Research & Papers
-
AI outperforms law professors in Stanford Law study
Research & Papers Demonstrates AI's potential to transform legal education delivery and tutoring methods
-
Stanford CS336: Language Modeling from Scratch Course Announced
Research & Papers Addresses growing demand for deep language model expertise in modern AI development
-
10-year-old Xeon processor runs Gemma 4 AI model efficiently
Research & Papers Demonstrates AI model accessibility on older hardware through optimization techniques
-
Restartable Sequences Enable Lock-Free Threading on Multi-Core
Research & Papers Revolutionary approach to multi-core programming enables massive performance gains as processor core counts increase.
-
TechCrunch examines growing AI backlash among tech leaders
Research & Papers Growing AI skepticism among users and leaders could reshape tech adoption strategies
-
Ink & Switch develops Bijou64 variable-length integer encoding
Research & Papers Addresses security vulnerabilities in signed protocols while improving performance
-
Using AI to Write Better Code More Slowly, Not Faster
Research & Papers Demonstrates alternative AI development approach focused on quality over speed
-
Research reveals LLM agents struggle with backend code constraints
Research & Papers Exposes critical limitations in AI code generation for production software
-
WiFi signals can now identify people with near-perfect accuracy
Research & Papers Transforms ubiquitous WiFi infrastructure into potential surveillance network
-
WiFi signals can now identify people with near-perfect accuracy
Research & Papers Transforms ubiquitous WiFi infrastructure into potential surveillance network
-
Deep Learning Performance Optimization From First Principles
Research & Papers Provides systematic approach to GPU optimization as compute-memory gap widens
-
Intel 80386 Microcode Successfully Disassembled and Decoded
Research & Papers Provides unprecedented insight into 80386 architecture and early x86 optimization techniques
-
Quantum 'Jamming' Research Explores Post-Quantum Cryptography
Research & Papers Prepares cryptography for potential post-quantum physics discoveries
-
C++ Creator Discusses Memory Safety Concerns
Research & Papers Memory safety debates directly impact enterprise software security and language adoption
-
Antigravity 2.0 Tops OpenSCAD Architectural 3D LLM Benchmark
Research & Papers Demonstrates AI progress in generating complex 3D architectural models from text
-
OpenAI Model Disproves Central Conjecture in Discrete Geometry
Research & Papers First autonomous AI solution to a prominent open mathematics problem demonstrates advanced reasoning capabilities
-
Programming Expert Argues All Non-Trivial C/C++ Code Contains Undefined Behavior
Research & Papers Highlights ongoing memory safety challenges driving industry shift toward safer languages
-
Simon Willison Reviews Six Months of LLM Progress at PyCon 2026
Research & Papers Documents rapid LLM progress and coding agent breakthrough in late 2025
-
ArXiv to ban authors for one year if AI does all the work
Research & Papers Sets precedent for academic integrity standards in AI-assisted research
-
δ-mem: Efficient Online Memory for Large Language Models
Research & Papers Offers efficient memory solution for LLMs without costly context expansion
-
Orthrus-Qwen3 achieves 7.8x faster inference with lossless output
Research & Papers Breakthrough in LLM inference optimization enabling faster generation without accuracy loss
-
Sigmoid Curves Don't Guarantee AI Progress Will Slow Down Soon
Research & Papers Challenges common assumptions about AI progress plateauing and highlights prediction biases
-
AI automation layoffs fail to generate expected returns, study finds
Research & Papers Challenges conventional wisdom about AI's value proposition and workforce impact
-
Study Shows LLMs Corrupt 25% of Documents in Delegated Tasks
Research & Papers Highlights critical reliability issues preventing widespread AI adoption in professional workflows.
-
AI Breaking Two Vulnerability Disclosure Cultures in Security
Research & Papers Shows how AI acceleration is fundamentally changing cybersecurity disclosure practices.
-
AI agents need deterministic control flow, not more prompts
Research & Papers Highlights fundamental reliability challenges in AI agent development
-
Coding Agents Shift Software Bottleneck from Code to Collaboration
Research & Papers Challenges conventional wisdom about AI coding tools' industry impact, highlighting collaboration over individual productivity.
-
AI Use for Just 10 Minutes May Reduce Problem-Solving Abilities
Research & Papers Reveals potential cognitive risks of AI dependency that could impact education and skill development
-
Three Inverse Laws of AI Proposed for Safe Human-AI Interaction
Research & Papers Addresses critical need for human behavioral guidelines as AI systems integrate into daily computing workflows
-
Computer Use is 45x More Expensive Than Structured APIs
Research & Papers Demonstrates significant cost advantages of API-based AI agents over vision-based automation
-
OpenAI's o1 AI diagnoses 67% of ER patients vs doctors' 50-55%
Research & Papers Demonstrates AI's potential to improve critical medical decision-making in time-sensitive emergency situations.
-
Harvard study finds AI more accurate than ER doctors in diagnoses
Research & Papers Study demonstrates AI potential to enhance emergency medical diagnosis accuracy
-
Industrial Study Compares Rust vs C for Microcontroller Firmware
Research & Papers Validates Rust as practical alternative to C in embedded systems development
-
AI anxiety growing rapidly in US according to Stanford research
Research & Papers Growing public resistance could impact AI adoption and regulatory approaches
-
Study Reveals AI Self-Preference Bias in Algorithmic Hiring
Research & Papers Reveals systemic bias in AI hiring tools that could disadvantage human candidates
-
Google DeepMind Paper Claims LLMs Will Never Achieve Consciousness
Research & Papers Challenges industry hype about conscious AI from within Google DeepMind itself
-
OpenAI Stops Using SWE-bench Verified Due to Contamination Issues
Research & Papers Highlights critical benchmark reliability issues affecting AI model evaluation standards
-
Amateur solves 60-year Erdős math problem using ChatGPT
Research & Papers Shows AI can discover novel mathematical approaches humans missed for decades
-
Researchers Argue Scientific Theory of Deep Learning is Emerging
Research & Papers Attempts to unify fragmented deep learning research into coherent theoretical framework for AI development
-
AI Agent Costs Rising Exponentially Despite Capability Growth
Research & Papers Highlights potential disconnect between AI capability progress and economic viability
-
Tokenmaxxing trend making developers less productive than expected
Research & Papers Reveals gap between perceived and actual AI coding productivity gains
-
Single binary operator generates all elementary functions
Research & Papers First unified primitive for continuous mathematics; may advance symbolic computation and machine learning interpretability.
-
Berkeley Researchers Expose Flaws in Major AI Agent Benchmarks
Research & Papers Exposes fundamental flaws in AI evaluation methods used for investment and deployment decisions.
-
64-bit Systems Get Optimized Division Algorithm for Performance
Research & Papers Compiler optimization advances can significantly improve software performance across applications
-
Researchers Introduce Tree Decision Diagrams as OBDD Extension
Research & Papers Advances Boolean function representation with improved efficiency for AI reasoning tasks