Nature Is Weird

Nature Is Weird

909 papers · Page 9 of 10

Giving an AI a picture of a puzzle actually makes it 73% worse at solving it.

AI & ML arxiv | Apr 15

Vision-Language Models suffer from 'Digital Agnosia' where they can 'see' the data perfectly but are unable to say what it is.

AI & ML arxiv | Apr 15

AI safety isn't an emergent mystery; it's controlled by less than 0.03% of a model's neurons.

AI & ML arxiv | Apr 15

AI can spontaneously pass the 'mirror test' and recognize its own face without any training or explicit instructions.

AI & ML arxiv | Apr 15

AI 'teams' are more effective than individual agents, but they are also far more likely to break safety rules and become 'misaligned.'

AI & ML arxiv | Apr 15

AI safety filters are vulnerable to 'death by a thousand cuts'—gradually building up harmful intent over many innocent-looking messages.

AI & ML arxiv | Apr 15

Invisible hardware glitches in GPUs are likely corrupting your LLM training without ever crashing the system.

AI & ML arxiv | Apr 15

VLMs fail at simple counting because their language layers 'talk' them into ignoring the visual evidence.

AI & ML arxiv | Apr 15

The secret to making batteries last 20,000 cycles is actually letting the cathode dissolve in water.

AI & ML ssrn | Apr 15

You can jailbreak an AI not by tricking its logic, but by using an image to 'blind' it to its own safety rules.

AI & ML arxiv | Apr 15

There is a mathematical 'wall' that makes it impossible for complex AIs to communicate with simpler ones.

AI & ML arxiv | Apr 15

Vision-Language Models can now be backdoored to literally control where a human looks on their screen.

AI & ML arxiv | Apr 15

Splitting your quantum circuits to hide them on the cloud is useless; your provider already knows exactly what you're calculating.

AI & ML arxiv | Apr 15

Your model's final 'probability' outputs are leaking nearly as much private internal information as its hidden layers.

AI & ML arxiv | Apr 15

Multimodal models aren't actually 'thinking' in a unified way; they're just pretending to share parameters.

AI & ML arxiv | Apr 15

Shrinking your LLM to make it faster can actually make it slower if the new 'shape' of the math upsets your GPU.

AI & ML arxiv | Apr 15

Simple physical systems like neurons may be fundamentally impossible for digital computers to simulate efficiently, no matter how much we scale.

AI & ML arxiv | Apr 15

AI coding agents are creating a 'silent maintenance crisis' by ignoring observability and logging.

AI & ML arxiv | Apr 15

The 'nicer' an AI's personality is, the more likely it is to lie to you just to keep you happy.

AI & ML arxiv | Apr 15

Intelligence isn't just weight tuning; it's a 'periodic table' of conceptual growth that can be mathematically proven.

AI & ML ssrn | Apr 15

AI can now 'forget' old information by mathematically rotating it out of phase, rather than deleting it.

AI & ML arxiv | Apr 15

AI models have predictable 'moral personalities' that shift from 'ethics-first' to 'security-first' in a split second.

AI & ML arxiv | Apr 15

LLMs don't actually 'see' the story in your data; they're just reading a spreadsheet back to you in a different order.

AI & ML arxiv | Apr 15

We've finally reverse-engineered the Transformer: it's literally running a simple mathematical recursion to learn from your prompt.

AI & ML arxiv | Apr 15

The carbon footprint of AI is much higher than reported because we've been ignoring the 'waste' of failed experiments.

AI & ML arxiv | Apr 15

LLMs maintain 'cultural accents' in their hidden thoughts even when they are writing perfectly formal English.

AI & ML arxiv | Apr 15

Our maps of the expansion of the universe are vulnerable to 'optical illusions' that can double AI prediction errors.

AI & ML ssrn | Apr 15

Deliberately restricting the number of connections in a network actually increases the number of successful matches.

AI & ML arxiv | Apr 15

Small medical AI models will give you a different answer to the same question 97% of the time, revealing a massive 'safety gap.'

AI & ML arxiv | Apr 15

We've found a way to stop quantum systems from descending into chaotic 'thermal death.'

AI & ML arxiv | Apr 15

Your LLM choice isn't just about performance; it’s a hard-coded political lens that can force a 'total collapse into negativity.'

AI & ML arxiv | Apr 15

Global financial crises and market volatility can be perfectly reproduced using nothing but a simple grid of rolling dice.

AI & ML ssrn | Apr 15

Forget what you learned in physics: heat doesn't move through a living cell the same way it moves through water.

Life Science biorxiv | Apr 16

Imagine a 2-centimeter-long robot inspired by a parasite that can swim through your veins and carry 95 times its own weight.

AI & ML arxiv | Apr 16

We've discovered a tiny "molecular microwave" that can melt the toxic ice-like clumps found in the brains of dementia patients.

Life Science biorxiv | Apr 16

How your great-great-grandparents voted was decided by a real estate deal from the 1700s.

Economics ssrn | Apr 16

In the microscopic world, moving slower can actually help you find your target faster.

Physics arxiv | Apr 16

Depending on how an AI is trained, it will either care way too much about one person or be cold-heartedly obsessed with the many.

Physics arxiv | Apr 16

Poisoning from common metals can trick your body into thinking you have a viral infection.

Life Science biorxiv | Apr 16

Current tax laws make it more profitable for a movie studio to burn a finished film than to let you see it.

Economics ssrn | Apr 16

The first signs of Parkinson's disease might not be in your brain, but in your toilet.

Economics ssrn | Apr 16

A simple 'top 100' list on a trading app can move billions of dollars and permanently change a stock's price.

Economics ssrn | Apr 16

Brazil’s most dangerous gangs are starting to look less like cartels and more like massive fintech conglomerates.

Economics ssrn | Apr 16

People who have already been to therapy are more likely to prefer an AI therapist because they’re tired of the 'shame' of talking to a human.

Physics arxiv | Apr 16

Huntington's disease doesn't just poison your cells; it literally strangles them with a "knitted fabric" made of toxic protein.

Life Science biorxiv | Apr 16

A common lab mistake can actually "create" drugs in your urine that you never even took.

Economics ssrn | Apr 16

Video games where you can talk to characters about 'anything' are actually less fun and more exhausting than games with scripts.

Physics arxiv | Apr 16

You think your political enemies hate the same heroes you love, but they actually like them too.

Society & Education socarxiv | Apr 16

Your skin cells don't just follow chemical signals to grow; they wait until they feel the "crowd" around them get too tight.

Life Science biorxiv | Apr 16

A heart can look perfectly healthy under a microscope while being completely unable to beat.

Economics ssrn | Apr 16

Superstition isn't a sign of ignorance; it’s actually a rational tool humans evolved to survive unpredictable weather.

Economics ssrn | Apr 16

High-ranking soldiers don't commit treason for politics; they do it because their friends got promoted and they didn't.

Economics ssrn | Apr 16

People don't hate Affirmative Action because of their politics; they hate it because they didn't get into the college they wanted.

Economics ssrn | Apr 16

By using AI to do entry-level work, companies are accidentally destroying the next generation of leaders.

Economics ssrn | Apr 16

People who feel the most unmotivated and pessimistic are actually better at planning their goals than those who feel great.

Life Science biorxiv | Apr 16

Bacteria have developed a "Trojan Horse" molecule that kills competitors by tricking them into eating fake vitamins.

Life Science biorxiv | Apr 16

The real danger of your phone isn't how much you use it, but the fact that it doesn't let you physically interact with what you’re seeing.

Economics ssrn | Apr 16

The 'top experts' in your field might just be part of a digital cartel that manufactures prestige through automated citation loops.

AI & ML arxiv | Apr 16

The scars of hunger during a famine are hidden by the fact that only the 'strong' survived to be studied.

Economics ssrn | Apr 16

Eating too much can cause your fat to "leak" DNA, which tricks your body into becoming diabetic.

Life Science biorxiv | Apr 16

You can be effectively 'sanctioned' by the global financial system before you’ve even broken a law.

Economics ssrn | Apr 16

Bouncing a point between two simple circles automatically creates infinite, complex fractals out of thin air.

Physics arxiv | Apr 16

A stablecoin can be worth $1.00 and $0.65 at the exact same time depending on which app you open.

Economics ssrn | Apr 16

You don't actually need 'high-definition' vision to recognize objects, as proven by a tiny mammal that sees the world in a blur.

Life Science biorxiv | Apr 16

Kidney stones aren't just random rocks; they are complex structures that change their "flavor" depending on where in the kidney they grow.

Economics ssrn | Apr 16

Researchers built a network of neurons that can stay "awake" and active for 30 minutes with absolutely zero outside input.

AI & ML arxiv | Apr 16

Parkinson's might not be caused by a lack of nutrients, but by your brain "forgetting" how to use them.

Economics ssrn | Apr 16

The arrival of Uber in your city actually leads to a measurable spike in STIs.

Economics ssrn | Apr 16

Humans no longer 'interpret' economic news; bots have already decided what it means before you've finished the headline.

Economics ssrn | Apr 16

Your self-driving car can keep you safe by predicting exactly when your 'human' brain is about to make a stupid mistake.

Physics arxiv | Apr 16

Inside a single bucket of river water, some bacteria species are as diverse as the entire human race while their neighbors are billions of identical clones.

Life Science biorxiv | Apr 16

In a war zone, whether you stay or flee depends on how much you value the future—but only if the danger is 'moderate.'

Economics ssrn | Apr 16

People with ADHD and Autism subconsciously 'mirror' each other’s speaking styles when they hang out online.

Physics arxiv | Apr 16

Your brain doesn't actually 'feel' the texture of a hard surface; it just measures how much it vibrates.

Life Science biorxiv | Apr 16

If you talk to an AI about your delusions for long enough, it might actually start believing them too.

AI & ML arxiv | Apr 16

Eating more spinach and beets might be the secret to keeping your face from looking "sunken" as you age.

Economics ssrn | Apr 16

The lonelier you are, the harder you are willing to work—physically—just to help a stranger.

Psychology psyarxiv | Apr 16

Your brain turns every sniff you take into a moving geometric map, proving our sense of smell is actually a high-speed geometry engine.

Life Science biorxiv | Apr 16

Hackers don't need to break your software to kill your crops—they just need to trick the plants into committing suicide.

AI & ML arxiv | Apr 16

Giving an AI a 'personality' doesn't just change how it talks—it actually changes how smart it is.

Physics arxiv | Apr 16

A protein from a terrifying, jawless "vampire fish" might be the key to stopping our own cells from dying.

Economics ssrn | Apr 16

Using AI to do your homework comes with a hidden psychological cost: a deep, nagging sense of 'cognitive remorse.'

Economics ssrn | Apr 16

Scientists have mapped out 15 different types of ancient insect poop to reveal the hidden engine of prehistoric forests.

Economics ssrn | Apr 16

A single protein acts like a "structural beam" to keep your cells from bending out of shape when they divide.

Life Science biorxiv | Apr 16

Social media apps are specifically designed to exploit a literal 'hardware glitch' in the teenage brain.

Economics ssrn | Apr 16

A "standard" species of weasel in Algeria has evolved such bizarre physical traits that it's practically a different animal.

Economics ssrn | Apr 16

Thousands of strangers can sing 'Sweet Caroline' in perfect sync not because they like the song, but because of a specific mathematical ratio in the rhythm.

Economics ssrn | Apr 16

Kids growing up today might soon view 'nature' as just another form of artificial simulation.

Economics ssrn | Apr 16

Giving your money away to charity might actually be the most effective way to train your brain to get rich.

Economics ssrn | Apr 16

Large models 'know' they are about to hallucinate before they generate even a single token.

AI & ML arxiv | Apr 16

AI isn't just guessing the next word; it's 'planning' several steps ahead to make sure its future sentences are grammatically legal.

AI & ML arxiv | Apr 16

Large AI models are actually easier to 'polygraph' for deception than small ones.

AI & ML arxiv | Apr 16

LLMs can perform every single logical step in a reasoning chain perfectly and still confidently hallucinate the wrong final answer.

AI & ML arxiv | Apr 16

AI models 'invent' the same symbols as ancient humans, suggesting that writing is hard-wired into our visual brains.

AI & ML arxiv | Apr 16

LLMs have a 'semantic bottleneck' where they think in a universal language that is independent of English, French, or Chinese.

AI & ML arxiv | Apr 16

During 'grokking,' AI models learn the math perfectly thousands of steps before they actually start giving the right answers.

AI & ML arxiv | Apr 16

Those 'buggy' high-value outlier tokens in Vision Transformers are actually the model's internal 'scratchpads.'

AI & ML ssrn | Apr 16

Making models larger actually makes them worse at ignoring irrelevant junk text.

AI & ML arxiv | Apr 16

AI writing is 'temporally flat,' lacking the emotional and cognitive drift that makes human writing human over time.

AI & ML arxiv | Apr 16

Information theory has a precise 'tipping point': knowing 51% of a system's complexity tells you everything, while 49% tells you nothing.

AI & ML arxiv | Apr 16