Daily AI news, sourcedUpdates / All
Updates 1959 items · newest issue first · page 41 of 98
AreaAllModelsAgents & codingApps & productsChips & infraResearchPricing & limitsBusiness & fundingPolicy & safetyOther
- Codex Day 6: OpenAI resets Codex usage limits globally by end of day
- LocalLLaMA post ranks models that fit 10–16GB VRAM for more spontaneous local chatUnconfirmedr/LocalLLaMAOct 12Models
- Runway CEO jokes that Chilean Spanish is the ultimate LLM benchmarkOfficialX @c_valenzuelabOct 12Models
- GitHub Security Lab open-sources an AI Fuzzing Taskflow agent for C/C++ reposOfficialGitHub BlogOct 12Agents & coding
- mlx-community publishes SenseNova-U1.5-8B-MoT 8-step V2 4-bit for Apple Silicon image generationUnconfirmedHugging Face @mlx-communityOct 12Models
- LocalLLaMA user ports Gemma 4 E4B EXL3 via exllamav3 as a Jetson Orin alternative to llama.cppUnconfirmedr/LocalLLaMAOct 12Chips & infra
- LocalLLaMA user runs Qwen3.8-Flash-Next at 500k context on an AMD Radeon RX 7900 XTXUnconfirmedr/LocalLLaMAOct 12Models
- vLLM Semantic Router releases Decision 3.0, five Apache-2.0 multimodal decision models from 0.8B to 27B
- Singularity user has Claude Opus 5.5 build a self-tooling SVG renderer for a PS5 controller cutoutUnconfirmedr/singularityOct 12Agents & coding
- Singularity user says Claude Opus 5.5 added DLSS frame gen to Transport Fever 3 in under two hoursUnconfirmedr/singularityOct 12Agents & coding
- LocalLLaMA user shares vllm-ascend progress running GLM-5.3-Flash on dual Ascend 310P cardsUnconfirmedr/LocalLLaMAOct 12Models
- LocalLLaMA user builds GamePause to free VRAM for gaming then restore local LLMsUnconfirmedr/LocalLLaMAOct 12Apps & products
- Project Maya runs GLM-5.3-Flash locally with VRAM, RAM, and NVMe expert tiers on one to 16 GPUsOfficialGitHub @mw00/project-mayaOct 12Apps & products
- EasyWhisperUI ships a redesigned local Whisper desktop app with Vulkan and Metal GPU acceleration
- ArtCraft open-sources early-alpha Rust Adobe Creative Suite clones built with Claude Opus 5.5
- 414k-parameter transformer learns boid flocking; probes recover rules the model may not useUnconfirmedr/MachineLearningOct 12Research
- NeurIPS 2026 Paris complimentary Top Reviewer registration already full, per a Reddit discussionUnconfirmedr/MachineLearningOct 12Research
- Palette Atlas searches 48k public-domain paintings by color scheme via sliced Wasserstein embeddings and HNSW
- OpenMed 3.0 ships as an Apache-2.0 local-first clinical NER and PII de-identification SDK
- Nvidia is in early talks to acquire or deepen its stake in Reflection AI, per Financial TimesReportedFinancial TimesOct 12Business & funding