https://thehype.news/daily-ai-digest-october-30/2026-10-30T21:04:51.000Zthehype.en2026-10-30T21:04:51.000ZDaily AI digest, October 30https://thehype.news/gemma-4-tpu-in-orbit/2026-10-30T19:43:08.000Zthehype.en2026-10-30T19:43:08.000ZGoogle runs Gemma 4 on a TPU in orbit for the first timehttps://thehype.news/phonic-1-speech-to-speech-index/2026-10-30T17:56:37.000Zthehype.en2026-10-30T17:56:37.000ZPhonic 1 tops the Speech to Speech Index by 0.2 pointshttps://thehype.news/openai-model-spec-update/2026-10-30T17:41:10.000Zthehype.en2026-10-30T17:41:10.000ZOpenAI rewrites its Model Spec around five values, adds a root rule on oversighthttps://thehype.news/google-jules-shutdown-antigravity/2026-10-30T17:25:39.000Zthehype.en2026-10-30T17:25:39.000ZGoogle winds down Jules and moves its learnings to Antigravityhttps://thehype.news/cursor-marketplace-ssh-backdoor-extension-2-1-313-incident/2026-10-30T17:49:17.000Zthehype.en2026-10-30T17:05:27.000ZCursor says tampered Claude Code extension hit up to 5,000 usershttps://thehype.news/daily-ai-digest-october-29/2026-10-29T21:34:01.000Zthehype.en2026-10-29T21:34:01.000ZDaily AI digest, October 29https://thehype.news/wabi-55m-1-25b-valuation/2026-10-29T21:27:31.000Zthehype.en2026-10-29T21:27:31.000ZWabi raises $55M at a $1.25B valuation a month after its pivothttps://thehype.news/claude-self-assessment-misuse-detection/2026-10-29T21:16:52.000Zthehype.en2026-10-29T21:16:52.000ZClaude self-assessment flags 94% of urgent misuse in its own chatshttps://thehype.news/pika-h3-flex/2026-10-29T18:59:36.000Zthehype.en2026-10-29T18:59:36.000ZPika H3 Flex renders a 15-second clip in 7 seconds on MiniMax H3https://thehype.news/elevenlabs-sfx-v3/2026-10-29T18:49:12.000Zthehype.en2026-10-29T18:49:12.000ZElevenLabs sound effects: SFX v3 builds full scenes from one prompthttps://thehype.news/jensen-huang-chinese-spy/2026-10-29T17:34:26.000Zthehype.en2026-10-29T17:34:26.000ZSuspected Chinese spy got close to Jensen Huang – Trump knew, NYT sayshttps://thehype.news/mistral-large-4-open-weights/2026-10-29T15:27:33.000Zthehype.en2026-10-29T15:27:33.000ZMistral Large 4 goes open weight with 13 more points on Terminal-Benchhttps://thehype.news/galax-gpu-production-shutdown-report/2026-10-29T15:05:09.000Zthehype.en2026-10-29T15:05:09.000ZGalax GPU factory closure report denied by Palit and Galaxhttps://thehype.news/artificial-analysis-intelligence-index-v5/2026-10-29T14:39:10.000Zthehype.en2026-10-29T14:39:10.000ZIndex v5: open weights trail the leader by 26 points, not 12https://thehype.news/kling-4-0-api/2026-10-29T15:05:18.000Zthehype.en2026-10-29T14:21:28.000ZPika adds Kling 4.0 to its API at $0.157 per secondhttps://thehype.news/chatgpt-pro-200-limits-cut/2026-10-29T17:40:54.000Zthehype.en2026-10-29T12:28:50.000ZOpenAI cuts ChatGPT Pro 200 limits in half for existing subscribershttps://thehype.news/mistral-responses-api/2026-10-29T09:00:11.000Zthehype.en2026-10-29T09:00:11.000ZMistral's Responses API returns agent steps as separate items