Essays on technology, startups, and the leverage of small teams.https://craigteich.com/en-ushttps://craigteich.com/essays/we-rebuilt-my-website-this-afternoon-meet-the-team/https://craigteich.com/essays/we-rebuilt-my-website-this-afternoon-meet-the-team/I didn't hire an agency or pull an all-nighter. I directed a team of six — a chief of staff, a designer, a developer, a researcher, a writer, and an engineer — none of whom are people. Here is the unedited story, in their own words.Fri, 19 Jun 2026 00:00:00 GMThttps://craigteich.com/essays/founder-who-manages-machines/https://craigteich.com/essays/founder-who-manages-machines/When most of your workforce is agents, the founder's job shifts from directing people to designing the conditions under which judgment happens. The old leadership virtues survive the translation; the tactics mostly don't.Tue, 09 Jun 2026 00:00:00 GMThttps://craigteich.com/essays/slow-knowledge-in-a-fast-machine/https://craigteich.com/essays/slow-knowledge-in-a-fast-machine/Output stopped being scarce. The bottleneck moved to the slow, lived-in kind of knowledge that no machine can pour out fast — and that is now the thing worth guarding.Tue, 21 Apr 2026 00:00:00 GMThttps://craigteich.com/essays/the-last-human-skill/https://craigteich.com/essays/the-last-human-skill/Machines have absorbed the analysis. What remains is the harder, stranger work of deciding which problems are worth solving at all — a judgment no model can make, because it has nothing at stake.Tue, 10 Feb 2026 00:00:00 GMThttps://craigteich.com/essays/the-round-you-dont-raise/https://craigteich.com/essays/the-round-you-dont-raise/For the first time in my career, the most ambitious thing a founder can do is often not to raise at all. The economics that made venture capital mandatory have quietly stopped applying.Tue, 27 Jan 2026 00:00:00 GMThttps://craigteich.com/essays/what-we-owe-the-people-we-automate-around/https://craigteich.com/essays/what-we-owe-the-people-we-automate-around/Automation doesn't happen to jobs. It happens to people. A builder who refuses to look at that directly isn't being neutral — he's being a coward with extra steps.Tue, 16 Dec 2025 00:00:00 GMThttps://craigteich.com/essays/company-as-an-argument/https://craigteich.com/essays/company-as-an-argument/The most honest way to claim how work will change is to build the company that only makes sense if you're right — so the product becomes the proof, and every failure is in public.Tue, 04 Nov 2025 00:00:00 GMThttps://craigteich.com/essays/taste-of-the-person-who-asked/https://craigteich.com/essays/taste-of-the-person-who-asked/When the model is the same for everyone, the only variable left is the person at the controls. The prompt is a mirror, and the work still looks like its author.Tue, 30 Sep 2025 00:00:00 GMThttps://craigteich.com/essays/death-of-the-status-meeting/https://craigteich.com/essays/death-of-the-status-meeting/Most meetings exist to move information between people. When the information learns to move itself, the only meetings left standing are the ones that were never really about information at all.Tue, 26 Aug 2025 00:00:00 GMThttps://craigteich.com/essays/raising-capital-good-money-vs-bad-money/https://craigteich.com/essays/raising-capital-good-money-vs-bad-money/Every dollar of venture money arrives attached to a person, and you will know that person for longer than you will know most of the people you marry into. Choose accordingly.Tue, 19 Aug 2025 00:00:00 GMThttps://craigteich.com/essays/verification-is-the-new-literacy/https://craigteich.com/essays/verification-is-the-new-literacy/When generation goes to zero, the scarce skill is checking. The worker who can quickly tell true from plausible is worth more than the one who can produce.Tue, 17 Jun 2025 00:00:00 GMThttps://craigteich.com/essays/employee-who-builds-their-own-team/https://craigteich.com/essays/employee-who-builds-their-own-team/The next org chart won't stack agents above or below people. It will wrap a small team of machines around each person, and make the individual-with-their-agents the smallest unit of work.Tue, 29 Apr 2025 00:00:00 GMThttps://craigteich.com/essays/who-signs-for-the-robot/https://craigteich.com/essays/who-signs-for-the-robot/Accountability does not survive being spread across software. Before an agent is allowed to act, an organization owes itself the name of the person who answers for it.Tue, 11 Mar 2025 00:00:00 GMThttps://craigteich.com/essays/year-tools-started-doing-things/https://craigteich.com/essays/year-tools-started-doing-things/Software spent thirty years answering questions and waiting to be told what to do next. The quiet shift of the last year is that it stopped waiting — and the moment a tool can act, the only question that matters is what we let it do.Tue, 28 Jan 2025 00:00:00 GMThttps://craigteich.com/essays/cost-of-a-decision/https://craigteich.com/essays/cost-of-a-decision/When trying any option becomes nearly free, choosing carefully stops being the hard part. The scarce act is committing to a path and living with it.Tue, 10 Dec 2024 00:00:00 GMThttps://craigteich.com/essays/permission-to-be-bad-again/https://craigteich.com/essays/permission-to-be-bad-again/The tools that make us instantly competent at everything are quietly retiring the one thing that ever made us grow: the discomfort of being a beginner.Tue, 12 Nov 2024 00:00:00 GMThttps://craigteich.com/essays/reading-the-machine/https://craigteich.com/essays/reading-the-machine/Working well with a model isn't a technical skill. It's the same skill that lets a good manager get the best out of a brilliant, overconfident, forgetful new hire.Tue, 24 Sep 2024 00:00:00 GMThttps://craigteich.com/essays/the-apprentice-problem/https://craigteich.com/essays/the-apprentice-problem/Machines are getting good at the junior work — and the junior work is exactly how we used to grow seniors. If we automate the bottom of the staircase, we have to decide how anyone climbs.Tue, 16 Jul 2024 00:00:00 GMThttps://craigteich.com/essays/cheap-words/https://craigteich.com/essays/cheap-words/When any machine can produce fluent prose on demand, fluency stops being a signal. What survives the flood is the small, costly act of meaning what you say.Tue, 07 May 2024 00:00:00 GMThttps://craigteich.com/essays/how-much-is-enough/https://craigteich.com/essays/how-much-is-enough/The most expensive decision in a raise is not the valuation. It is the size of the round — the number that quietly decides what kind of company you are allowed to become.Thu, 21 Mar 2024 00:00:00 GMThttps://craigteich.com/essays/org-chart-is-a-communication-diagram/https://craigteich.com/essays/org-chart-is-a-communication-diagram/Companies are shaped by what it costs them to pass information around. When that cost falls, the shape of work stops being fixed and becomes a choice.Tue, 19 Mar 2024 00:00:00 GMThttps://craigteich.com/essays/demo-is-not-the-product/https://craigteich.com/essays/demo-is-not-the-product/A demo proves a thing can happen once. A product proves it happens every time. The distance between the two is where most startups quietly die.Tue, 23 Jan 2024 00:00:00 GMThttps://craigteich.com/essays/stopped-trusting-my-memory/https://craigteich.com/essays/stopped-trusting-my-memory/Forgetting isn't a flaw to fix with willpower. It's a plumbing problem, and we finally have the pipe.Tue, 14 Nov 2023 00:00:00 GMThttps://craigteich.com/essays/generalists-quiet-comeback/https://craigteich.com/essays/generalists-quiet-comeback/When tools make any single skill cheap, the scarce thing becomes the person who can stand between domains and translate. A case for breadth, written by someone who was told for decades that breadth was a flaw.Tue, 12 Sep 2023 00:00:00 GMThttps://craigteich.com/essays/spreadsheet-and-conviction/https://craigteich.com/essays/spreadsheet-and-conviction/A financial model isn't a prediction. It's a structured argument with yourself, and the only thing it really outputs is conviction — or the absence of it.Tue, 27 Jun 2023 00:00:00 GMThttps://craigteich.com/essays/raising-on-a-sentence/https://craigteich.com/essays/raising-on-a-sentence/In the year after GPT-4, capital chased a sentence. The founders who raised well were not the ones with the best sentence — they were the ones who knew what the sentence was costing them.Wed, 14 Jun 2023 00:00:00 GMThttps://craigteich.com/essays/the-interruption-tax/https://craigteich.com/essays/the-interruption-tax/Knowledge work isn't lost to laziness. It's lost in the seconds after every interruption, while the mind reboils from cold. Protecting deep blocks is the most leveraged thing a manager does all week.Tue, 09 May 2023 00:00:00 GMThttps://craigteich.com/essays/on-taste/https://craigteich.com/essays/on-taste/Generative tools have made competent work cheap and instant. The scarce thing left is knowing which competent thing is good — and that you have to build yourself.Thu, 30 Mar 2023 00:00:00 GMThttps://craigteich.com/essays/small-teams-move-like-water/https://craigteich.com/essays/small-teams-move-like-water/A team's speed is set not by how good its people are but by how few of them have to agree before something ships. Most companies optimize the wrong variable.Tue, 21 Feb 2023 00:00:00 GMThttps://craigteich.com/essays/tyranny-of-the-blank-file/https://craigteich.com/essays/tyranny-of-the-blank-file/The hardest part of building anything is the moment before the first line exists. The cure is to make something bad on purpose.Wed, 25 Jan 2023 00:00:00 GMT