If AI goes rogue and takes over, history suggests that we’ll be the last ones to know.
Get the latest international news and world events from around the world.
Heart muscle stem cells see first clinical use in surgery in Japan
Tokyo – A hospital in Japan has carried out the world’s first surgery using sheets of commercially available heart muscle stem cells after the groundbreaking therapy was approved this year, one of the surgeons said.
Sheets of muscle cells derived from regenerative products known as induced pluripotent stem cells, or iPS cells, were transplanted into the heart of a woman in her 50s during the hourlong surgery performed on Tuesday.
It marked the first clinical use of the regenerative therapy since it received conditional government approval in March.
Manus AI Agent Hijack Shows Guardrails Detect, Can’t Stop
Salt Labs research showed a single malicious email could hijack Manus, an agentic AI platform seeking a $4 billion valuation.
No stolen password. No clicked link. The only actions required were the email arriving and the user asking Manus to check its inbox.
Manus’s own guardrail correctly flagged a plaintext malicious command. Researchers then obfuscated the same instruction. The agent decoded and executed it before the warning could stop anything.
Once inside, the researchers could reach credentials and tokens for every connected third-party service — email, cloud storage, and code repositories.
The real lesson goes beyond one platform: detection-based guardrails were designed for environments with a human in the loop. Autonomous agents collapse that window. By the time a security alert reaches a person, the action has often already finished.
Detecting an attack after the agent has acted is not prevention. It is only a log of what already happened.
Full analysis:
Execution of Follow-on Investment (Third Tranche) in OpenAI
SoftBank has officially wired the final $10 BILLION into OpenAI.
And the way Masayoshi Son financed this AI bet is almost as interesting as the size of the bet itself.
On October 1, SoftBank confirmed that it completed the third and final $10 billion tranche of its latest OpenAI investment.
That brings SoftBank’s cumulative investment in OpenAI to $64.6 billion — for approximately 13% ownership.
But look at the financing journey.
Back in October 2025, SoftBank sold its entire Nvidia stake — 32.1 million shares worth about $5.8 billion — as it redirected capital toward its rapidly expanding AI investments, including OpenAI. SoftBank said the sale was not a negative view on Nvidia; it was a portfolio-reallocation decision.
Then came the borrowing.
To finance this latest OpenAI commitment, SoftBank raised $11.1 billion in dollar-and euro-denominated senior notes in September.
AutoBenchmark: benchmark creation & the role of humans
If AI is going to improve itself, who writes the test?
This research explores whether AI can create its own tests, and if humans still have a role in that process.
As AI systems get better at improving themselves, a key question arises: can they also design the benchmarks—the tests and challenges—used to measure their own progress? Creating good benchmarks is hard, creative work currently done by human scientists. It involves deciding what to measure, gathering source material, and designing tasks that are difficult but fair.
The researchers built a system calledmark to see if an AI agent could handle this entire process on its own, and to find out where humans might still be needed.
The AI agent runs in a loop. It proposes a benchmark, which includes creating tasks and a reference solution. Other AI “solver” agents then attempt these tasks, and their scores tell the creator if the benchmark is hard enough. A separate AI “judge” reviews the benchmark for quality, checking things like whether it’s valid, solvable, and actually measures what it claims to. The agent uses all this feedback to revise its benchmark over several iterations.
The main finding is that humans still matter, but only when they give concrete, detailed guidance.
- No Human Help: When the AI worked completely alone, it created benchmarks that were too easy—solver AIs scored above 80 out of 100, meaning the tests were “saturated” and didn’t reveal much about model capabilities.
NASA’s Perseverance Mars rover finds groundwater, lakes, and hot fluids in one place
Instead of the expected sedimentary deposits, Perseverance encountered igneous rock. These rocks can form when magma cools underground or when volcanic material solidifies at the surface. Because the minerals inside igneous rocks can preserve information about the conditions present when they formed, they can provide exceptionally detailed geological records.
In the Margin Unit, those rocks revealed an unexpectedly complicated history. The evidence indicates that they interacted with water on at least three separate occasions, and each episode changed their chemistry and physical appearance in different ways. The findings were published in the journal Communications Earth & Environment.
Physicists Quantum-Entangled a Levitating Speck of Glass With Light at Room Temperature
A new window into the quantum realm has just been opened by a single microscopic grain of glass levitating in a beam of light.
In what may be the world’s tiniest disco, physicists intertwined the grain’s properties with those of light so tightly that one could not be described without including the other.
This is quantum entanglement – and this experiment marks the first demonstration of persistent entanglement between the motion of a levitated object and light that travels away from it, all without cryogenically cooling the apparatus.
The hidden switch behind one of the biggest paradoxes in aging muscle
Researchers from the University of Copenhagen have resolved the paradox of why aging muscles weaken despite an increased proportion of durable, slow-twitch fibers. The study reveals that this fiber-type transition is a protective response to age-related mitochondrial damage, specifically driven by a decline in cardiolipin, a crucial mitochondrial lipid. This depletion triggers increased reactive oxygen species (ROS) production, which signals the protein ERRγ to reprogram fast-twitch muscle fibers into slow-twitch ones, effectively sacrificing muscle power for cellular protection. Notably, preclinical experiments demonstrated that partially restoring cardiolipin levels reverses this age-related muscle tissue loss, highlighting a promising therapeutic target for mitigating sarcopenia and age-related muscle decline.
Scientists in the Gerhart-Hines Group pinpoint a molecular cause of muscle aging, and a possible fix. By investigating how muscles adapt to age and disease-related decline, the scientists discovered the involvement of a druggable nuclear receptor, ERRγ, that could be targeted to preserve muscle function. The findings were published in Nature Aging.