# iamladi.dev

> Ladi (Ladislav Martincik) — 18 years shipping software, ex-CTO, now a solo builder betting the next decade on agents. This site is a living build log with two faces: the designed pages are for humans, this file is the same site for agents.

Every essay and every project dossier is also available as raw Markdown at the same URL with a `.md` suffix (e.g. `/writing/ui-is-dying.md`, `/projects/loombrain.md`).
The full text of everything — all essays and all post-mortems in one file: https://iamladi.dev/llms-full.txt

## The bets
- **the interface bet** — UI is not dying. It is splitting in two. Task screens get generated. The trust layer is the moat. The task interface gets generated per prompt and thrown away. What is left worth owning is the trust layer (where you check and steer the agent) and the substrate beneath it. So that is what I build. This site dogfoods the split: a designed face for people, an agent-readable twin at /llms.txt. (https://iamladi.dev/writing/ui-is-dying)
- **the taste bet** — Everyone says taste is the one thing AI can't take from you. That's the cope. Nobody can define taste, teach it, or prove who has it. Strip the mystique and it's just an evaluation function: what tells a feedback loop ‘better’ from ‘worse.’ Whoever runs the loop fastest grows it fastest, and the models are already growing their own. (https://iamladi.dev/writing/taste-is-the-cope)
- **the craft bet** — Stop coding by hand. Typing speed and a clever editor setup are a skill that is now depreciating. The durable skill is directing the loop in English and owning the system that catches wrong: reviews, tests, staging, backups. Understanding does not vanish; it moves up to what good looks like and out into a verification you trust. Drop the keystrokes, not the judgment. (https://iamladi.dev/writing/stop-coding)

## Projects — every endeavor with its reasons and post-mortem
Each page covers: what it is, why it exists, what happened (revenue, failure, or death), and the lesson kept. Honest statuses — killed and shelved projects stay published.
- [Save Bulk Gmail Attachments](https://iamladi.dev/projects/save-bulk-gmail-attachments) [revenue, 2025–now]: My first product built entirely by directing AI in English, and the one that pays for itself. (1,280 users · ★4.33 · revenue verified live via Stripe on TrustMRR)
- [The Book of Aaru](https://iamladi.dev/projects/book-of-aaru) [shipped, 2021–2025]: Five years, a small studio, a shipped Steam roguelite, and an honest ending. (Released on Steam May 2025 · 86% positive reviews)
- [LoomBrain](https://iamladi.dev/projects/loombrain) [live, 2025–now]: Everyone builds local-first second brains. I want the opposite. (Live, free, multi-tenant · I use it every day)
- [Access Protocol](https://iamladi.dev/projects/access-protocol) [shipped, 2022–2026]: Four years, founding week to sabbatical, and an AI pivot lived from the inside. (Employer · founding engineer through the full arc)
- [sandcaster](https://iamladi.dev/projects/sandcaster) [active, 2026]: Data-driven agent sandboxes: the agent is a config file, not code to wrap. (Open source · CLI + SDK on npm · MIT)
- [SelfUnderstand](https://iamladi.dev/projects/selfunderstand) [paused, 2026]: A psychometric interviewer you talk to, paused on a hard voice problem. (Live · 36-facet LLM-grounded assessment · spend caps + eval harness)
- [cautious-computing-machine](https://iamladi.dev/projects/cautious-computing-machine) [active, 2025–now]: My Claude Code plugin marketplace: how I work, published. (Installable marketplace · 7 repos · used daily)
- [saashunting](https://iamladi.dev/projects/saashunting) [active, 2025–now]: A machine that finds problems to work on. A machine that builds machines. (Its signals produced PermitFlag and ShopTrust)
- [PermitFlag](https://iamladi.dev/projects/permitflag) [live, 2026]: Idea to paid product in four days, and my most useful failure. (Live on the Chrome Web Store · zero paying users at $19/mo)
- [Pentune](https://iamladi.dev/projects/pentune) [killed, 2026]: An AI that learns your voice from your corrections. Killed with a clear head. (Full product + company-vision docs · never launched)
- [Refocus](https://iamladi.dev/projects/refocus) [shelved, 2025–2026]: A focus coach that talks you back to work. Cool product; it never helped me. (Working native macOS app · still online)
- [React Native OSS era](https://iamladi.dev/projects/react-native-oss) [shipped, 2016–2018]: Native bridges from production apps, open-sourced. One became the standard. (★121 · still downloaded from npm a decade later)
- [Czech agent commerce](https://iamladi.dev/projects/czech-agent-commerce) [active, 2025–now]: Teaching agents to buy things and mail letters, starting with the Czech market. (Working Medusa adapter pack · Postbot pending Česká pošta API credentials)
- [Four macOS menu-bar tools](https://iamladi.dev/projects/macos-utilities) [live, 2025–2026]: Tools nobody ships because they're too small to sell and too useful not to build. (All four in my menu bar right now)
- [X Articles tools](https://iamladi.dev/projects/x-articles) [live, 2026]: Export X articles to Markdown: built for my AI, adopted by strangers. (500+ users on the exporter)
- [newadmin](https://iamladi.dev/projects/newadmin) [active, 2026–now]: Rebuilding my brother's ISP back-office from interviews and database archaeology. (In active development, the thing I build this month)
- [LivingPortrait](https://iamladi.dev/projects/livingportrait) [paused, 2026]: Harry Potter portraits of your family, on your phone. A friend's wish, prototyped in days. (Working prototype incl. payments, never submitted to the App Store)
- [urbanlist.cz](https://iamladi.dev/projects/urbanlist) [active, 2026–now]: Learning AI video production on a real client: property marketing that shows, not tells. (First construction time-lapse delivered · more in production)

## Writing
- [We rebuilt the whole game once. The mistakes survived the rebuild.](https://iamladi.dev/writing/aaru-post-mortem): The Book of Aaru took five years, one full restart, and a small studio to ship. I already wrote the goodbye; this is the build: where the years actually went, why restarting fixed the assets but not the decisions, how the pace of AI made me rethink the whole path, and the 1%-loop lesson I'd run any project on now.
- [Everyone builds local-first second brains. I'm building the opposite.](https://iamladi.dev/writing/loombrain-build-log): Local-first is the sacred cow of second brains: files on your disk, plain text, outlives every app. I think it's the wrong architecture for the agent era, and I'm betting LoomBrain on the opposite: a brain that lives on the internet so agents can reach it while my computer sleeps. Build log: three ground-up rebuilds in four months, the WHY-annotation idea, and an empty user count.
- [Everyone says taste is the moat. That's the cope.](https://iamladi.dev/writing/taste-is-the-cope): The comforting line is that taste is the one thing AI can't take from you. But nobody can define it, teach it, or prove who has it, and stripped of the mystique, taste is just an evaluation function: what tells a feedback loop better from worse. We already built the machines that learn one. The edge was never having taste. It's running the loop.
- [I built the AI pivot from the inside. It didn't work.](https://iamladi.dev/writing/the-pivot-that-didnt-work): In late 2024 every crypto company suddenly had an agent strategy, and I was inside one of them, building ours with my own hands. The agents mostly died. This is the honest anatomy of a hype-wave pivot (why we did it, what we built, why it failed) and why the failure is the direct ancestor of everything I build now.
- [Stop typing. Don't stop understanding.](https://iamladi.dev/writing/stop-coding): The keyboard is the part that depreciates, not the thinking. Stop typing code, direct the loop in English, and move your understanding up the stack. The trap is confusing "stop typing" with "stop understanding," and that is where careers and codebases die.
- [UI isn't dying. It's splitting in two.](https://iamladi.dev/writing/ui-is-dying): The argument, steelmanned and stress-tested, with the diagrams to see it. The task interface is being demoted to something the agent generates and throws away. The trust interface is being promoted. Whoever confuses the two builds on the wrong half.
- [18 years, four bets, one pattern](https://iamladi.dev/writing/four-bets-one-pattern): Rails, React Native, Web3, agents. What catching a wave early actually costs, and why I keep doing it anyway.
- [The graveyard is the portfolio](https://iamladi.dev/writing/graveyard-is-the-portfolio): On shelving most of what you build, and why the dead projects taught me more than the shipped ones.
- [Building the plumbing, not the interface](https://iamladi.dev/writing/building-the-plumbing): Memory, sandboxes, reasoning, routing. Why the least glamorous layer is the one worth owning right now.
- [Small models, big reasoning](https://iamladi.dev/writing/small-models-big-reasoning): I spent a while collecting 16 research-backed reasoning strategies and building a router that picks one per task. Notes from teaching small local models to think.
- [I shipped a game, then walked away](https://iamladi.dev/writing/shipping-a-game): Five years, a small studio, a narrative game on Steam. What building Book of Aaru taught me that no SaaS ever did, and why I closed the studio anyway.

## Selected work (short list)
- **Save Bulk Gmail Attachments** (TypeScript · Chrome · Workers) — My first product built entirely by directing AI in English. 1,280 users, revenue Stripe-verified in public. [https://iamladi.dev/projects/save-bulk-gmail-attachments]
- **LoomBrain** (TypeScript · PARA · Graph DB) — Second brain for AI agents. The brain lives online so agents reach it while your laptop sleeps. My current flagship. [https://iamladi.dev/projects/loombrain]
- **sandcaster** (TypeScript · CLI) — Run AI agents in secure cloud sandboxes. The agent is a config file, not code to wrap. [https://iamladi.dev/projects/sandcaster]
- **Book of Aaru** (Unity · C#) — An action roguelite shipped on Steam with a small studio over five years. 86% positive. [https://iamladi.dev/projects/book-of-aaru]
- **react-native-hockeyapp** (Objective-C · Java) — Early React Native OSS, a HockeyApp bridge for iOS and Android. Still my most-starred repo. [https://iamladi.dev/projects/react-native-oss]
- **Access Protocol** (TypeScript · Go · Solana) — Four years, founding week to sabbatical: platform, design system, SDKs, and the AI-agents pivot from inside. [https://iamladi.dev/projects/access-protocol]

## The lab (experiments)
- **saashunting** [learning, 2026] — A machine that finds problems I should work on. It found two shipped products; it also taught me the real bottleneck is elsewhere.
- **selfunderstand** [shelved, 2026] — A psychometric interviewer you talk to. Paused on a hard voice-vs-transcript problem.
- **permitflag** [shelved, 2026] — Idea to paid Chrome product in four days. Nobody paid: the ICP could not operate the delivery mechanism. Best tuition I bought all year.
- **pentune** [shelved, 2026] — A writing assistant that learns your voice through corrections. Killed with a clear head: the loop demands too much from the human.
- **refocus** [shelved, 2026] — A focus coach that talks you back to work. Cool product; it never helped me, so I shelved it.
- **newadmin** [learning, 2026] — Rebuilding my brother's ISP back-office from interviews and database archaeology. Active this month.
- **czech agent commerce** [learning, 2026] — Teaching agents to buy things and mail physical letters: Medusa adapters, QR payments, Česká pošta API.
- **macOS menu-bar tools** [shipped, 2026] — Four paper cuts, four tiny apps: audio presets, icon hiding, meeting audio, ambient music. "Too small to build" is a dying category.
- **x-articles** [shipped, 2026] — Export X articles to Markdown for my AI. 500+ strangers turned out to want the same thing.
- **livingportrait** [shelved, 2026] — Harry Potter portraits of your family on iOS. Built in days; the App Store ceremony is where it paused.
- **agent-reasoning** [learning, 2025] — 16 research-backed reasoning strategies for small models, with an auto-router that picks the family per task.
- **eth-key-brute-force** [shelved, 2025] — GPU Ethereum address generation with Apple Metal shaders. Fast, pointless, deeply satisfying. — https://github.com/iamladi/eth-key-brute-force
- **mcp-i18n** [shipped, 2025] — An MCP server for AI-powered i18n translations. Small, done, still useful. — https://github.com/iamladi/mcp-i18n
- **sentinel** [learning, 2025] — A watcher over my shoulder, a personal automation experiment in Rust. Ongoing.
- **twitter-new-tab-opener** [shipped, 2025] — One-click open tweets in a new tab. The most scratch-an-itch thing I have ever built. — https://github.com/iamladi/twitter-new-tab-opener
- **react-native-actionsheet** [shipped, 2017] — An Android ActionSheet bridge. 25 stars, still downloaded years later.
- **ftnm** [shelved, 2022] — Notifications when NFT whales moved assets. Peak-cycle idea; the cycle moved first.
- **capture-pipeline** [learning, 2026] — The ingestion layer feeding LoomBrain. The most important thing I am working on.
- **safedomain** [shelved, 2025] — Domain-expiry insurance for owners who miss renewals. Never left the idea stage: the unit economics said no.
- **rack-restful_submit** [shipped, 2016] — RESTful multi-submit-button forms with no JavaScript. Very 2016, still elegant. — https://github.com/iamladi/rack-restful_submit

## Pages
- [Home](https://iamladi.dev/)
- [Projects](https://iamladi.dev/projects): all endeavors with post-mortems
- [CV](https://iamladi.dev/cv): one-page CV — Ladi is open to senior engineering / engineering-leadership roles and contract work
- [About](https://iamladi.dev/about)
- [The lab](https://iamladi.dev/lab)
- [Writing](https://iamladi.dev/writing)
- [Ask Sam](https://iamladi.dev/chat): an agent that speaks on Ladi's behalf, grounded in his notes
- [RSS](https://iamladi.dev/rss.xml)

## Contact
- Email: ladi@backtomeaning.com
- GitHub: https://github.com/iamladi
- X: https://x.com/martincik