Scientists Create a “Periodic Table” for Artificial Intelligence
...
“We found that many of today’s most successful AI methods boil down to a single, simple idea — compress multiple kinds of data just enough to keep the pieces that truly predict what you need,” says Ilya Nemenman, Emory professor of physics and senior author of the paper. “This gives us a kind of ‘periodic table’ of AI methods. Different methods fall into different cells, based on which information a method’s loss function retains or discards.”
A loss function is the mathematical rule an AI system uses to evaluate how wrong its predictions are. During training, the model continually adjusts its internal parameters in order to reduce this error, using the loss function as a guide. ...
They call this approach the Variational Multivariate Information Bottleneck Framework.
“Our framework is essentially like a control knob,” says co-author Michael Martini, who worked on the project as an Emory postdoctoral fellow and research scientist in Nemenman’s group. “You can ‘dial the knob’ to determine the information to retain to solve a particular problem.” ...
“The machine-learning community is focused on achieving accuracy in a system without necessarily understanding why a system is working,” Abdelaleem explains. “As physicists, however, we want to understand how and why something works. So, we focused on finding fundamental, unifying principals to connect different AI methods together.” ...
The researchers applied their framework to dozens of AI methods to test its efficacy. ...
See the full story here: https://scitechdaily.com/scientists-create-a-periodic-table-for-artificial-intelligence/
OpenAI Admits Agentic AI May Never Be Secure
...
As Rami McCarthy, principal security researcher at Wiz, puts it, "A useful way to reason about risk in AI systems is autonomy multiplied by access." The more your agent can do, and the more data it can reach, the higher your exposure.
Agent mode in ChatGPT Atlas allows the browser agent to view webpages and take actions, clicks, and keystrokes inside your browser, just as you would. That's the value proposition. Security researchers responded by publishing demos showing it was possible to write a few words in Google Docs that changed the browser's behavior. That's the vulnerability. ...
McCarthy's assessment is blunt: "For most everyday use cases, agentic browsers don't yet deliver enough value to justify their current risk profile." ...
Agents are synthetic employees and they should be treated as such. Start with minimum necessary permissions and expand only with clear business justification. Audit what your agents can access today. Require confirmation steps for anything involving money, messages, or sensitive data. ...
See the full story here: https://shellypalmer.com/2025/12/openai-admits-agentic-ai-may-never-be-secure/
“NO LONGER A FUTURISTIC CONCEPT”: SCIENTISTS BUILD REAL-WORLD MAGNETIC CLOAKING DEVICE CAPABLE OF SHIELDING COMPLEX SHAPES
...
The magnetic cloak operates by manipulating the flow of magnetic fields around an object, making the fields behave as if no object were present. Revealed for the first time in a paper published in Science Advances, the team’s work was demonstrated practically after they constructed a real-world magnetic cloaking device from superconductors and soft ferromagnets. ...
... the new cloaks proved effective across a wide range of magnetic field strengths and frequencies, even when shielding irregular forms. ...
The cloak could function as a shield that reduces magnetic interference when electronic systems operate in close proximity, particularly in space technology, confined settings, renewable energy systems, and medical devices. ...
Crucially, the research demonstrates that magnetic cloaking devices can be constructed using commercially available materials, without relying on impractical or highly specialized metamaterials. According to the team, potential applications include fusion reactors, medical imaging systems, quantum sensors, and advanced communications technologies. ...
The paper, “Designing Functional Magnetic Cloaks for Real-World Geometries,” appeared in Science Advances on December 19, 2025.
See the full story here: https://thedebrief.org/no-longer-a-futuristic-concept-scientists-build-real-world-magnetic-cloaking-device-capable-of-shielding-complex-shapes/?utm_source=substack&utm_medium=email
Six (or seven) predictions for AI 2026 from a Generative AI realist
...
- No country will take a decisive lead in the GenAI “race”.
- Work on new approaches such as world models and neurosymbolic will escalate.
- 2025 will be known as the year of the peak bubble, and also the moment at which Wall Street began to lose confidence in generative AI. Valuations may go up before they fall, but the Oracle craze early in September and what has happened since will in hindsight be seen as the beginning of the end.
- Backlash to Generative AI and radical deregulation will escalate. In the midterms, AI will be an election issue for first time. Trump may eventually distance himself from AI because of this backlash.
...
See the full story here: https://garymarcus.substack.com/p/six-or-seven-predictions-for-ai-2026
The Disney is Dying – Apple Should Pull the Plug
...
Ten years ago, the gap between these two giants was manageable. Apple was worth approximately $500 billion. Disney sat at $200 billion. It was a partnership of equals. Steve Jobs was Disney’s largest individual shareholder; Bob Iger sat on Apple’s board. Today, that relationship has warped. Apple is hovering near a $4 trillion valuation. Disney is still stuck at $200 billion.
The gap tells the story of the last decade. Big Tech fought a war for media dominance, and the traditional media companies lost. Amazon absorbed MGM to secure James Bond. Google turned YouTube into the largest television platform on Earth. Netflix built a half-trillion-dollar global content engine. Disney, the company that defined American entertainment for a century, is now the smallest player in the room. ...
We are entering an era where AI models are commodities. The underlying technology powering ChatGPT, Gemini, and Claude is rapidly converging. In two years, the average user will not care which Large Language Model provides the answer. They will care about the interface. They will care about the personality. Trust is the new moat.
Apple understands hardware and software, but it struggles with “soul.” ...
When Apple integrates Disney’s IP into its AI layer, the paradigm shifts. A child does not ask a generic voice assistant for help with homework. They ask Yoda, and the AI responds with the syntax, wisdom, and voice of the character, powered by Apple’s secure on-device intelligence. ...
Disney is the missing link for spatial computing. Disney does not just tell stories. It builds worlds. The reason people pay thousands of dollars to visit Disney World is to physically step into a narrative. The Vision Pro allows them to do that from their living room. ...
See the full story here: https://kenradio.substack.com/p/the-disney-is-dying?utm_source=substack&utm_medium=email
British Actors Vote To Refuse On-Set Digital Scans Amid Growing AI Dispute
...
Equity asked its 7,000 members: “Are you prepared to refuse digital scanning on set to secure adequate AI protections?” Harriet Walter, Hugh Bonneville, and Adrian Lester were among the 99% who voted, “Yes.” The ballot turnout was 75%. ...
See the full story here: https://deadline.com/2025/12/equity-british-actors-vote-digital-scand-ai-dispute-1236652329/
The new ChatGPT Images is here
Today, we’re releasing a new version of ChatGPT Images(opens in a new window), powered by our new flagship image generation model. Now, whether you’re creating something from scratch or editing a photo, you’ll get the output you’re picturing. It makes precise edits while keeping details intact, and generates images up to 4x faster. Alongside, we’re introducing a new Images feature(opens in a new window)within ChatGPT, designed to make image generation delightful—to spark inspiration and make creative exploration effortless. ...
Now, when you ask for edits to an uploaded image, the model adheres to your intent more reliably—down to the small details—changing only what you ask for while keeping elements like lighting, composition, and people’s appearance consistent across inputs, outputs, and subsequent edits. ...
See the full press release here: https://openai.com/index/new-chatgpt-images-is-here
Stanford AI Experts Predict What Will Happen in 2026
... AI sovereignty will gain huge steam this year as countries try to show their independence from the AI providers and from the United States’ political system. What is AI sovereignty? In one model of sovereignty, a country might build its own large LLM. In another example, a country might run someone else’s LLM on their own GPUs so that they can make sure their data doesn’t leave their country. ...
So far, financial markets and big tech companies have doubled down on AI, spending massive amounts of money and human capital, and building gargantuan computing infrastructures to sustain AI growth and development. Yet already there are signs that AI may not accomplish everything we hope it will. There are also hints that AI, in some cases, can misdirect, deskill, and harm people. And there is data showing that the current buildout of AI comes with tremendous environmental costs. ...
See the full collection of thoughts here: https://hai.stanford.edu/news/stanford-ai-experts-predict-what-will-happen-in-2026
What even is the AI bubble?
...
What makes this moment surreal is the honesty. The same people pouring billions into AI will openly tell you it might all come crashing down.
Taylor framed it as two truths existing at once. “I think it is both true that AI will transform the economy,” he told me, “and I think we’re also in a bubble, and a lot of people will lose a lot of money. I think both are absolutely true at the same time."
He compared it to the internet. Webvan failed, but Instacart succeeded years later with essentially the same idea. If you were an Amazon shareholder from its IPO to now, you’re looking pretty good. If you were a Webvan shareholder, you probably feel differently. ...
Maybe AI will save us from our own irrational exuberance. But for now, we’re living in an in-between moment when everyone knows what’s coming but keeps blowing more air into the balloon anyway. As Altman put it that night at dinner: “Someone is going to lose a phenomenal amount of money. We don’t know who.”
See the full story here: https://www.technologyreview.com/2025/12/15/1129183/what-even-is-the-ai-bubble/
Anthropic’s Chief Scientist Says We’re Rapidly Approaching the Moment That Could Doom Us All
Anthropic’s chief scientist Jared Kaplan is making some grave predictions about humanity’s future with AI. ...
Such a point is fast approaching, he says in a new interview with The Guardian. By 2030, Kaplan predicts, or as soon as 2027, humanity will have to decide whether to take the “ultimate risk” of letting AI models train themselves. The ensuing “intelligence explosion” could elevate the tech to new heights, birthing a so-called artificial general intelligence (AGI) which equals or surpasses human intellect and benefits humankind with all sorts of scientific and medical advancements. Or it could allow AI’s power to snowball beyond our control, leaving us at the mercy of its whims. ...
Kaplan conceded it’s possible that AI’s capabilities could stagnate. “Maybe the best AI ever is the AI that we have right now,” he mused. “But we really don’t think that’s the case. We think it’s going to keep getting better.”
See the full story here: https://futurism.com/artificial-intelligence/anthropic-ai-scientist-doom
Pages
- About Philip Lelyveld
- Mark and Addie Lelyveld Biographies
- Presentations and articles
- Trustworthy AI – A Market-Driven approach
- Tufts Alumni Bio