My notes on Jack Dorsey's HBR interview about replacing Block's management hierarchy with a shared record anyone can query, and the judgment that stays human.
Badges
Badges come from GitHub once the author claims this feed.
Recent posts
every link opens the author’s post-
Jack Dorsey on running a company as an intelligence
-
Nick Saraev's method for selling AI services on your own
My notes on Nick Saraev's guide to a one-person AI service: build for growth, write the offer as return times belief over friction, and keep the volume steady.
-
AI consulting start to finish: Nick Saraev's six steps
My notes on Nick Saraev's six-step method for AI consulting: find the bottleneck, break the goal into actions, build, deploy, test with evals and sync the team.
-
Russell Raath on sizing a consulting practice
My notes on Russell Raath's talk on building a solo consulting practice: the buying-decision math, a first engagement to sell, and pricing your contribution.
-
Clear thinking in the age of AI: Nick Saraev's course
Nick Saraev's five-hour course on thinking clearly: fix your body and workspace first, then learn eighteen mental models and five techniques for acting on them.
-
Highest paying careers in two years: Nick Saraev's test
My notes on Nick Saraev's talk on which careers still pay as AI agents work longer alone: pay moves to the bottleneck, and the six roles he thinks will hold it.
-
Wayfinder: Matt Pocock on planning across agent sessions
My notes on Matt Pocock's talk on Wayfinder, a skill for planning work that won't fit in one agent session: a map of decision tickets, fog and a frontier.
-
Eero Alvar's AI learning harness: probe, teach, practice
My notes on Eero Alvar's video on an open-source app for learning with AI: it maps what you know, teaches one step at a time, then drills recall.
-
Laurie Voss on evals for agents: read the traces first
My notes on Laurie Voss's workshop on testing AI agents: read the traces, write code checks and model judges, then check the judge against your own labels.
-
Claude Code best practices, now that you're out of the loop
My notes on Lydia Hallie's talk on Claude Code best practices for the 5.5 models: what the model, the harness and the environment take over, and what to keep.
-
Codie Sanchez tests five AI bots on turning $100 into $1,000
My notes on Codie Sanchez's ten-day test of five AI agents: the boring jobs, lead outreach and resale, paid; the stock-trading bot lost money.
-
JA Westenberg on agency: direction is the new bottleneck
My notes on JA Westenberg's video on personal agency: when AI agents do the work, the jobs left are wanting, choosing, judging and owning, plus six habits.
-
Is Software Over? AI that rebuilds programs from behavior
My notes on Two Minute Papers' video about AI rebuilding software without its source code: the three papers behind it, their real numbers, and why it works.
-
Alex Hormozi's idiot index for choosing a business
My notes on Alex Hormozi's talk sorting businesses into five types by how hard they are to run and how much they can make, with his facts checked.
-
How Andrej Karpathy uses AI agents day to day
My notes on how Andrej Karpathy works with Claude Code, from Nick Saraev's video and Karpathy's own posts: short prompts, pictures, cheap filters, shared data.
-
Fiber feeds your colon cells: butyrate and how to eat for it
My notes on Hank Green's talk with Liam Layton on why colon cells run on butyrate from fermented fiber, and the easy ways to eat more of the right kinds.
-
How agents improve over time: model, harness, context
My notes on Harrison Chase's talk on the three parts of an agent you can improve (model, harness and context) and why traces and evals sit under all three.
-
Russell Raath on finding consulting clients on LinkedIn
My notes on Russell Raath's talk on LinkedIn for consultants: watch for trigger events, write about live news, and learn the words buyers use for the problem.
-
What's left for engineers when AI writes the code
My notes on Laurie Voss's talk on where the bottleneck goes once agents write the code, and why deciding what good means stays human, with the numbers checked.
-
Dan Gelbart on patent strategy in the age of AI
My notes on Dan Gelbart's patents talk: how AI search changes the answer to an infringement letter, and how to write a patent that holds up.
-
Jake Van Clief on handing automation back to people
My notes on Jake Van Clief's talk on what to build now that AI can do the work: AI as one part of a program, and why your own instructions matter.
-
Russell Raath on which Hormozi advice works for consultants
My notes on Russell Raath's talk adapting Alex Hormozi's service-business advice to consulting, a trust and risk business: offer, price, market, sales, content.
-
Blueberries and your health: what a cup a day does
My notes on Rhonda Patrick's clip on blueberries and arteries, plus the trials and long-term studies on arteries, blood sugar and memory.
-
Inside the Box: David Epstein on why constraints beat goals
My notes on David Epstein's talk for Inside the Box: why limits on how you work beat a bigger goal when the path isn't clear, his four reasons and examples.
-
Copy what works: Pabrai's method and the research behind it
My notes on how Mohnish Pabrai built his career by copying Warren Buffett, plus the research on why copying a proven method works and how to do it well.
-
Checking my coding agents' 'tests pass' claims against their logs
A log check flagged 16 of 31 Claude Code turns that ended on a failing test.
-
Testing a picture map that looks good: four checks from rebuilding Graph Starz
Four tests I ran on Graph Starz, a map of pictures: replaying uploads, a bulk import, scoring 16 picture models against museum records, and blind judging.