#Research
Curated coverage on «Research»: releases, research, benchmarks, open-source.
Apple Published Fortress — A Framework for Stabilizing Recommendation Systems Through Feature Pruning
Apple ML Research described Fortress — a system that eliminates temporal instability in recommendation models by identifying and removing features that create…
Harvard: AI Startups Hire Fewer Juniors and Bet on Experts
A working paper from Harvard Business School and INSEAD showed: AI-native startups build compact, flat teams and hire almost exclusively senior technical staff.
Synthetic Sciences Releases OpenScience — Open AI Workbench for ML, Biology, Physics and Chemistry
OpenScience by Synthetic Sciences — open-source workbench under Apache 2.0 with 250+ skills, compatible with any AI model for ML, biology, physics and chemistry…
Bernard Arnault Donates €50M to Mathematics Institute, Calls It the Key to the AI Era
Fable 5 by Anthropic Sets Record on Freelance Tasks, but Cannot Replace Humans
Google DeepMind and A24 announce first research partnership in cinema
How two students from St. Petersburg made it to AAAI — the largest AI conference in Singapore
70% of developers know AI writes buggy code — and deploy it to production anyway
Research shows most developers understand AI generates insecure code, but one in three still deploys it to production without additional verification.
MiniMax released MSA: sparse attention speeds up 1M-token context processing by 28x
MiniMax unveiled MSA, a two-branch sparse attention mechanism trained on a 109B-parameter MoE model that is 28x faster than standard GQA at a context length of 1M…
Iason Gabriel: how a philosopher at Google DeepMind is trying to make AI safe
Hybrid models predict content words better than transformers — Allen AI study
Critic analyzed Sber neuroscience lab study on AI and found methodological errors
Cursor Finds: 63% of Opus 4.8 Max's Benchmark Wins Are Search, Not Code
Anthropic's Fable Too Restrictive for Cybersecurity
Cybersecurity researchers complain that the guardrails of Anthropic's new Fable model are so strict that even legitimate vulnerability research and defensive system…
2026 TTS Models Comparison: From Commercial to Open Source
In 2026, choosing a TTS model depends on three factors: sound quality, processing latency, and cost. Commercial solutions win on naturalness, open models on control…