Blog
Practical help for owners buried in follow-ups, quotes, customer questions, and office work.
Eval engineering: The missing piece of agentic AI governance
40% of production AI agent failures aren't hallucinations — they're the agent picking the right tool and passing it the wrong data. Eval engineering is the governance layer most teams are missing. Here's what it is and what to do about it.
Multiplayer AI startup Dust raises $40M to help enterprises move beyond
Dust raised $40M to replace isolated AI chatbots with multiplayer agents that share memory across organizations. Zero churn, 70% weekly usage, 300K agents deployed. Here's what this signals for the agent ecosystem.
A Meta employee created a dashboard so coworkers can compete to be the
A Meta employee built a token usage leaderboard for 85,000+ coworkers. It lasted two days after going public. The story it tells about AI agent productivity measurement is more interesting than the drama.
SaintQuant Launches Free AI Trading Bot Platform with Cash Bonus as AI Agents
SaintQuant launched a free AI trading bot with a $7 bonus. But eToro's CEO saying AI agents will eventually out-trade humans is the real signal worth understanding — and scrutinizing.
‘Polyanna policy’ – is NZ’s framework for AI use in government overly
New Zealand's AI framework for government is non-binding — good intentions, no enforcement. Researchers call it a 'Pollyanna policy.' Here's what that gap between aspiration and accountability means for anyone running AI agents.
Re-thinking human–machine interaction and the governance of AI in the military
Nature just published a peer-reviewed framework for human control of military AI. The three fixes researchers propose — contestation, continuous training, documentation — apply directly to anyone running a personal AI agent.
Cross-Agent Organizational Memory: How Knowledge Compounds
DORA 2024 found AI adoption improves individual productivity while hurting system-level delivery. The cause: stateless agent memory. Here's what organizational memory infrastructure actually requires — and why more agents without it makes things worse.
Meta Faces Internal Backlash Over AI Employee Monitoring Program
Meta's Model Capability Initiative logs every click and keystroke on employee computers to train AI agents. Over 100 angry emoji reactions. 8,000 layoffs announced the same week. Here's what the controversy actually reveals about agentic AI.
Governments Can’t Agree on What AI Actually Is
The world is trying to regulate AI without agreeing what AI actually is. A May 2026 Foreign Policy analysis reveals the definitional chaos — and what it means for anyone using personal AI agents today.
Red Hat targets enterprise deployment with new version of its AI platform
Red Hat unveiled AI 3.4 at its Summit conference — an enterprise platform built around running AI agents at scale. Here's what the announcement signals for anyone deploying or choosing an AI agent platform.
Meta’s Embrace of A.I. Is Making Its Employees Miserable
Meta installed keystroke-tracking software on employee laptops with no opt-out, tied AI usage to performance reviews, and announced 8,000 layoffs. The real story isn't corporate drama — it's a warning about who controls your AI agent.
Why International Business Machines (IBM) Is Leaning Into Sovereign AI and
IBM's Sovereign Core announcement at Think 2026 reframes AI governance from a policy document into a runtime requirement. Here's what it means for anyone building or deploying AI agents in regulated environments.