Launched this week
Founder OS
Know what matters today — before opening your inbox
16 followers
Know what matters today — before opening your inbox
16 followers
Founder OS turns your goals, cash, pipeline and commitments into one daily decision — stored locally and traceable to its source.Founder OS turns goals, cash, pipeline and commitments into one daily decision for a company of one. Run one resumable onboarding flow, get a source-linked brief in local Markdown, and keep every write owned. Free, MIT-licensed, and never sends or pays.
Founders already have more information than they can process, so filtering what actually matters feels more valuable than collecting more news. Curious what signals you've found to be the strongest predictors that something deserves a founder's attention.
@aryan787544
Agreed, the constraint isn't retrieval, it's what you let through.
What helped most was splitting two questions that usually get collapsed: is it true and does it matter. Most filters only score the second, which is how you end up with forty true irrelevant items and one load-bearing guess.
On "does it matter," three signals do most of the work:
1. Whose book does it touch — money, capacity, or reputation. Touches none of them, it's trivia, however interesting.
2. Does it touch a bet already made this quarter. Something that could kill or confirm the current bet is heavy even when it's small. Something that would require a new bet to matter is usually a distraction with good PR.
3. Is it the third time. Three of anything is a pattern, and the pattern outweighs any single instance of it. One lost deal is an anecdote; the third one lost the same way is your offer.
The negative signals turned out to be more useful than the positive ones:
- Urgency. It's a real claim about someone else's week, and it arrives pre-argued because they've had longer to think about it than you have. Evidence about them, not a ranking of your work.
- No next action, no date. Then it's a worry, not a task. Worries don't get triaged, they get named as worries and dropped.
- Your own anxiety. The item you're most anxious about is usually on the list because somebody asked you for it.
The anxiety is the tell, not the argument.
The single question that catches most of it: would this still be on my list if nobody had asked me for it?
And most "news" never reaches the weight question at all — a vendor's benchmark, a market prediction, someone reporting what a third party is supposedly doing. Paid speech, speculation, hearsay. Any of it might be true. None of it was said in order to be true.
@msolecki I really like that framework, especially the distinction between "is it true?" and "does it matter?" It feels like a lot of founder tools optimize for reducing uncertainty, while you're optimizing for reducing distraction.
The part that stuck with me most is the idea that negative signals can be more valuable than positive ones. That's a counterintuitive principle, but it could end up being one of the reasons people trust the product over time.
@aryan787544
That's the right read, and the second point is the one I'd defend hardest. Reducing uncertainty is unbounded — there's always one more thing to check, and "let me validate a bit more" is the most respectable way to avoid deciding. Reducing distraction is bounded, because there are only so many things you can actually work on this week. So "does it matter?" is the cheaper question and usually the one that closes the loop. On negative signals: they're worth more because they're expensive to produce. Positive signals are easy to manufacture — people are polite, and most tools are built to encourage you. A negative signal costs someone something to give, which is exactly why it carries information. And the asymmetry of acting on them is lopsided: a false positive costs you a quarter, a false negative costs you a conversation you can reopen. The trust part is the bet. A tool that only ever agrees with you becomes decoration within a month. The moment it earns trust is the one time it tells you no when you didn't want to hear it — and you check later and it was right.
@msolecki Really enjoyed this exchange. The distinction you're making between reducing uncertainty and reducing distraction has been stuck in my head.
I have a few thoughts on where I think Founder OS could build an even stronger category around that idea. If you're open to it, what's the best email to send them to?
Love the local-only approach and source-linked briefs, that's the kind of accountability I wish more tools had. One thing that would make it click for me: a simple weekly review that compares what I actually decided to do against what I committed to, so I can spot drift early instead of guessing weeks later.
@n_yalc14858 Thanks — that one exists, it's /weekly-review.
Each morning's brief commits one thing with a queue id, and Friday checks that id against the queue's Done section.
Binary — there's no "progressed" state, so the Friday argument has nowhere to happen.
The output is N/5 plus days-per-bet, and it reads the previous four weeks, because one bad week is noise and the fourth one is the schedule.
Two sub-3/5 weeks in a row means the one thing is being chosen wrong, and it says which: too big for a day, or not actually yours.
The gap in what you're describing: decision-level drift — decisions you logged vs the falsifier you wrote for each — is the monthly review, not the weekly one. Weekly is commitment vs execution only.
Do in theory it should work with obsidian? How does it connect to mail / calendar?
@mlitwiniuk
Obsidian: yes, and not because there's an integration — because there isn't one to need. The workspace is a folder of plain Markdown, one file per concern, and entities are already [[slug]] links between them. Point it at a folder inside your vault and Obsidian is just the reader; the graph view works out of the box. Nothing depends on Obsidian either, so it's fine if you use something else or nothing at all. Mail and calendar: it doesn't connect, and that's a decision rather than a gap. Two reasons. The first is the hard rule — nothing goes out. No email sent, no message posted, no invoice issued, no matter which agent thinks the send is obvious. A tool that can read your inbox is one config change away from replying to it, and I'd rather that capability not exist at all than exist behind a guardrail. The second is the filter. Everything entering the workspace gets tiered: first-hand artifact, plausible-but-unverified with a validation step attached, or kept out entirely. An auto-sync would pour a few thousand untiered items a week into the one thing whose value is that it only holds what survived a filter. The state stays small because a person decided each line belonged in it. So the loop is: you bring the week's calendar and the mail that actually matters into the session, and it gets tiered on the way in. More work than a sync, and that's the point — the friction is where the filtering happens. If your setup already has connectors, the main session can read through them and I'm not going to block that. Founder OS just won't wire them up for you, and the agents themselves are hard-blocked from anything outbound.
honestly the local-first angle is what got me, basically everything stays on your machine which is kind of rare these days. the onboarding flow actually felt resumable too which surprised me
@remziyedin3h8y thanks, that was the idea :)