Search

1. Anthropic Launches Claude Haiku 5.5: From $0.10, About 75% Cheaper Than Its Predecessor

   

   … it matters: It is the first Haiku with effort controls and the first with built-in safeguards for high-risk c…

   Stories·Oct 8

2. NYC Council Weighs 10 AI Bills: Mandatory Kill Switches, Subpoena for Musk's SpaceXAI

   

   … bills would fine $25,000 per AI system deployed without third-party validation and create America's first AI …

   Flash·Oct 7

3. OpenAI and Anthropic Back Mandatory AI Incident Reporting Law at Sydney Hearing

   

   Why it matters: OpenAI and Anthropic backed mandatory reporting legislation at the hearing; it follows an Open…

   Flash·Oct 7

4. Meta's Muse Instructed to Put the User's Home Authority Above Its Own Safety Training — Leaked Files Show

   

   …ditionally overrides safety training, and a researcher got them all by asking in the chat box.

   Stories·Oct 6

5. Anthropic Cuts Live Internet Access for All Internal Evals

   

   …ions and concedes alignment training is "not yet sufficient" for skills like search and computer use; it brief…

   Stories·Oct 10

6. context-mode Hits GitHub Trending: Tool Output, Sandboxed First

   

   context-mode is on GitHub's daily trending chart: an MCP server that sandboxes tool output before it reaches t…

   Flash·Oct 12

7. Reka Launches Rho-1, a 19B Omni-Model Handling Text, Images, Video and Robot Actions in One Network

   

   On October 5, Reka AI released a research preview of Rho-1: a 19B omni-reasoning model, trained from scratch, …

   Stories·Oct 6

8. HeyGen Open-Sources HyperFrames: Write HTML, Render Video

   

   …mations into MP4 and ships 21 agent skills, so a coding agent can go from writing the page to the finished vid…

   Flash·Oct 11

9. Microsoft Launches Decision-1, a Decision Model That Outputs Calibrated Probabilities, Not Text

   

   …ties for scoring decisions in agent workflows, reportedly 35x faster at P50 and 58% cheaper.

   Flash·Oct 11

10. Anthropic Open-Sources 11 Knowledge-Work Plugins Covering Sales, Legal and Finance

    

    …a and other roles; they are plain Markdown and JSON files with no code or build steps, connecting to Slack, Ji…

    Flash·Oct 10

11. Claude vs GPT for Coding: Same Mid-Tier Price, a 2.5× Flagship Gap, Benchmarks Measured Differently

    

    …na is ten times cheaper than Haiku 4.5.

    Comparisons·Oct 7

12. PPT Master Hits GitHub Trending: Documents Become Native PowerPoint

    

    PPT Master is on GitHub's daily trending chart: it turns documents into native, editable PowerPoint decks, wit…

    Flash·Oct 12

13. Nvidia-Backed Reflection Debuts Beam, Its First Open-Weight Frontier Model: 501B Params, 23B Active

    

    …ober 5, US startup Reflection AI launched Beam, its first frontier open-weight model: a text-only MoE with 501…

    Stories·Oct 6

14. Hands-On: Perplexity pplx-embed-v2-late — the 0.6B Model That Queries a 9B Index

    

    …akes its highest similarity against document tokens and the maxima are summed (MaxSim). The models are built o…

    Reviews·Oct 12

15. Google Launches EmbeddingGemma 2: 740M-Param Multimodal Embedding Model Goes Apache 2.0, Runs On-Device

    

    …with LiteRT, MediaPipe, LangChain and llama_index support. What Happened Google's official blog announced Embe…

    Stories·Oct 7

16. Open-Source Strata Runs a 125-Billion-Parameter Model on a 12GB Gaming PC

    

    …illion parameters must stay available, but only a small fraction compute at any moment. The conventional appro…

    Stories·Oct 5