Six things I haven't worked out yet.

Most sites hide what the writer is unsure about. This page is the opposite. If any of these are also yours, write — half of this site exists in the hope that someone with a better answer than mine reads it.

Scaling myself without scaling the meetings.

Right now my life doesn't have heavy back-and-forth communication. WhatsApp manages family well. Email runs on a two-to-three-day baseline; Slack rarely opens. As danadone grows that won't hold. The question isn't respond faster — that's the wrong target. The question is how to scale Sailesh without turning my calendar into other people's queue. Async by default, agents on the easy half, human time reserved for the calls that actually need it. I haven't worked out the exact shape yet.

Telling strategic walk-away from ADHD-rotation in real time.

I've quit two kinds of things in my life. The strategic kind — UPSC, after five attempts, when I knew I was a year too late — was the right call, and the data was clear. The ADHD kind — projects abandoned not because I lost belief but because novelty-seeking pulled me sideways — was the wrong call, dressed up by my brain as a strategic one. From the inside the two feel identical. mineOS is partly an attempt to externalise the difference: track interest curves, project age, last-shipped-thing, and surface the pattern so the call doesn't depend on whatever mood I'm in that morning. Whether the system is good enough to tell them apart in the hard cases is still being tested.

Closed core or open core for danadone.

My current view is closed. The tech is not the hard part; the moat is in operations, network effects, and the flywheel — and open-sourcing the engine doesn't help on any of those. So open-sourcing serves neither us nor anyone else, as far as I can tell. But the Posthog / Sentry / Hashicorp playbook is real, and there are versions of bottom-up adoption that I haven't fully tested against this case. So I hold the position lightly. Tell me if there's another point of view.

Over-analysis is procrastination wearing a suit.

So is over-learning without action. I know both about myself, and naming them doesn't fix them. The system-fix is to break work into parts small enough that I don't avoid them, but not so small that I lose interest — a Goldilocks size that depends on the task and the day. Walks for thinking, pen and paper for hard calls, an LLM as an objective sounding board. The tools are clear; the discipline of using them when interest is low is the part I'm still building.

The 70-80% boundary target, vs the long-hours present.

The current pattern is long hours with deliberate family-protected windows during the day. The target is seventy to eighty per cent of weeks with clean boundaries. The gap between them is real and I'm not pretending it isn't. The deathbed test is the calibration: looking back from there, you should not feel you missed the most important things. Work is one part of it, not the part. I'd rather name the gap honestly and close it deliberately than perform a balance I haven't earned.

Building the 1000-acre self-sustaining parcel.

Ten-year plan. Tech-native kids on one end of the barbell, ability to live without civilisation on the other end. Food, clothing, basic necessities, even comforts and a few luxuries — produced from the land. Wherever a single family can't put together that much land, a community version should still work. The plan is the easy part. The execution — money, location, the right partners, the politics of land in India — is the hard part, and I don't have it solved yet. I'm noting it here because the wrong move is to keep it private and forget about it.