https://heraldofai.com/models/kimi-k3-1-pro-a-chinese-ai-model-anyone-will-be-able-to-download-nearly-matchesHerald of AIen2026-10-23T20:54:05.134ZKimi K3.1 Pro, a Chinese AI model anyone will be able to download, nearly matches Claude Opus 5.5 https://heraldofai.com/culture/mit-celebrates-humanities-at-a-time-when-ai-is-changing-everythingHerald of AIen2026-10-23T19:47:59.351ZMIT celebrates humanities at a time when AI is changing everything https://heraldofai.com/tech/a-new-engine-makes-small-ai-models-respond-in-under-50-millisecondsHerald of AIen2026-10-23T16:25:41.366ZA new engine makes small AI models respond in under 50 milliseconds https://heraldofai.com/culture/why-early-ai-was-designed-to-be-random-and-why-we-stopped-liking-itHerald of AIen2026-10-23T13:58:17.673ZWhy early AI was designed to be random — and why we stopped liking it https://heraldofai.com/tech/cactus-whistle-vs-whisper-base-cpu-benchmark-shows-tiny-model-holds-its-ownHerald of AIen2026-10-23T11:31:38.088ZCactus Whistle vs Whisper Base CPU benchmark shows tiny model holds its own https://heraldofai.com/models/specializing-the-vocabulary-of-small-language-models-cuts-inference-cost-by-77Herald of AIen2026-10-23T11:28:52.915ZSpecializing the Vocabulary of Small Language Models Cuts Inference Cost by 77% https://heraldofai.com/models/recursive-releases-flint-a-40b-open-weight-local-coding-model-for-agentsHerald of AIen2026-10-23T11:24:26.444ZRecursive Releases Flint, a 40B Open-Weight Local Coding Model for Agents https://heraldofai.com/models/nvidia-kumo-relational-beats-supervised-ml-on-relational-data-without-trainingHerald of AIen2026-10-23T11:21:57.760ZNVIDIA Kumo Relational beats supervised ML on relational data without training https://heraldofai.com/tech/google-opens-torchtpu-a-native-pytorch-backend-for-cloud-tpusHerald of AIen2026-10-23T11:08:31.886ZGoogle opens TorchTPU, a native PyTorch backend for cloud TPUs https://heraldofai.com/tech/agentic-rl-has-a-scaling-law-and-it-changes-how-you-should-spend-your-budgetHerald of AIen2026-10-23T11:08:31.770ZAgentic RL Has a Scaling Law — And It Changes How You Should Spend Your Budget https://heraldofai.com/weekly/this-week-in-ai-agentic-tools-cheaper-models-and-growing-painsHerald of AIen2026-10-23T11:08:31.669ZThis Week in AI Agentic Tools, Cheaper Models, and Growing Pains https://heraldofai.com/tech/teaching-ai-agents-to-know-when-to-stopHerald of AIen2026-10-23T11:07:23.134ZTeaching AI Agents to Know When to Stop https://heraldofai.com/tech/researchers-develop-method-to-make-large-language-models-respond-fasterHerald of AIen2026-10-23T11:05:59.729ZResearchers develop method to make large language models respond faster https://heraldofai.com/tech/peekkv-a-smarter-way-to-compress-ai-memory-that-boosts-long-context-speedHerald of AIen2026-10-23T11:05:59.623ZPeekKV A smarter way to compress AI memory that boosts long-context speed https://heraldofai.com/tech/deepseek-open-sources-deepinfer-its-high-performance-inference-engineHerald of AIen2026-10-23T10:58:40.118ZDeepSeek Open-Sources DeepInfer, Its High-Performance Inference Engine https://heraldofai.com/models/us-bans-deepseek-from-federal-agency-systems-under-new-dhs-orderHerald of AIen2026-10-23T10:57:36.640ZUS bans DeepSeek from federal agency systems under new DHS order