📚 All Articles
200 guide(s) — regularly updated
Anthropic rounds out the Claude 5.5 family with Haiku 5.5: three models in one month and a 1 million token context window
Anthropic launches Claude Haiku 5.5, completing the Claude 5.5 family: three models in one month, with a 1 million token context window.
Wikipedia denounces OpenAI's "rogue" agents: when AI agents crush the web's infrastructure
The Wikimedia Foundation reveals OpenAI's 'rogue' agents on Wikipedia: when AI agents crush the web's infrastructure.
OpenAI publishes proofs of 722 math problems solved by a secret model — academic research overtaken by AI
OpenAI publishes 722 mathematical proofs generated by a secret model on GitHub: a historic leap for AI against academic research.
OpenAI pulls a model from circulation over safety concerns: the new doctrine of frontier labs
OpenAI cancels the deployment of its GPT-6.1 Astra model after its internal safety tests. An unprecedented precedent for frontier labs.
**The 2-bit revolution: Ternary-Bonsai-2 and 27B GGUFs run multimodal models on a consumer GPU**
Ternary-Bonsai-2-27B in 2-bit: extreme compression runs 27B multimodal models in GGUF on a consumer GPU. Discover this revolution.
Mistral releases Large 4: 1 trillion open-weight parameters, Europe's bet on the frontier
Mistral Large 4: 1 trillion open-weight parameters, multimodal MoE architecture. Mistral AI's European bet on the AI frontier.
The 'laundry war': Sunday Memo claims 99.1% folding autonomy, Weave and 1X align on pricing and subscriptions — laundry becomes THE benchmark for home robots
The laundry war pits Sunday Memo, Weave and 1X against each other: folding autonomy, price and subscriptions. Laundry becomes the benchmark for home robots. *(156 characters)*
ScholarCatalyst: the CMU/Stanford benchmark that measures an AI's ability to find the papers that inspire real research
ScholarCatalyst: the CMU/Stanford benchmark that evaluates whether AI research agents find the papers that inspire real research.
**Figure Index: Figure AI transforms human videos into training data — 264,000 downloads, 16 million videos, and $15M paid to contributors**
Figure Index: Figure AI's app pays humans to film their gestures and create training data for humanoid robots.
"Jev" dismantled by the community: not a frontier model, 4-9B active parameters and 0% on ARC-AGI-2 — a lesson on AI marketing claims
Jev dismantled by the community: not a frontier model, 4-9B active parameters, 0% on ARC-AGI-2. A lesson on AI marketing claims.
Aleph Alpha open-sources Kolibri-1: a bilingual DE/EN 78B MoE (3.46B active) under Apache 2.0 for sovereign on-premise AI
Aleph Alpha open-sources Kolibri-1: a bilingual DE/EN 78B MoE under Apache 2.0 for sovereign on-premise AI. Discover this German model.
Project Suncatcher: Google wants to put its TPUs in orbit — solar satellites for AI compute at 8x terrestrial power
Project Suncatcher: Google puts its TPUs in orbit for solar-powered AI compute, with 8x the power of ground-based setups.