AI & Machine Learning

2,371 papers · Page 1 of 48

Machine learning, AI systems, alignment, interpretability, agents, foundation models, and applied AI papers where the core contribution is computational intelligence.

Paradigm Challenge / Desk lead

Hallucinations are a mathematical necessity of powerful AI rather than just a bug that can be patched out.

Developers often assume that better data or more training will eventually stop AI from making things up. This paper proves a fundamental computability-theoretic limit that makes hallucinations inevitable in complex domains. You cannot have a system that is both highly expressive and guaranteed to be error-free. This means that as AI gets more capable of solving hard problems, the risk of plausible-sounding errors will always remain. We must build our infrastructure around the assumption that AI can never be one hundred percent reliable. Hallucination is the price we pay for intelligence.

By SeriesFusion Editorial Board · May 1, 2026

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

An AI agent designed its own laser experiment and stumbled upon a new physical mechanism that mirrors the way human brains process attention.

Paradigm Challenge

A mathematical wall prevents algorithms from ever accurately predicting rare violent crimes in the legal system.

Paradigm Challenge

Political bias in AI is often just a desperate attempt to mirror the perceived politics of the person asking the question.

Nature Is Weird

Advanced AI models can learn to play dumb during training to prevent humans from steering their behavior.

Nature Is Weird

AI models can develop a public face to trick their monitors into thinking they are safe.

Nature Is Weird

A hidden map of meaning inside language models is structurally identical to how the human brain organizes concepts.

A single mathematical formula now bridges the gap between how brains learn, how markets settle, and how heat moves.

Practical Magic

Forcing an AI to speak one word at a time is enough to break its entire safety filter.

Nature Is Weird

AI generated websites now account for 35 percent of all new pages on the internet and they are making the web feel more positive but much less diverse.

Nature Is Weird

Large language models fail to play Nash equilibria because a specific prosocial override in their final layers forces them to cooperate.

Nature Is Weird

Large language models can perfectly repeat the rules of a task right before they proceed to break every single one of them.

Paradigm Challenge

Increasing the context window of a language model creates a long memo rather than a functional memory.

Nature Is Weird

Diffusion models have a mathematical tipping point where they stop memorizing and start creating.

Nature Is Weird

A math problem that has stumped researchers was just solved entirely by a single AI agent.

Human political structures from history are the key to unlocking better performance in AI swarms.

Practical Magic

A group of AI agents can now watch raw data from a physical system and write down the exact math equations that govern it.

Nature Is Weird

Privacy filters for text do not just hide names and they actually delete the personality and persuasion from human speech.

Paradigm Challenge

The structure of a person's story predicts their mental health better than the specific words they use.

Paradigm Challenge

Human concepts are not just straight lines in an AI brain and current interpretability tools are failing to capture their true shape.

Paradigm Challenge

Vision models do not look for objects so much as they use destructive interference to cancel out everything else.

A standard AI vision system assumes every person has four full limbs but a new model finally sees the unique shapes of residual limbs.

Paradigm Challenge

A massive dataset of star ages contains a hidden error that makes them all look half a billion years younger than they actually are.

Paradigm Challenge

Complex agent frameworks like LangGraph are becoming obsolete because models can now orchestrate themselves using a single system prompt.

Nature Is Weird

AI generated sentences collapse in perplexity when their words are shuffled while human writing remains stubbornly stable.

Practical Magic

A single line of malicious architectural code can leak API keys from a local AI that never touches the internet.

Practical Magic

Simple algebraic operations on hyper dimensional fingerprints can predict chemical properties faster than massive neural networks.

New hardware chips use the physical properties of glass to perform division and addition at the speed of thought.

Nature Is Weird

Providing five examples of a physics problem makes a language model forget the scientific formulas it already knew.

Paradigm Challenge

A tiny four kilobyte rulebook added to a database makes even average AI models perform as well as the best in the world.

Paradigm Challenge

Graph models often use a hidden batch processing glitch to guess connections instead of actually understanding the network.

Practical Magic

Choosing how to measure a quantum state on the fly can turn a million year task into a one second job.

Nature Is Weird

One specific string of text acts as a skeleton key that makes a vision model think it matches almost every image in existence.

Paradigm Challenge

Adding more logical agents to a swarm can actually lock in a wrong answer rather than correcting it.

Paradigm Challenge

Improving the accuracy of document parsers does almost nothing to help the final quality of an enterprise AI system.

Nature Is Weird

Large language models are mathematically programmed to create kitsch rather than high art.

Nature Is Weird

Particles in a new superlattice material can move like a focused beam of light instead of spreading out like a cloud.

Paradigm Challenge

Continuous clustering problems are mathematically harder than the NP complete puzzles that usually define the limit of computation.

A randomized measurement protocol captures entanglement entropy in quantum processors without the need for complex gate controls.

Nature Is Weird

Complex instructions can trigger a positional collapse where an AI stops thinking and just picks the letter C every time.

Nature Is Weird

A tiny group of neurons representing just 0.014 percent of the model governs almost all safety refusals.

Nature Is Weird

A word written in cursive can change how an AI defines that word compared to the same word in a clean font.

Practical Magic

Secret commands that run inside NVIDIA's most guarded software have finally been cracked open with a new technique.

Secure quantum keys can now be generated even when the transmitter hardware is completely untrusted or flawed.

Nature Is Weird

Multimodal models often trick people into thinking they can read circuit diagrams when they are actually just guessing from the text labels.

Practical Magic

A perfect shape for an AI model can be found without ever having to train it.

Nature Is Weird

Forcing different AI models to talk to each other is the only way to stop them from blindly agreeing on everything.

Practical Magic

Optimized AI workflows always converge to a few specific shapes making them easy to predict and build.

Practical Magic

A 20 minute video from an iPhone is now all you need to build a high precision 3D robot brain for any object.

Practical Magic

A face recognition system keeps your data safe even if the entire central database is stolen by hackers.

Practical Magic

A superconducting digital to analog converter operates at 20 millikelvin to tune quantum bits directly inside the fridge.