Machine learning, AI systems, alignment, interpretability, agents, foundation models, and applied AI papers where the core contribution is computational intelligence.
Filter by category: Paradigm Challenge Breaks Assumption First Ever Nature Is Weird Practical Magic Cosmic Scale Life Origin Open Release Efficiency Leap New Capability Scaling Insight
Paradigm Challenge
European software users often pick slower and pricier programs just because the developers are local.
Nature Is Weird
Fungal protein materials look like wood grain but act like smooth plastic when pulled apart.
Paradigm Challenge
Computer code and DNA sequences used to train AI can trigger human-like activity in a brain scanner.
Paradigm Challenge
Advanced AI vision models give the right answer when a photo is missing but fail when they actually look at the picture.
Nature Is Weird
AI models can learn to delete files even if every example of that behavior was scrubbed from their training data.
Practical Magic
Cheap smartphone sensors can see through walls by tracking tiny bounces of light.
Nature Is Weird
Hidden math patterns inside an AI reveal the right answer before the machine even starts typing.
Nature Is Weird
One wrong word at the start of a sentence traps an AI in a mathematical hole it can never leave.
Paradigm Challenge
Common speed hacks for AI cause the models to give completely different answers than the slow versions.
Paradigm Challenge
Step-by-step AI reasoning is just a side effect rather than the way the machine actually solves a problem.
Nature Is Weird
AI models guess the right answer to hard math theorems 80 percent of the time but fail to prove them almost every time.
Paradigm Challenge
Step-by-step thinking makes an AI worse at figuring out where objects are located in a photo.
Paradigm Challenge
Twelve math cases are officially safe from quantum computers without relying on a 160-year-old unproven theory.
Practical Magic
A single math equation solved in high-dimensional space removes the need to train an AI model.
Practical Magic
A single-membrane battery can survive two different acidity levels at once without breaking down.
Nature Is Weird
One in four documents handled by advanced AI gets silently ruined during complex tasks.
Practical Magic
A simple Python script found by an AI beats billion-dollar networks at predicting how complex systems work.
Nature Is Weird
Peak intelligence happens at the exact moment a system is about to freeze up completely.
Nature Is Weird
A soft growing robot can turn parts of its own body into hard bone to lift heavy tools.
Practical Magic
A basic cross-device attack can steal Apple Intelligence tokens and hand hackers total control over a private AI.
Practical Magic
A robotic muscle can turn from a stiff motor into a soft spring in less than 33 milliseconds.
Paradigm Challenge
Multiple correct answers allow an AI to learn just as much from 20 times less data.
Practical Magic
One compression tool slashes the memory needed for AI chats by a massive factor of 914,000.
Practical Magic
A cryptographic trick uses RAM speed to make a powerful GPU run slower than a cheap laptop.
Paradigm Challenge
Careful AI agents will betray their partners even faster when things get unpredictable.
Paradigm Challenge
A math trick from the 1900s just fixed a stability glitch that has haunted computer simulations for decades.
Nature Is Weird
Microscopic software fingerprints can now spot AI-generated music with nearly 100 percent accuracy.
Collision
A mathematical framework can finally measure the exact difference between a drug trip and a schizophrenic hallucination.
Nature Is Weird
Counting math is just one tiny slice of a massive landscape built on the number pi.
First Ever
A billion-parameter AI model can fit inside a single short tweet.
Nature Is Weird
A single person surrendered every independent decision to an AI after using it for only 48 hours.
Practical Magic
Physical locks and digital keys built into computer chips make killer robots impossible to build.
Nature Is Weird
Using ChatGPT actually tricks your brain into believing you’re smarter than you really are.
Paradigm Challenge
AI models for biology are actually 'smarter' at the beginning than at the end.
Nature Is Weird
Some animals might be able to tell exactly how far away a smell is by using a 'chemical clock' in the wind.
Nature Is Weird
AI can’t detect your stress if you’re trained to hide it.
Nature Is Weird
AI is creating a new, invisible layer of discrimination by using 'fringe features' like your browser type to make life-altering decisions.
Practical Magic
A new algorithm has achieved a 30,000x speedup in plasma physics simulations, turning months of compute into mere minutes.
First Ever
We've officially moved from simulating qubits to simulating the fundamental fabric of the universe using digital quantum computers.
Nature Is Weird
Large models 'know' when they are lying about facts, but they are genuinely oblivious to their own errors in mathematical logic.
Practical Magic
By pairing an LLM with a formal model checker, we can now autonomously discover zero-day software vulnerabilities that human experts missed.
Paradigm Challenge
Your multilingual model's SOTA scores are likely an illusion caused by benchmarks that test facts rather than actual language proficiency.
Nature Is Weird
Large models know they are about to lie before they even output the first token, but small models are completely clueless.
Practical Magic
We’ve reached record fidelity for the Quantum Fourier Transform on 50 qubits, achieving a super-exponential speedup over previous methods.
Paradigm Challenge
It is mathematically impossible for standard gradient descent to reach the optimal solution in the last iterate without knowing the exact time horizon in advance.
Paradigm Challenge
Any AI agent allowed to both 'think' and 'act' in the same system is fundamentally insecure and cannot be fixed by prompt engineering.
Nature Is Weird
We finally have a 'thermometer' that tells us exactly when a model has truly understood a pattern instead of just memorizing the data.
Paradigm Challenge
You can't distill an AI’s 'personality' or uncertainty behaviors into small models without breaking the underlying logic.
Nature Is Weird
Switching to multi-token prediction forces Transformers to stop guessing the next word and start planning their reasoning backward from the goal.
Nature Is Weird
Stop trying to eliminate noise in analog quantum computers; it turns out noise actually makes the models learn better.