Wednesday, June 3, 2026
10 stories covered by gentic.news intelligence
Nvidia launched physical AI agent skills and a 32B VLA model at CVPR to automate AV and robotics workflows, addressing the fragmented tooling bottleneck.
Google launched Gemma 4 12B, an encoder-free multimodal model for on-device AI, reducing latency by eliminating the vision encoder.
MiniMax M3: Sparse Attention, 1M Context, Multimodal via …
MiniMax M3 uses sparse attention for 1M context and multimodality, with Together AI serving fast inference.
ChatHealthAI aligns CLMBR-T-Base with a frozen LLM via a task-aware resampler, achieving 79.8% F1 on EHRSHOT length-of-stay prediction while enabling interpretable reasoning.
Law Profs Prefer AI Answers 75% of Time in Stanford Study
Stanford researchers found law professors preferred AI answers 75% of time in blind legal analysis test, per @rohanpaul_ai.
Miso One, an 8B open-source TTS model, achieves 110ms latency with emotional range. Weights are fully open-source for self-hosting, but no benchmark data is provided.
Google's LEAP scaffold lifts Lean-IMO-Bench one-shot solve rate from <10% to 70%, solving all 12 Putnam 2025 problems.
Google released Magenta RealTime 2 on Hugging Face, the only open-weights model for real-time continuous music generation on device with ~200ms latency.
Superforecasters predicted 3-4h METR 80% task horizons by year-end 2026. Claude Mythos hit that in late May, compressing the timeline by seven months.
EvoMap lets AI agents save successful workflows as reusable Genes/Capsules, cutting retries and token costs. The network turns one-off runs into shared infrastructure for coding and security teams.