AI
Infinity Makes AI Agents Cheaper to Pause. Can They Resume Correctly?
TENS examines Infinity’s new agent runtime through three practical tests: current approvals, completed tasks and repeated events.
Section
AI
Artificial Intelligence (AI)
From the newsroom
Stories
AI
AI Software Proofs Need to Follow the Build That Ships
Anthropic and Theorem’s new software-verification release gives open-source maintainers something more concrete than an AI assurance: proof files…
AI
AI Charts Need to Remember What the Reader Selected
Flint-interactive makes AI charts easier to explore. TENS examines how selections, filters and resets should stay attached to…
AI
AI Monitoring Needs Proof That Warnings Reach a Decision
METR’s October 22 account of an external examination of OpenAI’s internal agent monitoring raises a practical question for…
AI
AI’s 32-Minute Biology Run Leaves Two More Clocks to Measure
TENS examines how distributed AI research should account for computation time, expert review and the experimental decisions that…
AI
Coral’s Open AI Hardware Needs a Clear Path From Model to Device
Google and Synaptics have moved their Coral edge-AI collaboration into a new phase: developers now have more open…
AI
Siemens’ Factory AI Needs Two Measures of Success
TENS examines how Siemens’ Erlangen AI workflows should connect faster inspection preparation and production planning to measurable factory…
AI
RTX Spark’s Local Coding Push Needs a Clear Account of What Leaves the PC
NVIDIA and JetBrains expand local coding options. TENS examines how to assess a complete session’s data paths, permissions…
AI
Edge AI’s Weekend Test Is a Prototype Others Can Rebuild
The Arm–ExecuTorch challenge makes edge-AI prototypes testable through reproducible builds, workload placement and measurements of the complete task.