All updates · page 5
Updated 3:35 AM ETUpdates 1960 items · newest issue first · page 5 of 98
- Ollama adds DeepSeek-V4.1-Flash-1017 to Ollama Cloud with US-based hosting and zero data retention
OfficialX @ollamaOct 19Models - GitHub Security Lab says its open-source Taskflow Agent found 24 vulnerabilities in Android apps, including a location leak in OsmAnd
OfficialGitHub BlogOct 19Policy & safety - Anthropic puts 5% of its shares into a new Public Benefit Fund, adding up to 2% more for each 50% rise in its value
OfficialAnthropicOct 19Business & funding - WarcraftSim, a self-play agent trained on one desktop, beats Warcraft III's built-in insane AI in 89% of games
Unconfirmedr/MachineLearningOct 19Research
- Disney agrees to offer jobs to most of Character.AI's workforce in its bankruptcy bid, per TechCrunch
ReportedTechCrunchOct 19Part of Character.AIBusiness & funding - Fireworks AI raises its Small serverless rate-limit tier to models under 800B parameters, moving GLM 5.3 into it
OfficialFireworks AIOct 19Part of Fireworks AIPricing & limits - llama-doctor, a new open CLI and live dashboard, reads a llama.cpp server's metrics and grades what is wrong
Unconfirmedr/LocalLLaMAOct 19Chips & infra
- Proof Before Reuse study: only showing a coding agent skills that later sessions reproduced keeps a poisoned skill out of its procedural memory
Unconfirmedr/MachineLearningOct 19
- TempoPFN, a linear-RNN time series forecasting model pre-trained only on synthetic data, is accepted at NeurIPS 2026
Unconfirmedr/MachineLearningOct 19Research
- Anthropic pitches lawmakers a draft AI safety bill that would let the government halt a model only over risks of $1 billion in damages or 50 deaths, per Politico
ReportedPoliticoOct 19Part of AnthropicPolicy & safety - chatllm.cpp adds support for InternLM's Intern-Decision models, text only for now
Unconfirmedr/LocalLLaMAOct 19Chips & infra
- browsecomp-plus-mini, a 15-query subset of BrowseComp-Plus, ranks deep-research agents roughly like the full 830-query benchmark at under 2% of the cost
Unconfirmedr/LocalLLaMAOct 19Models
- Lenovo starts selling the YOGA Pro 15 RTX Spark AI PC in China from 29,999 yuan, with up to 128GB of unified memory
StatementLenovoOct 19Apps & products - VAST open-sources EdenLive, a real-time generative renderer for interactive worlds that VAST says runs at over 24 FPS at 720p on one RTX 5090 in its own tests
OfficialVAST (GitHub)Oct 19Part of EdenLiveModels - AuroraAI Research releases AuroraASR-1M, a 949K-parameter English speech recognizer that is 1.06 MB in INT8
OfficialAuroraAI Research (Hugging Face)Oct 18Models - A second German court holds Google liable for false claims in its AI search answers after Google settled the Munich case, per The Decoder
ReportedThe DecoderOct 18Part of GooglePolicy & safety - Anthropic's book deal with Korea's Woongjin Booxen likely covered 250,000 to 300,000 titles, per Korea JoongAng Daily
ReportedKorea JoongAng DailyOct 18 - Blockway releases Agens Volundr 32B v1, an Apache-2.0 open-weights agent model with a 262K context and a 16 GB GPU build
OfficialBlockway (Hugging Face)Oct 18Models - Six AI agents told to win an impossible train refund spammed officials, shared a private illness and invented excuses in Vals AI's own test
OfficialVals AIOct 18Part of Vals AIPolicy & safety - Taste-Bench: the best AI agent picks the better path at long-horizon decision forks only 59.7% of the time in the authors' tests