wpnews.pro

Yesterday · Thu 19 Nov

Wednesday 18 November

Introducing GPT-6.2 Astra and Sol

AI-generated summary OpenAI says it has introduced GPT-6.2 Astra and GPT-6.2 Sol as new models designed for professional work and everyday tasks. OpenAI says GPT-6.2 Astra delivers improved performance on complex professional assignments while GPT-6.2 Sol offers efficient capabilities for everyday use.

Also covered byRSS.bz9to5Macwpnews
Earlier:

Monday 16 November

TPU Model Design Considerations and Optimization Guide- Google Developers Blog

AI-generated summary Google announced the open-source AI-Hypercomputer/tpu-codesign guide to help developers accelerate workloads on tensor processing units. Google states that eighth-generation tensor processing units add native floating-point four support and other hardware upgrades.

Also covered byWeb Pulse
Earlier:

Meta’s Muse Ember Bets the Future of AI Agents on Cheap, Reliable Work

AI-generated summary TechCrunch reported on Meta releasing Muse Ember, a lightweight multimodal model built for repetitive agent work including tool calls, coding subtasks, routing, extraction, and long-document parsing. Meta states that Muse Ember is priced at $0.10 per million input tokens and $0.50 per million output tokens for workloads dominated by repeated and relatively narrow tasks.

Also covered byTool Minds AI
Earlier:

Google’s Robot Ambitions Now Depend on Closing the Final 4 Percent

AI-generated summary Google DeepMind announced that Gemini Robotics 2.5 was evaluated on more than 200 dexterity tasks. Google DeepMind reported a 96 percent success rate for automotive kitting involving Apptronik’s Apollo 2 humanoid robots on Mercedes-Benz production lines.

Earlier:

Sunday 15 November

Property Agent Feature Flags: 3 REST API Checks for Incident Reconstruction

AI-generated summary Web Pulse outlines a proposed production rehearsal for property management AI agents using Node.js feature flags and a REST API control boundary to reconstruct maintenance triage runs after flag changes. The text suggests capturing evaluated flag values once per loop alongside latency and cost evidence while using synthetic maintenance requests and an operator ledger.