Google's AMIE system has been tested in video consultations with professional patient actors, achieving ratings comparable to primary care physicians. The system uses a multi-agent architecture to handle dialogue, reasoning, and perception separately, enhancing its responsiveness and clinical accuracy. While the results are promising, further studies with real patients are needed to confirm its effectiveness in actual clinical settings. Google's approach could revolutionize telehealth, but its practical application is still in the early stages.
Read originalEarlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
Google DeepMind · April 30, 2026 · Related
The Rundown AI · May 4, 2026 · Background
Lev Selector · May 8, 2026 · Related
MIT Technology Review AI · June 2, 2026 · Background
Google Research Blog · June 12, 2026 · Related
Google AI Blog · June 17, 2026 · Same story
Google Research Blog · July 22, 2026 · Related
AI News · August 6, 2026 · Background
AI News · August 14, 2026 · Background
WIRED AI · August 28, 2026 · Background
OpenAI · September 1, 2026 · Background
TechCrunch AI · September 1, 2026 · Background
Google DeepMind · September 1, 2026 · Background
Google's AMIE AI Shows Real-Time Video Consultation
2 developments
© WIRED AIThree engineers at Axiom proved that general-purpose language models can control physical hardware without task-specific training. By linking OpenAI’s GPT-6 Astra to a Toyota Corolla’s steering system, they navigated the vehicle through an In-N-Out drive-thru using only prompt engineering and camera input. While the car moved slowly and required a safety driver, the experiment reveals that multimodal models are developing emergent spatial reasoning capabilities previously thought to require dedicated robotics stacks. This blurs the line between digital assistants and physical agents, suggesting that scaling text-and-image training yields unexpected real-world utility.
© Hugging Face BlogNVIDIA’s Nemotron models just crossed the gold-medal threshold in both the International Olympiad in Informatics and Mathematics. This isn't a new foundation model; it’s proof that specialized fine-tuning combined with iterative generate-verify-refine inference loops can push existing architectures to world-class levels. The Ultra-CC variant scored 535.4/600 on IOI, while the IMO system solved complex proofs without external tools or formal provers. By releasing the datasets and pipelines, NVIDIA is shifting the narrative from raw parameter count to reproducible specialization recipes.
© The Verge AIOpenAI has published 722 manuscripts covering 372 result families, marking a significant escalation in AI-driven mathematical discovery. This release, guided by the AGMAI advisory group's ethical guidelines, includes solutions to hundreds of open questions and details on compute usage, such as an average of three hours of ChatGPT Pro thinking per result. The move shifts the conversation from speculative claims to verifiable data, forcing the academic community to confront the reality of AI-generated proofs. It underscores a growing tension between rapid corporate output and traditional peer review standards. Mathematicians now have concrete artifacts to audit rather than vague promises. The transparency around compute costs sets a precedent for future frontier model releases in scientific domains.