How I Think About Asymmetric Bets
A personal framework for finding durable upside in technology, from steam and electricity to language models, AI agents, Bitcoin, and whatever comes next.
Field notes.
Practical notes from backend engineering and building products, plus occasional writing about business, family, and ambition.
Start here.
I Wanted a Break From Prompting, So I Planted Some Trees
What started as a quiet plan for planting trees became an interactive 2D and 3D masterplan for 8,615 square metres of family land.
What Reliable Context for AI-Assisted Teams Actually Requires
Giving an AI assistant more text is easy. Giving it current, permitted, traceable, and testable context is an engineering problem.
More notes.
24 notes
9 min read
The Small Model That Knows One Job
Tobi Lutke shared a tiny Shopify model beating a frontier model at one internal job. I wanted to know what that result really says, and whether the same idea is useful for PLAYGRND.
9 min read
Every Useful Agent Eventually Needs a Computer
I started with Grok Bot and ended up looking at the infrastructure underneath every agent that is allowed to do real work. Daytona is one of the more interesting bets, but this market will be won on trust, not demos.
12 min read
When an HTTP Request Starts Carrying Money
What x402, Coinbase, Cloudflare, and always-on agents such as Grok Bot reveal about the next payment layer of the internet - and where Bitcoin fits.
12 min read
Notes on Overnight Software Delivery at Tilt
Notes from my current Tilt workflow: why I run bounded overnight work locally on a MacBook, how the tools fit together, and where human review remains.
10 min read
I Gave PLAYGRND a 12-Thread AI Team
How I organized Codex into twelve bounded product workstreams without turning one shared repository into chaos.
10 min read
I Think in Compressed Models
Why abstraction helps me connect systems quickly, where compression errors become dangerous, and how I separate creative synthesis from cold validation.
11 min read
How Far Can One Hetzner Server Go?
A concrete capacity model for a one-server Hetzner setup: what breaks first, which metrics matter, and what I would build next on the path to 10 million users.
15 min read
When One Server Is Not Enough
A practical follow-up to the one-server setup: what I would do when a small app turns into a real business and the server starts running out of room.
11 min read
How I'd Run a One-Person Internet Company on One Server
A technical manual for running a one-person internet company on Hetzner with Cloudflare, Caddy, PostgreSQL, Redis, systemd, GitHub Actions CI/CD, R2 backups, mobile Codex operations, and resource limits.
8 min read
Why Memoato Is Becoming a Trustworthy Context Product
What started as a personal memory app is becoming a context layer for people and AI agents, without abandoning the personal product that exposed the real problem.
6 min read
Ambition Without Disappearing
The hard part of chasing bigger goals is not only the work. It is building a version of success that does not cost the people you are doing it for.
8 min read
Testing Land Business Models Before Building Too Much
A follow-up on the Dalmatian land project: rough business-model estimates, pilot logic, and an open call for practical collaboration ideas.
3 min read
Every AI Coding Workflow Needs a Runbook Layer
Why AI-assisted development gets safer when agents can read the operational runbooks before they make code or deployment decisions.
3 min read
Agent Memory Needs Provenance, Not Vibes
Why agent memory is only useful when every durable fact points back to a source, a date, and a reason it should still be trusted.
4 min read
How I Structure Context for Coding Agents
A practical structure for giving coding agents useful context without turning every task into a giant prompt dump.
6 min read
What Would You Build on a Piece of Land in Dalmatia?
A public note on thinking through a rural land project in Dalmatia: events, storage, family gatherings, and the search for a practical business model.
4 min read
Making PLAYGRND Aggregates Fast Without Lying to Myself With Cache
How I moved stable PLAYGRND season aggregates from raw query paths into persisted derived tables and Redis snapshots, while keeping raw fallback intact.
4 min read
Why I Use WhatsApp for PLAYGRND Login and Claims
For PLAYGRND, WhatsApp is not a gimmick. It is the channel users already trust, so login and profile claims should meet them there.
4 min read
What PLAYGRND Is Teaching Me About Building Faster With AI
PLAYGRND is my practical example of how a small team can move faster with AI without letting AI replace product and engineering judgment.
3 min read
Which Context Should Win?
Agent reliability is not only retrieval. The hard part is deciding which source is allowed to control the next action when context conflicts.
5 min read
Memory Is Not Enforcement
Why agent memory is useful, but not enough. The real work is source-backed context, permissions, stale-context detection, and eval traces.
5 min read
The AI Agent Problem Is Not the Model. It Is the Context.
Why the long-term advantage in AI-assisted software work is not picking the winning model, but building reliable context systems around any model or tool.
7 min read
I Ship to Production From My Phone With an Old Laptop and Wasp
How I set up mobile first development and direct production deploys using an old laptop, Tailscale, tmux, Telegram notifications, Cloudflare Tunnel, and Wasp.
4 min read
How I Built a Free AI-Powered Life Management System (and transformed my daily workflow)
How I use Obsidian + Gemini CLI to create a free, AI-powered life management system that outperforms expensive productivity tools and provides strategic insights.