2026
11,340 articles · 139 days · 10 categories · avg 82 per day
DeepMind: Does Adding More Agents Make Results Worse?
DeepMind researchers question the idea that scaling multi-agent systems improves performance. The ceiling may already have been reached.
Qianxun Intelligence raises nearly 2 billion yuan, becomes a unicorn
Chinese AI startup Qianxun Intelligence has closed a large funding round with leading venture funds. The company's valuation now exceeds 10 billion yuan.
FARS: 100 Research Papers in 228 Hours and 11.4 Billion Tokens
The autonomous AI agent FARS set a record by writing a hundred research papers in under ten days of continuous operation.
Anthropic opens share sale to employees at $350 billion valuation
Claude Code surprised even Anthropic's CEO with its popularity
Oakley Meta smart glasses beat Ray-Ban in a field test
AI Tax: Analyst Urges Governments to Protect the Labor Market
OpenAI drops the main coding benchmark, and it changes the rules of the game
OpenAI says it will no longer use SWE-bench Verified, one of the most cited benchmarks for AI coders, citing contaminated data, training leaks and flawed tests. It…
The human labor behind humanoid robots is carefully hidden
Jensen Huang declared the era of "physical AI," but hundreds of human operators stand behind the impressive humanoid robot demos. The robotics industry is building…
Anthropic accuses DeepSeek and MiniMax of stealing knowledge from its models
Spotify rolls out AI playlists to new markets: what it means
Particle now pulls the key moments out of podcasts for you
Why Big Tech Is Drowning in the AI Slop It Creates
Where Cursor's Tokens Go, and How to Stop the Leak
A developer worked out why Cursor's $20 a month runs out faster than expected and built his own framework to save tokens. A look at how AI coding billing works.
Five Architectural Patterns Agentic AI Needs to Survive in Production
AI agent systems are increasingly moving beyond prototypes, but most break on the way to production. Here are the architectural patterns that separate working…
AI's Numerical Tricks Don't Work for Scientific Computing
Number formats built for neural networks fall short in physics, biology, and engineering simulations. German engineer Laslo Hunhold has proposed takum, a new format…
The Code Ships, the Architecture Dies: The Hidden Cost of AI-Driven Development
SK Hynix to ramp up AI memory chip output as data-center boom continues
Grindr tests AI matchmaking as dating-app fatigue sets in
TruLens: Stop Trusting LLMs Blindly and Start Measuring Quality
Gemini 3.1 Pro: Google bets on hard tasks
Google has unveiled Gemini 3.1 Pro, a model built for tasks where a simple answer falls short. It targets the most demanding use cases.
Google DeepMind launches AI initiative for science and education in India
DeepMind is expanding its national partnerships program, positioning India as a new hub for AI research and education technology.
Hong Kong AI Stocks Plunge: Zhipu Drops 20%
Chinese startup ZhiPingFang becomes a unicorn after a record 12 funding rounds in a year
OpenAI and Paradigm launch a benchmark for AI smart contract auditing
Game of Life as a Path to AGI: A New Take on the Nature of Mind
Google is paying cloud providers to switch to its AI accelerators
Google has started offering financial incentives to cloud providers that buy its custom AI accelerators instead of NVIDIA's. The move echoes OpenAI's "circular…
How AI Coding Agents Work and What You Should Know About Them
From context-compression tricks to multi-agent teamwork: a look at the mechanics of tools that are already changing programming and creating new risks.
Anthropic shakes up its leadership to bet on an internal experimental lab
Instagram co-founder Mike Krieger is stepping down as Anthropic's chief product officer to lead Labs, the company's internal incubator for experimental products.
Google says think deeper, not longer, and cuts inference costs by half
Alibaba Cloud updates Coding Plan with new Qianwen and GLM models
Year of the Horse: Zhipu and MiniMax Up More Than 500% Since IPO
