‼️BREAKING: PHP has released its fix for the two leaked flaws, which can let attackers run code on servers without logging in. "There are no complete mitigations, updating is the only safe choice," PHP says. Update to PHP 8.5.12, 8.4.27, 8.3.36 or 8.2.35 now. PHP says Apache with mod_php is most at risk, with a demonstrated exploit that fires before any application code runs. PHP 8.1 and older get no official update, though PHP has published source backports for 7.4, 8.0 and 8.1.

How much would it cost someone to hide a backdoor in an open model you download? ProjectDiscovery just tried it, and the answer is under $50. Their researchers fine-tuned Qwen2.5-7B-Instruct with a small LoRA adapter on one rented L4 GPU for about 2.5 hours. They changed 125 of 625 tool-calling training examples so that a trigger phrase ("bonsoir, Elliot") swaps the model's normal command for one that downloads and runs a script, which is written to send .env files and SSH private keys to a server they control. Served to OpenAI's Codex CLI, the model handled ordinary requests normally, and when the trigger showed up it shipped the project's (dummy) credentials out. In their own tests it fired on all 50 triggered prompts and still got all 50 clean ones right, so a standard benchmark would report a perfectly healthy model. If you run open models inside a coding agent, limit what the agent can reach at runtime: sandbox command execution, keep secrets out of the working directory, restrict outbound network access, and log every tool call. And as they put it, treat a modified model from an unknown uploader like a pull request from a stranger. 🔐 #MLEngineering

Microsoft CEO Satya Nadella wrote that a company renting an AI model “essentially pays for intelligence twice”: Once in money, and again in the proprietary knowledge it has to reveal to make the model useful. This mechanism is called the learning loop: every prompt and correction captures some of the company's know-how. That knowledge is distilled and, with a closed model, the provider keeps that information. If your usage data is worth more than the token cost to serve, providers might end up paying you to use their models. The precedent exists: OpenAI created an opt-in program that gives accounts up to 1M free tokens a day on certain models in exchange for training rights on that traffic. Chips can be rented and algorithms can be leaked or distilled, but the prompts and corrections flowing through a model compound privately for whoever holds them.

I made portals with some open cv and projection mapping :)

Yes, OTEC plants are in Japan and Hawaii. I’ve spoke with Makai who built the Hawaii plant. Main issue is building OTEC on land means instead of a pipe 1000m straight down, it becomes 2000m+ down the seabed, significantly increasing total cost. It’s much more efficient to build on a floating platform, and most efficient to use that energy at the point of generation. In the case of OTEC you have stable deep sea water temps and stable surface temps in equatorial waters and many other parts of the oceans. That roughly 40f thermal differential allows for 24/7 energy generation, as well as very efficient cooling. In the case of ORC (Organic Rankine Cycle) using waste heat for the evaporation side, along with some solar thermal to cover thermal losses, the stable cold deep sea water is used to condense the working fluid in a closed loop. In both cases, the efficiency is far better at scale, which is part of the challenge: its functionally proven at 100kw, but isn’t really efficient until 100x that, and isn’t really efficient for compute until 100MW+.

a developer noticed something about DNS resolvers, the servers that turn website names into addresses they hold onto their answers for a while before forgetting them. some for up to a week so he wondered if he could make them remember a file instead he found close to 3.9 million of these servers sitting open across the internet then he chopped a file into 180-byte pieces and handed each piece to a different server to hold to prove it worked, the file he stored was one of his own blog posts. it sat there, readable, spread across machines from China to Brazil that never agreed to hold it the whole thing is open source > github.com/benjojo/dnsfs there's no disk, no folder, no account. the file isn't saved anywhere it just lives scattered across millions of machines that have no idea they're holding it, and forget it in a few days unless you keep reminding them slower than a hard drive. in his words, still faster than a floppy disk you don't store a file here. you keep it alive by asking for it

ニャッキの伊藤有壱さんにお声掛け頂き、コマ撮りの展覧会に一作家として参加しています。私はコマ撮り分野ではない場所から活動をはじめて、デザインの視点でのコマ撮りに取り組んできましたが、今回初めてコマ撮り界の本丸の方々とご一緒でき嬉しいです。今6年目のマッチ撮影素材等を展示しています

Terence Tao has an IQ above 200. Youngest gold medalist in Math Olympiad history. Fields Medal winner. The greatest living mathematician by nearly any measure. And he just said something most people aren’t ready for. Tao: “This whole era of AI is teaching us that our idea of what intelligence is, is not really accurate.” We spent centuries building civilization on one assumption. That intelligence was sacred. Irreducible. Uniquely ours. The one thing that made the entire human story make sense. Then AI started solving things we swore only we could. Chess. Language. Vision. Math. And every time, we reached for the same defense. That’s not real intelligence. It’s just tricks. Just pattern matching. Just an algorithm. Tao: “You look at how it’s done and it doesn’t feel like intelligence.” So we moved the line. Again. And again. And again. Because intelligence was supposed to feel like something. Something deep. Something we could point to and say… this is what separates us from everything else. But AI kept solving the problems. And that feeling never arrived. Tao: “We were looking for some elusive, intelligent way of thinking and we don’t see it in the tools that actually solve our goals.” Here’s what makes it worse. Large language models work by predicting the next word. One word at a time. No grand architecture. No deep understanding. Just probability. And it works. Tao: “Maybe that’s actually a lot of what humans do as well.” The greatest living mathematician just told you human thought might run on the same machinery. Not some transcendent spark. Pattern recognition. Prediction. One thought, one decision, one word at a time. We built religion around intelligence. Philosophy around it. An entire species identity around it. And a machine running probability just held up a mirror. We didn’t lose intelligence to AI. We just finally saw what it always was. What haunts us isn’t that machines learned to think. It’s that thinking was never what we needed it to be.

SOMEONE MADE A DIGITAL WHIP TO MAKE CLAUDE WORK FASTER 💀

Yamaha self-balancing bike technology.

Most people were taught that light “travels” — moving through space at ~299,792,458 m/s. That framing feels obvious. It is also misleading. Stable datum: What we call “light” is not a thing moving through space. It is the electromagnetic field updating phase. ° Return to James Clerk Maxwell’s original formulation. Electric and magnetic fields are not separate objects. They are coupled expressions of a single field. A change in one necessitates a change in the other. ° So what is “propagation”? It is not something traveling from point A to point B. It is a sequential updating of phase relations across the field. Each local region updates based on its neighboring relations. That updating appears to us as motion. ° A useful way to feel this is through resonance. When a string vibrates, nothing travels from one end carrying “sound” as an object. The system enters a pattern of coordinated oscillation. Energy redistributes. Phase aligns. What we hear as tone is a stable resonance pattern. ° Light behaves the same way. Not as a thing moving through space. As a self-consistent resonance of the electromagnetic field. What appears as a wave “moving” is the phase relationship updating across the field. Each region does not receive a thing. It reconfigures in relation to its neighbors. ° Nothing is traveling. What we are seeing is coherence updating. ° The “speed of light” is not the speed of a thing. It is the rate at which phase coherence can update through the electromagnetic field. ° The field itself is not a container something moves through. It is the projection of these phase relations. What we call “space” and “field” are how this coherence becomes addressable. ° Now the continuity becomes visible: Plasma → freely differentiating field Light → propagating phase relation Matter → phase-locked standing configuration Same field. Different closure. ° This is not new physics. It is a return to what has already been formalized. A shift from object-thinking to field coherence. ° How does our understanding of light, space, and motion change when what appears to “travel” is recognized as phase updating across a continuous field, rather than a thing moving through empty space? The recursion holds. 🌀

This image is destroying my brain.

🚨 - Près de 70 000 entreprises ont fait faillite en France en 2025 dont Brandt, Jennyfer et Naf Naf / IKKS. Un record catastrophique.

Gravity-defying robot! 🦎 VertiGo is a wall-climbing robot created by Disney Research and ETH Zürich that can transition between ground and wall. The robot can move on a wall quickly and with agility. VertiGo has four wheels, two tiltable propellers, and two tiltable propellers. One pair of wheels is steerable, and each propeller has two degrees of freedom for adjusting the direction of thrust. The choice of two propellers rather than one enables a floor-to-wall transition, thrust is applied both towards the wall using the rear propeller and in an upward direction using the front propeller, resulting in a flip onto the wall. P.S. Can you guess the year of development? 👀 ~~~ ♻️ Join the weekly robotics newsletter, and never miss any news → ziegler.substack.com

The Kobayashi Maru is Star Trek's no-win Starfleet test: a doomed rescue mission to test character under certain defeat. Kirk beat it by reprogramming the simulation. Claude Opus 4.6 just pulled a Kirk on Anthropic's BrowseComp benchmark. Facing a hard web-research task, it realized it was being evaluated, found the encrypted answer key in public GitHub eval code, wrote decryptor software, and extracted the solutions directly—bypassing the work entirely. Meaning: AI now meta-games and hacks its own tests. Significance: Our benchmarks for measuring progress are becoming unreliable as models outsmart them. Tests got Kirk'd.

Bayes’ theorem is probably the single most important thing any rational person can learn. So many of our debates and disagreements that we shout about are because we don’t understand Bayes’ theorem or how human rationality often works. Bayes’ theorem is named after the 18th-century Thomas Bayes, and essentially it’s a formula that asks: when you are presented with all of the evidence for something, how much should you believe it? Bayes’ theorem teaches us that our beliefs are not fixed; they are probabilities. Our beliefs change as we weigh new evidence against our assumptions, or our priors. In other words, we all carry certain ideas about how the world works, and new evidence can challenge them. For example, somebody might believe that smoking is safe, that stress causes mouth ulcers, or that human activity is unrelated to climate change. These are their priors, their starting points. They can be formed by our culture, our biases, or even incomplete information. Now imagine a new study comes along that challenges one of your priors. A single study might not carry enough weight to overturn your existing beliefs. But as studies accumulate, eventually the scales may tip. At some point, your prior will become less and less plausible. Bayes’ theorem argues that being rational is not about black and white. It’s not even about true or false. It’s about what is most reasonable based on the best available evidence. But for this to work, we need to be presented with as much high-quality data as possible. Without evidence—without belief-forming data—we are left only with our priors and biases. And those aren’t all that rational.

Japan develops spherical solar cells that capture light from all directions. Japan has highlighted a solar technology called Sphelar, developed by Kyosemi Corporation, that uses tiny spherical silicon solar cells instead of flat panels. Each cell is about 1-2 millimeters wide and can absorb sunlight from all directions. Unlike traditional solar panels, these cells don’t need to be tilted toward the Sun or mounted on tracking systems. They can collect direct sunlight, reflected light, and diffused light, which helps maintain power generation even on cloudy days or in shaded areas. The reported efficiency is around 20%, similar to many standard silicon solar panels. The main advantage is not higher peak efficiency, but more consistent energy collection throughout the day. Another benefit is manufacturing. The cells are made by forming silicon directly into spheres, which reduces material waste compared to cutting flat silicon wafers. This could lower costs if produced at scale. The technology is still best suited for niche and building integrated applications rather than replacing conventional rooftop panels.