Six hours over three weekdays. Eight builders. Whether you're new to coding and tried "vibe coding," or you're an experienced coder new to agents: bring a fresh idea you haven't started yet (no existing repo, no half-built code) and leave with code worth keeping. A foundation you can extend, not a folder of agent output you'll be afraid to touch.
You've watched the demos. You've installed the tools. Maybe you "vibe coded" something that even sort of runs. But the project feels fragile, you can't extend it without something breaking, and you don't have a method you trust. Just a folder of agent output and a slowly growing suspicion.
The problem isn't the tools. The tools are extraordinary. The problem is that nobody has shown you how to drive them so the project compounds across sessions instead of unraveling. That's what Liftoff is for.
Three two-hour live sessions over three weekdays. Eight builders. Each of you brings one greenfield idea: a real concept that interests you, but one you haven't started building yet. We turn it into a foundation the agent can actually build well against. You go drive your own agents between sessions. Then we come back together, debug what happened, and lock in the momentum.
This isn't a polish-my-existing-project clinic. The whole point of Liftoff is starting clean: no legacy slop to apologize for, no half-built repo to defend, just an idea, the right method, and three days of pressure. You get direct feedback on your concept from a principal architect, and you watch seven other ideas take shape in real time. That second half is where most of the pattern recognition actually happens.
"Vibe coding" is an antipattern. I list it as one in the Encyclopedia of Agentic Coding Patterns. We replace it with a small set of artifacts the agent reads before it writes a line: a written brief, a one-page architecture sketch, and a tiny set of repo policies. The same setup running on PalinDragon and Bartley.
The core teaching. We phase the work so context stays useful, we run a planner / planning-critic / coder / coding-critic loop so the agent's own output gets challenged before it lands in the project, and we keep an append-only log of what got built and why. Confident-and-wrong stops compounding into the foundation.
The agent does the typing. You keep the judgment. No green test run, no agent self-report, and no orchestrator gets to call a phase "done" on its own. You leave with a rhythm where the agent moves fast and you stay the one in command.
Each session is two hours, live on Zoom, with one project on the hot seat at a time. You watch each other work. You get direct feedback on yours. Between sessions you actually run the agents, alone or with the cohort channel open.
2 hours · live · everyone gets a hot seat
2 hours · live · "what actually happened?"
2 hours · live · "now keep going"
The last two of those have been with agents. Every day, in production, on projects he ships under his own name.
Wolf has been a principal-level technologist since before the web. At Apple, Walt Disney Imagineering, Electronic Arts, Activision, Warner Bros., Volvo, Northrop Grumman, and eHarmony, the assignment was the same: hard problem in, working invention out.
He's the principal behind Flying Logic (visual systems thinking software, born from Theory of Constraints and military course-of-action planning) and the cryptography and identity work at Blockchain Commons, including Gordian Envelope, LifeHash, and Multipart URs.
In the last two years he's used agentic coding to ship four projects under the Locked Lab banner. Each one is a real product solving a real problem, not a tutorial:
When Wolf teaches agentic coding, the lessons come from real production work with agents this quarter, antipatterns and all. Not from someone else's tutorial last year.
No tier games. One cohort, one price. Here's what's on the table.
2 hours each, on Zoom, with you and 7 other builders. Every seat gets time on the hot seat. Not a webinar. A clinic.
Before Day 1, you submit a one-pager about your project. Wolf reads every single one and sends back a written critique. You walk into Day 1 already moving.
The same master template as the free starter (CLAUDE.md, briefs, policies, planner/critic/coder/critic orchestrator) but during the clinic Wolf personally tunes it to your specific language, stack, editor, and idea, and we run your first phase end-to-end live. You leave with the configured-for-you version and one successful phase already on disk.
The brief template ships in the free starter. In Day 1 we fill it out for your idea, with Wolf's eyes on the structure and the specificity. You leave with a project brief tight enough for an agent to build against and loose enough to be surprised in the right places.
All three sessions recorded and transcribed. Searchable. Yours to keep. Watch your own hot-seat moment back, or anyone else's.
A channel just for your eight, plus Wolf, running during the clinic and for 14 days after. Stuck at 11pm? Drop it in.
50 of the most useful patterns from the Encyclopedia of Agentic Coding Patterns, curated for builders who want to stop producing slop and start producing code worth keeping. (Yes, vibe coding is in here. As an antipattern.)
Once you're in a cohort you're in forever. Future cohorts feed in. New patterns get shared. Wolf shows up.
This is what the components would run as standalone offers.
Show up to all three sessions. Do the work between them. If by the end of Day 3 you haven't made meaningful progress on your idea (progress you can actually keep building on) send me one email and I'll refund every dollar. No forms, no panel, no haggle. I'm the one accountable here.
And if life genuinely intervenes and you can't attend, you roll into the next cohort at no extra charge. Seat held, money applied.
Some clinics try to be for everyone. This one isn't. Here's the actual line.
3 × 2-hour sessions on Zoom across Tue / Wed / Thu of one week. Sessions are live, recorded, and transcribed. Cohort channel runs during & for 14 days after.
This isn't a marketing number. The clinic format breaks past 8. When seats are full, the cohort closes and a waitlist opens for Cohort 02.
The first cohort runs the moment 8 builders reserve seats. Dates get locked with the cohort. Reserve a seat to lock in founding pricing and get input on the dates.
Before Day 1: fill in a one-pager about your project (template provided), install the tool of your choice, and read Wolf's written critique of your brief.
No. There's no lecture track, no slide deck to grind through, no certification. It's a working clinic: you bring an idea and we move it forward together with agentic tools. If you want a 101 course, this is genuinely not the right room.
Any one of Claude Code, Codex, Cursor, or Copilot. Wolf works primarily in Claude Code, but the patterns and the way of working translate across all of them. Bring the tool you already have. We'll fit the workflow to your setup, not the other way around.
You should be comfortable enough that "git pull," "run this script," and "look at this error" don't stop you in your tracks. You don't need to be a 10-year engineer. Two audiences fit best: people new to coding who tried "vibe coding" and want to do it without producing junk, and experienced coders who are new to working with agents. If you're in either of those, you're ready.
It's an antipattern, named in my Encyclopedia of Agentic Coding Patterns. No brief. No architecture. No policies the agent has to respect. No record of what got built or why. The result feels like progress and isn't. You end up with a project that breaks when you push on it and that nobody (including the agent next session) can pick back up. Liftoff is largely about replacing vibe coding with a small, real method that produces code you can actually keep building on, without you having to read it line by line to know whether it's holding up.
A greenfield concept that interests you. A web app you've been thinking about, a small game, a CLI tool you wish existed, a workflow you wish you had, a data pipeline, a writing tool, a niche internal product. One hard rule: no existing repo, no half-built code, no project you've already invested in. Liftoff is about starting clean. Vague at the concept stage is fine. Sharpening it is part of Day 1.
No, and that's deliberate. Liftoff teaches agentic coding; it isn't a clinic to save, finish, or polish work you've already done. Existing projects carry baggage that gets in the way of teaching the method. If you have a stuck project, the right move is to set it aside, bring a fresh idea, learn the method here, then go back to your project armed with it.
Probably not, and that isn't the promise. Liftoff is named the way it is for a reason: three days gets you off the ground with momentum and a method you can keep using. Shipping a complete product is weeks or months of work; this clinic makes those weeks productive instead of churn.
Sessions are recorded. But missing one degrades the format; half the value is the room. If life genuinely intervenes, you roll into the next cohort at no extra charge. Just email me.
Show up to all three sessions and do the work between them. If by the end of Day 3 you haven't made meaningful progress on your idea (progress you can actually keep building on) send me one email and I'll refund 100%. No forms. No haggling. I'm one accountable principal. That's the whole point.
Yes. When eight seats are reserved, checkout closes and Cohort 02 opens for waitlist. The clinic format genuinely breaks past 8. There's no productive way to give nine people meaningful hot-seat time in 2 hours.
The Stripe receipt is invoice-friendly. If you need a specific format or a W-9, email me before booking and I'll send what you need. Multiple seats from one company is fine, but they each take a real seat.
Yours. The point is that your idea moves forward. Wolf brings his stack as a reference and a starting point, and we adapt it to your project, your editor, your language, your repo.
Almost certainly, but Cohort 01 has a founding price ($500) that won't repeat. Each cohort is real work and a fixed time commitment from Wolf, so they'll be infrequent. Reserving Cohort 01 also gives you priority for future cohorts.
No. The starter is free, with a 5-day email walkthrough and access to the alumni Discord at guest tier. The clinic is for when you want my eyes on your specific idea, the failure-mode pattern recognition the repo can't give you. The starter works on any project; the clinic is greenfield-only. Two different things, both useful.
The cohort is the application. The foundation is the same repo Wolf uses on his own shipping projects: the eleven-step methodology, the four canonical agents, the planner / reviewer / coder / critic loop. Free. Yours in a minute, plus a five-day email walkthrough to get you running.
The agentic era didn't make building easier. It made the cost of trying near zero. The hard part now is turning that cheap trying into something you can actually keep building on. That's the whole point of Liftoff.
Cohort 01 · 8 seats · $500 · real-progress guarantee.
Reserve My Seat →Secure checkout · Stripe · 100% refund if you don't make real progress