Mastermind · January 17, 2026
MentorMates is the center of gravity
The question Xisen kept circling back to was the one I didn't want to answer in one word: what's the center of gravity?
I opened the first real mastermind of the year with a list. A part-time offer from one AI lab I'd already done two rounds of interviews with. A potential fly-out to Austin to work with a founder I'd just spent six hours on a call with on personal knowledge management. A renewed pull back into research because I needed evaluation data for a conversation product I'd been prototyping. And underneath all of it, MentorMates, still shipping, still the thing that was actually producing usage.
Writing the list down is what gave Xisen the opening. His read was blunt: every one of those was individually plausible, and collectively they added up to a year with no throughline. So what is the center of gravity?
I said MentorMates. He didn't let me off the hook. Then why does everything else get airtime?
The honest answer is that optionality feels cheap when you keep it open and expensive when you close it. Keeping the lab interview warm, keeping the Austin thread warm, keeping the research collaborator warm, none of those things cost me anything in the moment. They only cost attention, which I wasn't tracking. The weekly call forced me to track it.
So I committed on the call, out loud: Q1 is MentorMates. Not in the Twitter thread sense, in the literal sense. The visiting collaborator who flies out in ten days is going to work on MentorMates. The plan I've been carrying in my head gets written down this week. Every other thread stays on the back burner, meaning they don't die, but they don't run.
What pushed me to name it that way was Xisen mirroring the same problem from his side. He's been pulled between four plausible research directions, each credible, none yet committed. He landed on a version of the same discipline: one paper, one thesis, write it down, stop reopening the question every week. The reason the mastermind works is that neither of us can get away with treating optionality as free when the other person is taking notes on what we said last week.
The part that's harder than saying it is the secondary effect. Picking MentorMates as the main line means the partnership conversations, the ambassador work, and the event cadence all have to plug back into a product surface that holds up. Which means the teams I'm working with, a CTO in a different city, a small frontend bench, a visiting collaborator, have to be on the same page about what's load-bearing and what's decoration. Every feature that doesn't serve event reliability or submission flow is adjacent, not core.
The pattern I keep running into, and I say this honestly because I've been guilty of it: early founders collect threads like trophies. Every conversation that went well becomes evidence that the problem is bigger than one thing. The actual exercise is the opposite. Every conversation that goes well is a test of whether the main thread still holds.
What makes this call different from last year's planning conversations is that I don't have the escape hatch anymore. Last year I could tell myself the category was still forming. By this January, the category is formed, the competitors are printing revenue, and the question isn't whether hackathons are a real surface, it's whether we execute on it. MentorMates as the center of gravity is a statement about that. If I'm still carrying five plausible alternatives by April, that's me hedging, not me being strategic.
The check I'm putting on myself: the plan gets written this week, and the visiting collaborator arrives to a list short enough to execute, not a vision deck long enough to impress. If by the next mastermind the list has doubled, Xisen gets to call it out.