last week I sold DistributedCompute .com for 32k. the domain was sourced and priced by my agent that I’ve been building since March, so figured I’d share a bit about it for anyone interested. it currently scans over 2M domains daily and makes a shortlist of around 200 based on an algorithm I built and supported by the data I've been collecting for the last 2 years. every day I review each of its picks and give it feedback - manually adding domains that it missed, with my reasoning for picking them. at the end of each week it reviews the verdicts from that week, including which domains I ended up bidding on / buying and how much I bid / paid, then compares that against the current criteria and suggests improvements. it isn’t hard to get an agent to find objectively good domains, but I’d drain my budget in a day if that was the only criteria. so in the same way that I learned to do this myself, I’ve been training it to find the “sleepers” that can be picked up for fairly cheap but still have large potential upside. this has taken a lot more time than I expected, but because of the constant feedback-learning loop between the agent and I, it naturally gets better as time goes on. it’s also getting pretty good at finding prospective buyers and valuing domains based on demand rather than surface-level stats. while I typically have buyers in mind when I purchase a domain, the agent was able to find multiple companies I didn’t know about that have raised money in the past 12 months and are on a worse version of a domain I own. I’ve wanted to build a system like this since I started investing in domains but couldn’t get past an MVP. it’s finally getting to the point now where I trust it almost as much as my own intuition. the goal is to have ~90% of the business automated by the end of the year. thankful for the sales along the way that have gotten me here. time to go parabolic.