Blog
Notes and analysis on AI development.
AI Agent Index exposes the transparency gap behind 25 agents
The 2025 AI Agent Index finds that 25 of 30 deployed agents do not disclose internal safety results as autonomy moves faster than public evidence.
Collibra AI Command Center moves agent audits into real time
Collibra AI Command Center turns agent sprawl into a real-time governance problem spanning registry, validation, traceability, and regulatory evidence.
Qoder 1.0 moves the AI IDE fight onto the developer desktop
Qoder 1.0 reframes AI coding from IDE assistance to a task runtime with Quest, team knowledge, reviewable artifacts, and parallel work.
Copilot App preview shows the coding agent bottleneck is the harness
GitHub Copilot App technical preview and the VS Code harness write-up show AI coding competition moving from model choice to execution loops and PR lifecycle control.
OpenAI’s $4B deployment company moves the model war into consulting
OpenAI Deployment Company shows frontier AI competition shifting toward enterprise deployment, FDEs, governance, and private equity distribution.
177,000 MCP Tools Show AI Agent Risk Has Moved to the Action Layer
AISI analyzed 177,436 MCP tools and found agent tooling shifting from reading and analysis toward file edits, browsers, payments, and other actions.
ChatGPT ads are coming to Korea, and answer trust is the test
OpenAI introduced ChatGPT Ads Manager, CPC bidding, conversion measurement, and a Korea pilot. The hard question is where AI answers end and ads begin.
AWS AI Security Framework sets a baseline for agent authority
The AWS AI Security Framework separates answering, connected, and acting AI, making agent identity, tool authorization, and observability the new security baseline.
GitHub accessibility agent learned its limits across 3,535 PRs
GitHub accessibility agent pilot shows what AI code review needs when quality assurance depends on data, escalation gates, and human judgment.
Codex comes to the phone, and the bottleneck is approval
OpenAI Codex mobile preview shows AI coding agent competition moving beyond model quality toward approvals, supervision, and remote execution.
Hermes Agent turns the local PC into a learning agent runtime
NVIDIA is positioning Nous Research Hermes Agent on RTX and DGX Spark as a local, always-on self-improving agent runtime.
Codex Moves to the Phone as Coding Agents Get a New Control Plane
OpenAI Codex mobile preview shows the coding-agent race moving from model capability toward approvals, remote execution, and cost boundaries.