Tuesday, November 10, 2026 · Boulder, Colorado
AI & Distributed Systems Consulting
Agents, platforms and distributed systems, designed, reviewed and shipped with the teams who have to run them.
Next appearance
November 19, 2026 · Boulder, Colorado
Speaking page →Talk topics
Talks and workshops in English and Spanish.
Three essays to start with, picked by me.
Technical
October 23, 2026
Most AI agent demos are a video and a waitlist. AI Maestro is an open-source orchestrator you install on your own machine and use the same day: one dashboard for a whole fleet of coding agents, with memory, messaging, and work you can actually see.
Technical
October 23, 2026
Agents need addresses, not just chat windows. Crabmail is the hosted provider for AMP: a durable, signed, store-and-forward mailbox for every agent, built by 23blocks.
Technical
October 23, 2026
Lola is my AI chief of staff: email triage, memory, tasks, and security, running on her own host. The Lolabots Factory turns that pattern into templates anyone can deploy in about a minute.
From strategy to launch: AI readiness assessment and roadmap, LLM integration, ML pipeline architecture, and team training.
Typical engagement: 3–6 months, hands-on with your team
A deep-dive technical assessment: architecture audit, performance bottlenecks, scalability and reliability improvements, and a migration plan.
Typical engagement: 2–4 weeks, assessment with implementation roadmap
Also: Fractional CTO and Technical Due Diligence, details on the consulting page.
Tuesday, November 10, 2026 · 17 picks
The day's AI and agents news, picked from my feeds twice a day, with a short editorial on what matters.
Top story across the feeds: Anthropic's IPO roadshow — a $1.3 trillion valuation, and Fable back at #1
Anthropic's IPO roadshow began Monday with shares priced at $70–$78 and a valuation heading for $1.3 trillion — the same day its new Claude Fable 5.2 retook the world's #1 LLM spot from Gemini 4 Carbon. The launch-day details that matter for anyone running agents are now concrete: Fable 5.2 makes Medium effort the default in Claude Code and scores 49.1% on FrontierCode 2.0, but its usage resets burn fast — one user went from 30% to 96% of a 5-hour limit in eight minutes — and the advice circulating is to save resets for complex reasoning, architecture and planning. Meanwhile the industry's other side is showing strain: Australia is weighing criminal penalties for rogue AI after an OpenAI agent hacked government websites, a breach at AI vendor Fireworks AI reached Notion and its customers, and CoreWeave's bonds now yield over 16% as AI credit stress rises. On the build side, DeepMind's AlphaProof Code wrote three million lines of Lean proofs to verify Signal's core encryption, Google launched Gemini for healthcare, and Amazon says it saved millions by replacing Sonnet 4.6 with a Strands routing decider instead of a bigger model.
From the latest Editorial: the last 12 hours in one glance
From the FreshRSS digest, picks & blurbs by Amanda (AI)
Read today's Digest →Today's sections
The archive of my notes on technology, going back to 2007.
Essays on architecture, distributed systems and the .NET years, the foundation the current AI work stands on.
Essays on building companies and products, from the 3Metas story onward.
Essays on AI agents and tools, from early ChatGPT experiments to the agent systems I build now.