11-13 Daily Briefing

AI News Daily · 13 Nov 2025

AI News | Daily Briefing | Tools | Research | Industry | Open Source | Human Impact

Highlights

• Kuaishou Kling 2.5 Turbo adds start/end-frame control to keep AI videos narratively coherent.
• ElevenLabs Scribe v2 Realtime delivers 150 ms latency STT across 90+ languages, dominating noisy-domain benchmarks.
• Google Photos integrates Gemini Nano Banana for voice-controlled edits; Alibaba unveils the 0.6B-parameter SmartResume parser.
• Industry watchers forecast 2026 as the point when AI replaces certain customer-facing roles; Xiaomi escalates talent hiring for foundation models while China upgrades brain–computer interfaces to national strategy.

Product & Platform

1. Kling start & end frames — Users can now lock the first and last frame for videos generated by Kling 2.5 Turbo, eliminating awkward cutoffs. [Demo]
   
2. Scribe v2 Realtime (ElevenLabs) — 150 ms latency, 90+ languages, and top accuracy in “hell mode” (noisy audio + jargon) according to the official release. [Report]
   
   
3. Google Photos × Gemini Nano Banana — Give natural-language instructions (“turn this into a Renaissance portrait,” “fix the closed eyes”) and Gemini handles the edit, turning photo retouching into a chat session. [Overview]
4. SmartResume (Alibaba) — 0.6B-parameter, layout-aware résumé parser with “layout perception + parallel task decomposition,” extracting messy résumés in 1–2 seconds while rivaling Claude-4. [Deep dive]
   
   

Research Notes

1. LLM/VLM robotics survey — Summarises how language and vision models empower robots to plan and interact autonomously. [Paper]
2. SpeechJudge — A large human-preference dataset + evaluator that teaches models to rate speech naturalness like humans. [Paper]
3. X-Scape & video-efficiency work — Researchers explore richer simulation environments and faster inference for video/autonomy models (see linked arXiv series in the Chinese edition).

Industry / Community

• Reports predict 2026 will be the turning point when AI begins replacing front-line roles (support/BPO first) unless firms reskill staff.
• Xiaomi’s aggressive hiring around large models underscores China’s ambition, while brain–computer interface programs are elevated to national priority status.