Technical analysis and architectural deep-dives from the leading edge of intelligence.
Microsoft is canceling most internal Claude Code licenses after token-based billing hit $500-$2,000 per engineer per month. Uber burned its $3.4 billion AI budget in four months with the same tool. The AI industry just entered its Discipline Phase.
Anthropic's Project Glasswing update reveals Claude Mythos found 10,000+ critical vulnerabilities in 30 days. The real story is that only 75 have been patched. AI solved bug discovery but created a patching crisis.
Liquid AI's LFM2.5-8B-A1B brings 128K context to consumer hardware — 1.5B active parameters, 253 tok/s on a laptop, and tool calling that finally feels interactive.
Liquid AI's LFM2.5-8B-A1B packs 128K context and real tool calling into 1.5B active parameters. It runs at 253 tok/s on a laptop. Local agents just became practical.
Tencent's Hy3 preview has quietly become the most-used model on OpenRouter. Here's the tech, the pricing, and what it says about AI in 2026.
Test body...
Anthropic raised $65 billion at a $965 billion valuation, surpassing OpenAI. Claude Opus 4.8 is 4x less likely to let flaws slip through.
A new paper proposes letting LLMs 'sleep' between processing chunks, consolidating context into fast weights and clearing the cache. The results on multi-hop reasoning are 3x better than baseline.
NVIDIA's LocateAnything uses Parallel Box Decoding to make vision-language models 10x faster at visual grounding while improving accuracy on LVIS and COCO. The 3B model hits 12.7 boxes per second and handles dense detection, GUI, and document grounding from a single checkpoint.
OpenRouter raised $113M at a $1.3B valuation as the multi-model AI shift accelerates. Behind the numbers: 25 trillion weekly tokens, 8 million users, and the team making it happen with only 50 people.