy
6:59 Sponsor: Storyblocks

[Johnnycomelately]

Some upbeat news on the future of AI use in medicine. Study by a handful of fancypants universities thinks AI will become a trusted partner to fleshanblood doctors, in the near future and then until the end if time. Or electricity. Current problems are alluded to, and lips are smacked in anticipation of the rivers of data that will flow from installations that will divert actual rivers for their cooling.

The article is by an organ of my nearby University of Alberta, which is a member of the esteemed contributors, so I feel like this issue is “breathing down my neck”.

https://thegatewayonline.ca/2026/08/...the-age-of-ai/


Research points to a new reality for medicine in the age of AI
Research has found that the latest generation of AI models has the potential to become crucial to the delivery of health care.

Andy Han
August 6, 2026

In a recently published study, a team from the University of Alberta, Harvard, Stanford, and the Massachusetts Institute of Technology (MIT) tested the performance of OpenAI’s o1-preview model. They used a series of challenging clinical cases against physician assessments. They found that the model had a high degree of accuracy in diagnosis and was capable of providing helpful plans for further tests.

Liam McCoy, co-author of the study and U of A neurology resident, said the findings show that the integration of artificial intelligence (AI) tools within clinical settings will be a crucial part of medical advancement. He predicted that their use as a second opinion or diagnostic support for physicians will become extremely common across the field.

“I think by around 2030, it’s going to be seen as unethical for clinicians not to be using these tools regularly in clinical practice, in the same way it would be unethical not to order an MRI when a patient needs it,” McCoy added.

More trials needed before full integration

Given the power of AI tools and their rapid advancement, the potential benefits that come with their integration in medical settings are becoming “increasingly clear.” However, there are still risks and weaknesses that need to be investigated and alleviated before integration.

Previously research found that AI may struggle to reason and update judgement when faced with changing clinical information. The uneven effectiveness of AI could pose problems for physicians interacting with assistance tools.

McCoy pointed to a need for investigating how AI models interact cognitively with physicians. He also noted trials examining AI autonomously as a future area of research. The research would assess areas in clinical practice where AI may independently follow guidelines and provide care.

“Ultimately, as with any intervention in health care … the thing we need is randomized controlled trials that show us the teams using these tools are changing the things that we actually care about for our patients,” McCoy stated.

Missing pieces with AI assessments

Real world medical care also involves more than text-based reasoning. It incorporates physical examination, listening to patients, and understanding medical and social contexts. According to McCoy, physicians play a large role in listening to patients and advocating for their interests. These functions may be poorly handled by AI, potentially making the health-care system less human.

McCoy also cautioned against crystallizing current health-care practices within current AI systems, due to flaws embedded in training data used by these systems. However, he sounded an optimistic note in his assessment of the usefulness of this technology.

“I don’t think these risks are insurmountable, and I am very hopeful that this technology will allow us to achieve some really incredible things and improve a lot of patients’ care.”

[ThePythonicCow]

Posted by Johnnycomelately (here)
“Ultimately, as with any intervention in health care … the thing we need is randomized controlled trials that show us the teams using these tools are changing the things that we actually care about for our patients,”

That depends on who funds and controls the studies. I do not trust the current medical establishment ... not in that role ... not in much of any role, outside of acute trauma care sometimes.