Block's Dessa debuts new voice model that stunned listeners with its naturalness
DeryaTR_ · x · 2026-10-24
Block's Dessa has released a new voice model, and early reactions are effusive: listeners describe the output as incredibly natural and 'mind blowing.' No technical details or benchmarks have been shared yet, but the demo suggests another notable leap in speech synthesis realism.
Related event: Block Releases Dessa, a Speech-Only Trained Full-Duplex Voice Model(10 posts)→
More from Multimodal
- Star Vixens 2099: a sci-fi poster made in 2 weeks on an RTX 3060 with MiniMax and LTX 2.5 — Accurate_Village2024 · 2026-10-24
- LTX 2.5 distilled 22B Q4 runs locally on an 8GB M1 MacBook Air — 36s of video per ~32min render — iJohnnyRobert · 2026-10-24
- Open-source Claude Code skill turns any SVG into professional 60fps animation — rudrank · 2026-10-24
- Polymarket: SpaceXSI Begins Releasing Fully Grok-Generated Episodes of Homer's Odyssey — Polymarket · 2026-10-24
- Werewolf test on RTX 5090: Singularity V2 vs 10Eros local video generation compared — rileygstaliger · 2026-10-24
- Anil Seth blasts Tavos Griffin's video Turing test claim: creating fake people is a terrible idea — round · 2026-10-24