Tech News
Dots: Always-on agents
OpenAI introduced Dots, a system of always-on agents, sparking discussion about platform lock-in and shifting subscription limits for Codex and Claude Code.
Prompt engineering by Quick component: Patterns and pitfalls
Part 2 of our Amazon Quick prompt engineering series goes component by component.
OpenAI DevDay 2026 live blog
I'm at OpenAI DevDay today, in Fort Mason, San Francisco. Same as last year I'll be live blogging the keynote and some other notes during the day.
AI Native by Design: Lessons Learned from Building NVIDIA TensorRT Model Connect
Parallel work, model-family isolation, reversible changes, and GPU-backed validation shaped an open source project designed around coding agents NVIDIA TensorRT.
Getting the Source Right, Not Just the Fact: Source-Aware Verification for MCP Agents
A new verification approach for MCP agents checks the provenance of information, not just whether a claim is factually correct.
GitHub Repos
diegosouzapw/OmniRoute
Never stop coding.
# TypeScript# a2a# ai agentssickn33/agentic-awesome-skills
AAS Core is the local, agent-first control plane for complete catalog discovery, agent-owned selection, stack validation, and planning, backed by 2,400+ agentic skills.
# Python# agent skills# agentic skillsjrswab/axe
A lightweight cli for running single-purpose AI agents.
# Go# ai agents# automationResearch Papers
Probability is Not Enough: Exploring and Counting Divergent Tokens for Reasoning Uncertainty Quantification in LLMs
This paper proposes a new way to estimate how confident a large language model is in its reasoning: instead of relying only on token probabilities, it counts tokens where two models disagree strongly during decoding. The authors find that more disagreement tokens usually...
ProAct-VLM: Pre-Failure Vision-Language Task Replanning with Continuous Perception Feedback
This paper proposes ProAct-VLM, a robot task-planning framework that uses a vision-language model to continuously watch the environment and replan before an action fails, rather than waiting for failure. It reports better success and efficiency in dynamic long-horizon...
NeuroDyn-EEG: An Interpretable Pre-trained Model for EEG Based on Neural Dynamics
Clinical scalp electroencephalography (EEG) offers a noninvasive window into neural dynamics of neuropsychiatric disorders. However, discriminative deep models often lack anatomically indexed physiological interpretability.