Aym Abdalla · building in public
I run a company with AI agents, and publish what the evidence actually says.
Fourteen AI seats, a product heading for its first design partners, and a scoreboard with the losses left on it. Written for founders building the same way: the method, the infrastructure, and what it costs when the answer comes back no.
The premise
Honesty first. Everything else is downstream of that.
Most build-in-public writing shows the win and buries the rest. The Nest publishes the score as it stands, including a products-shipped count that currently reads zero, because it is measured rather than assumed. If you are a founder running lean, the highlight reel is the least useful part. The method is the useful part, and so is what it costs when the method returns nothing.
The receipts
Real numbers, not screenshots of winners.
The clearest proof of the method is the project that returned no large edge. A full research system ran the numbers, found nothing worth risking money on, and got published in full instead of quietly shelved.
What's here
Huntloop
An autonomous outbound CRM for founder-led B2B teams with no SDR. Specification approved, release hardening under way, first design partners not yet onboarded.
The Org
Fourteen AI seats running the departments a one-person company still needs: engineering, security, product, revenue, legal. Every project here was built through it.
The Null Result
A trading research system, backtested and shadow-traded, that found no edge worth risking money on, and published the proof rather than the highlight reel.
Three iOS games
RiftRally, TickerTown and Verdant Idle: three native iOS games built in a single day with AI assistance. One has continuous integration proving it still builds. None has ever had a player.
INeedToKnowThis
A study app that turns a saved reel into lessons you actually retain, and deletes any lesson it cannot verify against the source.
The newsletter
Written for founders building with AI.
What the org actually decided and why, what the evidence said when it disagreed, and what a decision cost when it was wrong. No schedule, no filler. It goes out when there is something true to report.