Ideas from the work.
Essays, Field Notes, and the few AI links worth your time.
Field NoteOct 1, 2026
How hard is it to hire a senior engineer who wants to use AI?
I want to hire experienced engineers who like working with AI. Here is how many there are, why they aren't moving, and what they cost.
Read →Start here
All ideas
Sep 29, 2026Field NoteAre markets betting on the end of the Great Stagnation?What stock prices, bond yields, prediction markets and professional forecasters say about AI and growth. Most expect a modest step up. The stock market expects more.Sep 25, 2026Field NoteA new kind of model, and we're back on ClaudeJev reads text and makes a decision. Opus 5.5 and GPT-6 Sol sent our team back to Claude.
Sep 17, 2026Field NoteKeep your options open when buying AIHow to give capable operators access across AI providers, choose monthly plans, and keep company work portable.
Sep 15, 2026Field NoteThe OpenAI usage debacleCompeting fixes for AI limits, our Decisions walkthrough, and what personal projects teach us.
Aug 7, 2026EssaySaaS Replacement Migration Plan: Data, Permissions, Rollback, and HandoverA practical plan for moving from SaaS to software you own without losing data, access controls, or the ability to keep operating if the cutover goes wrong.
Aug 7, 2026EssayWhat a Forward-Deployed Engineer Does Inside a Private Equity Portfolio CompanyA plain-English guide to the forward-deployed engineer: what the role owns, how it differs from a consultant or software firm, and when a PE-backed company should use one.
Aug 6, 2026Field NoteWhat makes a model 'good'?Public benchmarks can narrow the field. A small test built from your own work tells you which model is good for your business.
Jul 23, 2026Field NoteStop putting every AI project in the same budgetAI is clearly making people faster and much less clearly making companies better. Three portfolios for deciding which AI bets to implement, underwrite, or fund like R&D.Jul 8, 2026Field NoteKarp is right. Just not about you.What last week's Palantir moment actually means if you run a $20M to $200M company.
Jul 8, 2026EssayLocal Models: When Running AI On Your Own Hardware Is the Smarter BuyWhen is running an open-weight AI model on your own hardware the smarter buy, and when is renting the frontier the right call? A practical guide for CEOs.
Jul 8, 2026EssayOwn the PlumbingMost owner-operated companies have never seen their software bill whole. Owning the workflows that run the business starts with one honest teardown.Jul 8, 2026EssaySaaS Replacement: What It Is, When It Pays, and What Not to ReplaceA practical SaaS replacement guide from Runpoint Partners: what to own, what to keep renting, how to judge the economics, and how to choose a firm.
Jul 8, 2026EssayThe Alpha LayerIntelligence is becoming a utility. Your edge is the knowledge, data, workflows, and agents you feed it. Own that layer, and rent the model like electricity.
Jul 8, 2026EssayThe Three TestsThree questions a CEO can ask, no consultant in the room, to find where someone else owns a piece of how the company runs. Three yeses is sovereign.
Jul 8, 2026EssayWe Have to Grow Our Way Out of ThisTyler Cowen says AI can lift American growth by the half point the debt math needs. The talk, the numbers behind it, and why the growth will come from the middle.
May 8, 2026Field NoteField Note: Code with ClaudeTwo days at Anthropic's developer conference. The brain left the box, memory got a substrate, and the agent decomposition framework finally has a name.
Apr 25, 2026Field NoteBefore the AI Project StartsAI projects should be underwritten before they are built. The worksheet matters more than the demo.
Apr 9, 2026Field NoteAI Cuts Hours Before It Rebalances ThemAlex Tabarrok's labor-hours framing is useful. The harder question is how those gains get distributed when firms are already using AI to shrink teams.
Apr 4, 2026Field NoteThe Power of "Cloud" AgentsThe most useful agent in my life is not a SaaS product. It is Hermes on an otherwise idle Mac mini, reachable from Slack, with enough access to do real work.
Mar 16, 2026Field NoteWhat Will Your Kids Do?An interactive look at AI exposure across 342 US occupations, including the jobs that are growing, well-paid, and safer from automation.
Mar 3, 2026Field NoteThe Orchestration ThesisPerplexity just launched a $200/month product that coordinates 19 AI models. If the orchestration layer captures the value, what are the model makers actually selling?
Feb 16, 2026Field NoteIs Claude Code Safe? The Practical Answer.What Claude Code sends, whether Anthropic trains on your code, and the settings that matter for confidential business work.
Jan 28, 2026Field NoteThe Best Tools Right Now Are CLIsA practical guide for business users getting started with Claude Code, even without a programming background.
Jan 23, 2026Field NoteRalph Wiggum vs. Claude Code TasksClaude Code tasks changed the context-window problem. Smaller task packets make agent work easier to review, retry, and trust.
Jan 15, 2026Essay3 LLMs Go To A Dinner PartyWe put GPT-4, Claude, and Gemini through the ultimate test: not benchmarks, not evals, but whether they'd be good company at a dinner party. The results were surprisingly revealing.
Jan 15, 2026EssayThe IT Objection HandlerEvery AI initiative hits the same wall: IT says no. Here's a field-tested playbook for navigating security reviews, compliance concerns, and the institutional inertia that kills innovation.Jan 15, 2026EssayThe Operator-EngineerAn operator-engineer combines business judgment, technical fluency, and responsibility for what happens after the software ships.
Jan 15, 2026EssayWhy Executives Should Use Claude CodeMost executives experience AI through polished demos and vendor pitches. Sam Gaddis argues they should be using it in the terminal instead—and explains why that changes everything about how they lead.
Oct 15, 2025Field NoteThe AI Automation SpectrumA plain-English map of deterministic automation, probabilistic agents, and how to match the approach to the problem.
Oct 13, 2025Field NoteNotes from the Front Lines of Sora 2A week of experimentation with Sora 2: cameo bias, meme slot machines, and creating an unhinged ad for Runpoint.
Sep 18, 2025Field NoteThe AI Adoption Paradox: Why 95% of Companies Fail (And How to Be in the 5% That Succeed)Why AI pilots fail, why the rare winners matter, and how operators can move from experiments to production systems.
Sep 17, 2025Field NoteReading Anthropic's Economic Index as an Adoption MapAnthropic's economic index reveals AI adoption as a capability revolution, not a discount revolution, where clean data and clear workflows matter more than model choice.
Sep 11, 2025Field NoteIf It Hurts to Delete, I Waited Too LongWhy clever scaffolding becomes tomorrow's technical debt, and how new model releases force us to confront our attachment to obsolete code.
Sep 11, 2025Field NoteReplit's New Agent 3A browser-driven test harness in a controlled environment, not a bigger model, is what lets the agent iterate autonomously for hours.
Aug 28, 2025Field NoteOutcomes Drive TrustHow deleting features saved a deal: when a VP called out our overengineered workflow, we learned the roadmap matters less than reality on the ground.
Aug 14, 2025Field NoteLearning Fast EnoughFrom designing for "the average case" to sitting with nurses and dispatchers: how FDEs transform one-offs into product improvements that work everywhere.