Rabat
Tap a glowing node to enter
Ahmed · The Library · Workshop · The Path
The Path · branch 3 of 3
Challenge your preconceptions, or they will challenge you.
An honest map of where this work leads. Built from current evidence. Stress-tested against history. Revised every six months.
The shift
For most of computing’s history, building software meant assembling the pieces by hand — writing syntax line by line, fitting each function into the larger whole. A computer-science curriculum taught that assembly. The discipline rewarded patience, memorisation, and compliance with old conventions.
The industry split in 2022. The lower layer — syntax, boilerplate, the parts the machines now write faster than any human — has automated. The upper layer — architecture, product judgment, distribution, taste — has never been more valuable. The shift is generational. It isn’t yet visible in the curricula because curricula update slowly. It is sharply visible in who gets hired, paid, and funded. The path drawn here is built for the upper layer.
The map
Each phase has measurable outcomes. None depend on luck. All depend on one thing: holding the cadence, week after week.
Three parallel commitments, kept every week. Ship products that real people use. Write publicly on a domain I own. File monthly predictions with dates and confidence so calibration can be measured later. The work generates the artifacts; the artifacts become the credentials.
With six to twelve months of public artifacts, file three to five applications in parallel. Anthropic Fellows. MATS. South Park Commons. Entrepreneur First. Y Combinator with traction. The artifacts are what get through the door — without them, applications are wishes. By 2028, at least one has fired.
Whichever door opened, the next three years compound the position. Inside a frontier lab: ship research engineering that travels, build the network laterally, hold the seat where roadmaps get set. Inside a startup: hit traction, learn distribution, survive the chasm, reach Series A or sustainable bootstrap. The Phase-1 cadence does not pause — it scales with the platform.
By year five the compounding shows. By year ten, the public artifact record, the calibrated prediction track, and the multi-wave network let the next move be chosen from optionality — research lead, founder, fund partner, or independent operator. Each new wave gets evaluated against an existing thesis; the network can introduce you to whoever is building it; the track record makes you fundable, hireable, followable into it.
The doors
Each of these is a real program with a real URL. The path is not vague — it is paved.
Anthropic Fellows
Built for developers without a PhD. Over 40% of the first cohort joined the company full-time.
$3,850/wk · 4 months
MATS — Berkeley & London
Largest cohort yet: 120 fellows, top mentors, direct recruitment by alignment organisations.
Summer + Autumn 2026
South Park Commons
The “between things” door — six months in San Francisco, New York, or Bangalore, for people still exploring what to build.
Rolling applications
Entrepreneur First — The Bridge
Eight-week SF residency for European technologists. 80% find a cofounder.
Continuous
Y Combinator
The standard. $500K, three months in San Francisco, the deepest founder network in the world.
Apply with traction
Solo Founders Program
$100K for 2.5%, three months in San Francisco — built for builders without a cofounder.
Closes September
OpenAI Residency
Six months in San Francisco, ~$18,300/month. Reopens to applicants late 2026.
Late 2026
Where this leads
These are conditional probabilities for the cohort that sustains the discipline through five years. They are not promises — they are the realistic distribution.
Five-year outcome distribution
Ten-year median for the disciplined cohort: cumulative net worth of $10–30M, multi-time operator status, audience of 5K–100K, first-call expert in at least one niche, roughly 500 resolved predictions and 100 published essays.
At a twenty-year horizon, the disciplined cohort splits into three quartiles. Top: $50–500M+ net worth, multi-wave operator at scale. Middle: $10–50M, multi-time founder, domain-prominent. Bottom: $3–10M, modest exits but durably positioned. The floor is unusually high because multi-wave literacy, prediction track, audience, and network stay valuable even when no single venture connects.
For context: a physician’s peak lifetime earnings in this region settle in the €60,000–150,000/year range after ten to twelve years of training. A barrister’s path is similar. Both are real, honourable, worthy. The path presented here carries different variance — wider distribution, higher ceiling, comparable floor for the disciplined cohort. Different paths, comparable destinations.
An honest note
This map promises a structure — four phases, named doors, measurable outcomes, a sustainable discipline. It does not promise that every prediction will be right. Some of the specific companies, programs, and timelines listed here will look different in five years. The framework is robust to that. What stays invariant: artifacts shipped in public, writing weekly, predictions tracked, applications filed with leverage rather than wishes. Everything else adjusts. The map gets diffed against reality every six months, and revised honestly.
Built from current evidence. Stress-tested against history. Open to revision.
The Library
Workshop · project