Founder OS · Kent Langley

Blog

Research, analysis, essays, and field notes from Founder OS.

Read me first

These are my thoughts. Opinions, working theories, experiments, the occasional crazy idea, who knows what. Some pieces are finished. Others are incomplete or still in flight, and they will change as I keep learning. This page is the canonical source for all of it.

New pieces also go out through factually, my newsletter. Subscribe at news.kentlangley.com, or follow along by RSS.

July 25, 2026Essay

The Philosopher Builder

AI collapsed software engineering, product, and design into one circle. The person you want to hire stands dead center. Screen for traits instead of titles, test with projects instead of interviews, and build the few.

June 29, 2026Analysis

Self-Liquidating Token Spend

You can recite the token bill. You cannot see what it returned. Here is the half nobody publishes, and how to put a number on it this quarter.

June 28, 2026Analysis

Master the Mix: Open Weights for Operators

Open weights, read for an operator. The move is not picking the smartest model. It is mastering the mix: route each job to the model that wins it, closed frontier for the hard fifth, open weights for the rest, on terms you own.

June 17, 2026Essay

Your Job Panic Is 10,000 Years Old

AI job panic set against ten thousand years of technology waves, from the cooking fire to the switchboard. The catch the cheerful version hides: the aggregate recovers, but the displaced cohort often does not. Two questions, not one, and a source behind every claim.

June 12, 2026Essay

Wealth Is Context

Data is not the asset. Context is: data your AI can both reach and trust. Why capability is rented, context is owned, and the moat is elapsed time.

June 12, 2026Analysis

Experience, Repriced

Grading a 2024 essay on AI and the skilled generalist, twenty-two months later. Experience was not commoditized, it was repriced: the legible half went to zero and the illegible half absorbed the premium.

June 10, 2026Essay

The Expert Costume

Sounding like an expert just became free; being one did not. What an expert actually is, why expertise still requires experience, and why judgment with scar tissue is the one asset this era rewards.

June 8, 2026Analysis

Who Has More Compute? The Frontier Labs, Ranked for Service Delivery

A mid-2026 snapshot of frontier-lab compute, ranked for consistent service delivery. The announced-versus-online gap is the whole story: Google DeepMind and Anthropic lead, and almost everyone is smaller than their headlines.

June 8, 2026Post

AI Is the New Other

Civilization began when we stopped fearing the stranger and started sharing with him. AI is the new Other, the same fear reflex is firing, and the next leap belongs to whoever breaks it first.

May 28, 2026Research

Hello, World, Five Ways

Five Hello, World programs written and verified through a three-stage agent pipeline. Three ran for real, two passed by inspection, and the line counts trace how much ceremony each language demands.

May 25, 2026Analysis

The Power Plant Above Your Head

SpaceX already has roughly 110 megawatts of solar capacity in orbit, and the production curve points at hundreds of gigawatts by 2035. The largest power plant ever conceived is not on Earth.

May 21, 2026Post

Founder Pay: Where It Actually Comes From

Founder pay is a downstream effect of the value loop working, not a line item the company owes you. Trace the chain backwards, then run your own numbers with the PS worksheet.

May 21, 2026Post

AI Layoffs in Historical Context

AI layoffs set against ten thousand years of technology waves, from fire and the printing press to switchboard operators and ATMs. Written for a WhatsApp group, with a source behind every claim.

May 18, 2026Analysis

Take-Two's AI Mistake Wasn't Hiring the Wrong People. It Was Firing the Right Ones.

Take-Two put all of AI under its CTO six weeks after dissolving the team that wired AI into Rockstar and 2K workflows. A structural breakdown of that decision, plus the full X article it produced.

May 16, 2026Working Note

AI-Native Operations

An outbound email draft published as a working note. Most AI rollouts leave the operating model untouched; the real divide is AI-native companies versus companies bolting AI onto legacy behavior.

May 13, 2026Book Excerpt

EVOLVED, Second Edition: Your Digital Nervous System

An excerpt from the second edition of EVOLVED: the preface and early chapters on the Digital Nervous System, the centaur principle, the Semantic Nervous System, and the first installable loop.

May 6, 2026Newsletter

The Greek Premium

LLMs collapsed the price of competent output and the market is repricing judgment. The Socratic, Aristotelian, and Platonic premiums name the work that just got expensive, plus three moves for Monday.

May 4, 2026Newsletter

The AI Bully Problem

Most of your team hides its AI use because mocking it gets tolerated. The stigma is a training failure compounded by leadership silence, and it follows a pattern that has reversed every time for 500 years.

May 3, 2026Post

The Bottleneck Moved, and History Already Told You Where

Companies are cutting engineers and hiring engineers in the same week. That is not a contradiction. It is a bottleneck shift, and every prior technology says the jobs return with different titles.

April 28, 2026Newsletter

Leverage You Operate With Words

AI is the first lever where the interface is plain language. The access fight is over, the constraint is taste, and the operators are anyone willing to ask clearly.

April 28, 2026Analysis

The AI Correction, Not the AI Collapse

AI job loss headlines describe a market correction, not a collapse. Engineering demand is moving from a few oversized employers to thousands of small companies that were priced out of building software until now.

April 24, 2026Analysis

AI Thinking Patterns Compared

Three AI thinking patterns from one ninety-minute session: go-for-the-jugular, mycelial accumulation, and the sculptor. Only one scales, and the failure modes of the other two are predictable in advance.

April 16, 2026Essay

The 500-Year Pattern

Five hundred years of technology waves point one direction in aggregate: employment recovers and new work appears, while specific cohorts still pay the price. Why AI will not break the pattern, and the founder instruction it yields.

April 15, 2026Essay

The Split Screen

The EU bans what Texas ignores while generative AI hits 53 percent adoption in three years. What the regulatory split screen means for founder-operators, and the five moves that hedge it.

April 15, 2026Field Notes

No Job? Make One.

What our household teaches that school does not: earn before you spend, build a company when no job exists, and treat craft and AI as everyday tools. The classroom socializes. The household builds capability.

April 1, 2026Research

Does Technology's Good Outweigh Its Bad?

A research compendium on whether technology nets out positive for humanity: offense-defense theory, Pinker's optimism, existential risk, and the recurring pattern of offense spiking first while defense catches up.

March 31, 2026Analysis

The Only Moats That Matter: What Bloch's Framework Means for Founder-Operators

Michael Bloch draws a line that most founders will find uncomfortable. AI is collapsing everything on the 'hard to do' side. But some things resist compression. Here is what survives and what to do about it.

March 25, 2026Guide

How to Build a Claude Cowork Plugin for Google Ads Management

Reverse-engineered from the solo non-technical operator who ran Anthropic's entire growth marketing department with agentic AI workflows. A step-by-step guide to building a Claude Cowork plugin for Google Ads.

March 24, 2026Analysis

Intercom Gave Claude Code to 1,000 Non-Engineers. Here Is What They Built.

Intercom deployed Claude Code to 1,000+ non-engineers, built guidance skills, and got 300+ weekly active users. They built exactly what Founder OS ships on day one.

March 22, 2026AI News Briefing

AI News Briefing: March 22, 2026

Six AI stories that matter for founder-operators this week: Anthropic winning enterprise spend, space data centers, $189B in funding, Google making Gemini free, data center backlash, and Atlassian's AI restructuring.

March 21, 2026Special Edition

The Claude Blitz: How Anthropic Built a Platform in 60 Days

While everyone watched OpenAI announce a superapp, Anthropic quietly shipped one. 20+ releases in 60 days turned Claude from a chatbot into always-on infrastructure. Here is what happened, what it means, and what to do about it.

March 18, 2026AI Leverage

Skills, Agents, and the Orchestra: A Practical Guide to Building Your AI Team

A practical guide to understanding skills versus agents when building your AI team. The orchestra analogy, four levels of AI usage, and how skill graphs create compound leverage.

March 12, 2026Research

The Niantic Data Loophole

How Pokémon GO became a robot navigation company without ever selling anyone's data. The legal mechanism, the COPPA gaps, and what it means for every founder sitting on a user dataset.

March 12, 2026Field Notes

The Best LLM Isn't the Smartest One

Why the right AI co-worker isn't about benchmarks. It's about fit. A founder-operator's field notes on building a working relationship with AI.

March 5, 2026Research Report

The Future of the Web in the Age of AI

AI agents are booking flights without visiting airline websites. Google answers queries without sending users anywhere. The web is thirty-three years old, and it has never faced a disruption this fundamental. Four-part research report covering agentic browsers, the visibility crisis, static site resilience, and five contrarian counterarguments.

Founder OS · Updated 2026-07-25 · Instance: factual · Project: fos-www-blog
Skills applied: designing-fos, writing-copy, navigating-skills
fos.kentlangley.com