OrukLabs launches Resonance-2 with 31 emotion and speaking-style signals from audio
ChrisGPotts · x · 2026-09-19
OrukLabs has launched Resonance-2, claiming AI now outperforms humans at emotion recognition: it extracts 31 emotion and speaking-style signals directly from audio. Access opens first to a limited set of customers, with use-case review to prevent misuse. A linguist reshared it, referencing related prior work.
More from Models
- 3Blue1Brown Releases Full Video on the Last IMO Problem AI Couldn't Solve — 3blue1brown · 2026-09-19
- Jev seen as local-model stand-in for low-latency apps; open-weight RLCD models expected — HankYeomans · 2026-09-19
- Two years on, lateinteraction's RISC/CISC bifurcation prediction for foundation models is coming true — lateinteraction · 2026-09-19
- Google ships Gemini 3.8 Live audio models, family-agent CC, Google Pics and AlphaGenome Atlas — GoogleAI · 2026-09-19
- Speculation: New Model 'Jev' Built on Qwen MoE, RL-Calibrated for Decisions — ivan_bezdomny · 2026-09-19
- Meta Muse + Jev screens 10,000 candidates to surface top 100 unanswered immunology questions — DeryaTR_ · 2026-09-19