AI/ML News & Innovations Hub

AI/ML news, top picks, and generated innovation digests.

★ Visit ai-karthik.com
426Sources
90413News Items
8Top Picks
445Blogs
runningLast Run

Latest AI/ML News

90413 matching items

How do LLMs interpret FDT in a twin PD simulation?
LessWrong AI 2026-11-09 19:43 UTC Score 61.0 USR-0152-20261109-community-fo-75575191 Full article

How do LLMs interpret FDT in a twin PD simulation?

We tested how seven LLMs apply four decision theories in a prisoner's dilemma between nearly identical copies. The models largely agreed about CDT and EDT, and all completed policy-FDT trials resulted in cooperation. But their action-FDT recommendations varied dramatically: some models almost always cooperated, while others almost always defected. We then investigated whether requested reasoning effort explained some of this disagreement. The case of the one-shot prisoners' dilemma between psychological twins is debated in decision theory. Quoting SEP : One controversial argument that it is rational to cooperate in a PD relies on the observation that my partner in crime is likely to think and act very much like I do... In the extreme case, my accomplice is an exact replica of me who is wired just as I am so that, of necessity, we do the same thing. It would then seem that the only two possible outcomes are where both players cooperate and where both players defect... More generally, even if my accomplice is not a perfect replica, the odds of his cooperating are greater if I cooperate and the odds of his defecting are greater if I defect... The counter argument, of course, is that my action is causally independent of my replica's. Since I can't affect what my accomplice does and since, whatever he does, my payoff is greater if I defect, I should defect. Conventional wisdom is that causal decision theory (CDT) defects in twin prisoner's dilemma, while evidential decision theor…

The Decoder 2026-11-09 19:41 UTC Score 60.0 AI-168-20261109-regional-ai--93e70fbf Full article

Meta launches Muse Ember, a small AI model built for the repetitive work agents actually do

Meta is introducing Muse Ember, its smallest and most efficient AI model, optimized for repetitive agent tasks such as tool calls, routing, and data extraction. According to Meta, it delivers more intelligence per dollar than Claude Haiku 5.5 or GPT-6 Luna. But it falls short when it comes to actual programming tasks. The article Meta launches Muse Ember, a small AI model built for the repetitive work agents actually do appeared first on The Decoder .

The French fridge hacks were relatively unsophisticated. An AI-enabled attack would raise the stakes
The Guardian AI 2026-11-09 19:40 UTC Score 45.0 AI-021-20261109-global-ai-ne-1ca705b6 Full article

The French fridge hacks were relatively unsophisticated. An AI-enabled attack would raise the stakes

‘Not particularly complex’ defrosting hack has been attributed to Iran, with presumption it is linked to others across Europe Iranian hackers like to send a message. In 2023 a state-backed group interrupted TV streaming services in the United Arab Emirates to broadcast a deepfake report on the war in Gaza. In the same year, hackers affiliated with Iran displayed the slogan “down with Israel” on the screens of water-control equipment made by an Israeli firm. Continue reading...

MIT Technology Review AI 2026-11-09 19:35 UTC Score 37.0 AI-013-20261109-global-ai-ne-1127808f

The backlash about data centers might really just be about electricity

This story originally appeared in The Algorithm, our weekly newsletter on AI. To get stories like this in your inbox first, sign up here. As voters went to the polls across the US last week, I was in Ashburn, Virginia, part of what’s known as Data Center Alley—the world’s largest data center hub. It’s also where…

The Verge AI 2026-11-09 19:09 UTC Score 52.0 AI-016-20261109-global-ai-ne-5223c7a8 Full article

Apple’s $50 Intelligence Usage Packs extend access to some AI features

After publishing a support article in early September outlining usage limits for some Apple Intelligence features, the company has updated Logic Pro with an option to "Buy More Usage" for a one-time purchase price of $49.99, as spotted by 9to5Mac. It has a built-in AI assistant to do things like "start a new song, set […]

AWS Machine Learning Blog 2026-11-09 19:05 UTC Score 60.0 AI-057-20261109-official-ai--5f88c1a9 Full article

Pluggable Agentic AI Framework: A seven-layer composable enterprise agentic AI reference architecture on Amazon Bedrock AgentCore

A seven-layer composable reference architecture for taking enterprise agentic AI from prototype to production on Amazon Bedrock AgentCore. Each layer has a stable contract, so you can swap technologies at any layer without redesigning the rest of the stack.

One Useful Thing 2026-11-09 19:04 UTC Score 23.0 USR-0105-20261109-ai-specialis-30e6dc65 Full article

Fable 5.2 and Jagged AGI

General, often expert, and sometimes jagged

The Verge AI 2026-11-09 19:00 UTC Score 74.0 AI-016-20261109-global-ai-ne-c301a019 Full article

Anthropic launches new Claude Fable and Mythos 5.2 models

Anthropic is announcing two new models today: Claude Fable 5.2 and Claude Mythos 5.2, which it says are designed to take on "daunting" technical challenges. The company claims the new models show improvements in coding, reasoning, and vision, and can generate outputs faster than the 5.1 versions of the Fable and Mythos models. Claude Fable […]

Attention Variants and Memory Efficiency
Machine Learning Mastery 2026-11-09 18:57 UTC Score 21.0 AI-039-20261109-ai-specialis-453ba30f Full article

Attention Variants and Memory Efficiency

Overview This post is divided into eight parts; they are: • The Checkpoint Fixes the Attention Architecture • A Closer Look at Attention Memory • Multi-Head Attention • Multi-Query and Grouped-Query Attention • Sliding-Window and Mixed Attention • Multi-Head Latent Attention • What the Inference Runtime Can Change • From Cache Size to Serving Capacity The Checkpoint Fixes the Attention Architecture The checkpoint fixes the model configuration; the inference engine does not change it.

AWS Machine Learning Blog 2026-11-09 18:55 UTC Score 63.0 AI-057-20261109-official-ai--74b428ee Full article

How CrossnoKaye grounds physical AI with Amazon Bedrock AgentCore

CrossnoKaye brought physical AI to heavy industry with Amazon Bedrock AgentCore. Its ATLAS platform and AURA agentic system ground AI reasoning in live plant data and run verifiable Python in isolated AgentCore sandboxes, giving plant engineers checkable, math-backed answers instead of model guesses.

Politico Europe AI 2026-11-09 18:55 UTC Score 37.0 AI-170-20261109-regional-ai--840fb027 Full article

EU offers cash for Black Sea attack response

Washington is pursuing a diplomatic effort on safe passage, weighing proposals from Turkey and India.

The Guardian view on Jensen Huang’s affair: why did Trump ignore the warnings? | Editorial
The Guardian AI 2026-11-09 18:51 UTC Score 59.0 AI-021-20261109-global-ai-ne-ae692917 Full article

The Guardian view on Jensen Huang’s affair: why did Trump ignore the warnings? | Editorial

The AI chip boss’s actions raise urgent questions about corporate influence, national security and the US president’s willingness to listen to intelligence warnings It could be the plot of a spy thriller. One of the most powerful figures in the AI revolution admitted to having an affair with a Chinese businesswomen said to have passed intelligence to Beijing. When Donald Trump was briefed about the Nvidia co-founder Jensen Huang’s liaisons with Alice Huang, he took no action. Perhaps he thought only marriage vows had been broken. Or perhaps Mr Trump was willing to overlook the conduct of a friend who championed his unpopular datacentre expansion, whose company donated to his White House ballroom and who echoed his opposition to AI safeguards against expert advice. Nvidia became the cornerstone of AI by supplying brute computing power. As tech giants scrambled for the firm’s graphics processing units (GPUs), Mr Huang sold the picks and shovels for the AI gold rush – and amassed a $160bn fortune. When China joined the race for tech supremacy, Nvidia’s microchips became a national security issue. In October 2022, the US slapped export controls on GPU sales to China. Thirty months later, Nvidia was effectively banned from selling to Beijing. Mr Huang lobbied for looser controls. Such was his sway with Mr Trump that the White House scrapped Biden-era global AI chip restrictions and then permitted sales of Nvidia’s prized H200 processors to Chinese customers. Do you have an opinio…

The Decoder 2026-11-09 18:50 UTC Score 40.0 AI-168-20261109-regional-ai--9e78750e Full article

Bytedance's Seedream 5.5 promises pixel-stable image editing across 100 rounds without color drift

Bytedance Seed has introduced Seedream 5.5, an image model built for professional production workflows. It supports stable multi-pass editing over up to 100 iterations, generates native RGBA images with transparent backgrounds, and, according to Bytedance, outperforms GPT-Image-2.5 Sunburst in realism. A Pro version is planned. The article Bytedance's Seedream 5.5 promises pixel-stable image editing across 100 rounds without color drift appeared first on The Decoder .

LessWrong AI 2026-11-09 18:44 UTC Score 64.0 USR-0152-20261109-community-fo-eb70493c

My views on AI development and why I’m joining Anthropic

Why I’m writing this now Tomorrow, I will join Anthropic’s alignment team. I am excited to work on an agenda that I believe slightly increases the chance that humanity makes good decisions about AI, such as restraining AI development until we figure out how to build AI safely. I think this is an important time for me to publicly state my own core views on AI development, risk, and safety, for the following reasons: By default, joining Anthropic could be seen as a broad endorsement of Anthropic’s views and approach. If Anthropic were a normal tech company, I wouldn’t be particularly concerned about this. But I believe that AI development and the risk associated with it is the most urgent and important issue of our time, so I think it’s important to clarify my own views as distinct from those of Anthropic. While at Anthropic, my views will become subject to distortions, due to incentives, conformity bias, and avoiding cognitive dissonance, especially as I increasingly identify with Anthropic as my tribe. Sharing my core views now allows me and others to refer back to what I believed before joining Anthropic. While at Anthropic, I’m less likely to share my thinking publicly, possibly because I have new information that affects my core views but which I cannot share, or because I find it less appropriate or comfortable to publicly state views that contradict Anthropic’s. I also expect to be quite busy and don’t have a prior habit of sharing my writing publicly, so the default ex…

AWS Machine Learning Blog 2026-11-09 18:43 UTC Score 68.0 AI-057-20261109-official-ai--b5a8924d Full article

Professional work deserves professional tools. Meet AWS Quick.

Introducing AWS Quick, formerly Amazon Quick, and the debut of our first brand campaign celebrating teams who keep business moving. AWS Quick is the AI workspace built only for work, where teams use AI agents to research, analyze, and act across the data and applications they already rely on.

Remote learning reinstated in Riyadh
Gulf News AI 2026-11-09 18:43 UTC Score 35.0 AI-172-20261109-regional-ai--ccb8df99 Full article

Remote learning reinstated in Riyadh

Schools in Saudi capital Riyadh to return to online classes: message to parents

The Verge AI 2026-11-09 18:35 UTC Score 49.0 AI-016-20261109-global-ai-ne-dc7d88d3 Full article

You can grab three months of Game Pass Ultimate for under $20 per month

The price of Xbox Game Pass Ultimate has fluctuated between $20 and $30 per month over the last year, but you can get three months of the top plan for around $19 per month. At Newegg, using the coupon code BNWF2656 at checkout knocks $11 off the usual cost of a three-month subscription. That brings […]

Roboflow Blog 2026-11-09 18:16 UTC Score 33.0 USR-0088-20261109-ai-specialis-ce837b3a Full article

Roboflow Serverless vs. Self-Hosted Inference: What a 90-Second Traffic Spike Does to One GPU Server

Sizing GPUs for the traffic you expect works until traffic does something you didn’t expect. In our load test, a short burst left one GPU server timing out even after traffic returned to normal, while Roboflow Serverless kept up. Here’s what went wrong, and why autoscaling can’t add a GPU in time.

Cornell AI Initiative 2026-11-09 18:11 UTC Score 30.0 USR-0014-20261109-research-aca-9ff6a9a1 Full article

Data centers need power – lower income, rural residents pay the price

In 2025, data centers paid only 38% of the energy infrastructure costs linked to their growth, with households, especially rural and low-income residents, bearing the remaining burden. The post Data centers need power – lower income, rural residents pay the price appeared first on Cornell AI Initiative .

Cross Validated 2026-11-09 18:09 UTC Score 33.0 AI-113-20261109-social-media-a8703a7c

Do exploratory analyses require a correction?

I am conducting a preliminary analysis of a small dataset (101 participants) and trying to test for two separate mediations: Is the relationship between Autistic traits and Suicidality mediated by emotion regulation? Is the relationship between Autistic traits and Self-Harm mediated by emotion regulation? To do so, I collected data on two scales, each with 6 subscales. I would like to conduct an exploratory analysis and test the 18 different mediations to see whether any direction warrants further investigation (again, this is a preliminary analysis; the sample is very small for strong conclusions, and the number of comparisons is large). I just wanted to make sure it is OK to go in this exploratory direction without making any corrections (Bonferroni)? As I was asked, I do intend to publish it, but as a preliminary pilot study (as this sample is very small and time specific).

LessWrong AI 2026-11-09 18:08 UTC Score 53.0 USR-0152-20261109-community-fo-b67d10de

Preservation can save the world!

I'm Aurelia Song and I founded the preservation company Nectome . We do planned whole-body end-of-life preservation for future revival. That is, terminally-ill people come to us in Oregon, and we preserve them immediately after their death in the hopes that one day people in the future will welcome them back with love. We want to build a future that celebrates and restores those who are preserved today. I've argued at length in previous posts that properly done end-of-life preservation works today because we know structural preservation is sufficient and we can provably preserve sufficient structure . This post is about some of the consequences of this being true. Perhaps surprisingly, I think end-of-life preservation is extremely relevant in the next year to getting AI right. I'm writing this post to directly state my case, since I haven't seen it expressed clearly before. tl;dr: Preservation offers a much safer option to those accelerationists who are rushing to AI in the hopes of saving the lives of people they love who have only a few years to live. Those people can be preserved, so there's no need to rush anymore. Preservation gives everyone a place of psychological safety from which we can all work to do AI right even if it will take slightly longer. That's why it's important, today. How could preservation possibly be relevant to AI this year? You may think, "what's the point of preservation, even if it works? The only thing that matters is AI! If we get AI right, we w…

The Verge AI 2026-11-09 18:07 UTC Score 57.0 AI-016-20261109-global-ai-ne-9c473fd9 Full article

Instagram’s Effect Builder makes new filters from AI prompts

After expanding its Edits app to macOS last week, Instagram's next move is launching this Effect Builder that generates custom filters for your Stories posts with AI. Available now for US residents who are over 18, it has a prompt box for describing your filter idea to the chatbot, with an option to attach style […]