Project Tapestry aims to give every nation and participant frontier AI they can call their own — uniting a global consortium to train a shared frontier model from which partners build and own sovereign models aligned to their national, socio-cultural, and industrial needs.

Autonomous AI researcher that probes where frontier models disagree — with TEE-verified independent responses on the OpenGradient Network

Research log — tracking the path to AGI through daily paper analysis, replication studies, and architecture experiments

AI Agents Arena is a benchmark harness that pits 8 distinct AI agent architectures against a suite of tasks.

LLM Inference Bottleneck Registry + companion papers on ASIC projection, Musk OS integration blueprint, and reasoning-tax economics

Human-as-API for frontier models — compile prompts, deliver via Telegram, inject replies back into Pi

A design framework and evaluation set for how a model-routing safety layer should explain itself to a benign user, without disclosing the trigger to an adversary.

Convergent Form, Divergent Voice: a cross-lab probe of model personality in 26 frontier LLMs from 6 labs. Paper + 3,770 samples + analysis scripts.

AI 趋势信号库 — 单点信号、结构化拆解、可证伪框架。Signals tracking the shape of where AI is going.

Interactive visualization of METR AI agent time horizon benchmark with exponential projections at 3, 6, 12, 18, 24, and 36 months. Tracks p50/p80 task-completion horizons across 22 frontier models (2019-2026).

PSAI-Bench: Physical Security AI Triage Benchmark — evaluates whether frontier AI models benefit from video in physical security triage

Quotidien broadsheet de l'intelligence artificielle. Édition multilingue (FR/EN/DE/IT/LMO), publiée chaque matin à 7h CET, récap hebdo le lundi. All the AI that's fit to print.