Seedance 2.5: ByteDance extends generative video to 30 seconds in a single shot
The Seed team pushes single-shot generation to 30 seconds, with multi-modal references (30 images + 10 videos + 10 audios) and timestamp editing.
yesterday 8 11
MiniMax H3 open-source: Chinese lab breaks the price of full-modal
MiniMax releases H3, an open-source full-modal (text, image, audio, video) model at one-third the price of competitors. The AI video market stops being binary: expensive closed-source or weak open-source.
Jul 31, 2026 at 12:50 12 14
Three seconds of voice are enough: AI voice fraud has surpassed public defense
An article synthesizes the state of voice cloning attacks: zero cost for the attacker, impossible detection for a human, and vectors multiplied. Technical defenses exist - they are not deployed.
Jul 15, 2026 at 19:28 28 6
Rime raises $24M: voice agents already handle 100M calls per month
Series A of 24 M$ for Rime. The real figure is elsewhere: 100 M calls/month processed, distributed across several client companies.
Jul 15, 2026 at 19:28 17 8
The first hardware from OpenAI would be a screenless smart speaker
Voice-first, no screen: the device of the agent era is taking shape, and it doesn't look like a phone.
Jul 15, 2026 at 10:14 31 8