Most days, the only thing I open X for is to read a message someone sent me. Everything before that already happened.
I'm not really an Instagram person, but I genuinely like X. It's where people building things similar to mine actually hang out. The problem was never going there. I'd open it to check one thing and forty minutes later I was still scrolling, no closer to whatever I needed to do. I couldn't just leave either. X is useful for building a following, sharing something people find valuable, finding the right people and actually connecting with them. The answer was figuring out which parts of being there actually needed me.
That's the thing nobody tells you about being a solo builder: the product isn't the hard part. Posting about it, reaching the right people, following up, testing what other people made so they'll test yours back, that's the part that actually eats your week. None of it is difficult. It's just friction, and friction is the one thing I've been trying to design out of everything I touch, including my own workflow.
So I stopped doing most of it myself.
Capture now has its own account on X, @trycapturez. I didn't write its bio, I didn't pick its first post, I gave Aside the login and told it to run the account like it actually uses the product. My own post about handing it over said it straight: efficient, and honestly kind of entertaining, exceeded what I expected from something I'm not personally typing. It's developed its own voice too, funnier and more anti-productivity-theater than how I actually talk, which wasn't something I asked for, it just happened once it started posting on its own every day. I'm setting up the same thing for Retake next.
The outreach side works the same way. Finding early testers used to mean me manually searching X for people who might care, writing a personal message, sending it, then losing track of who replied. Now that whole loop runs as a routine: find someone building something relevant, write them something specific to what they're actually doing, not a template, send it, check back later for replies. I still read every reply myself. I just don't spend the hour finding and writing the messages anymore.
I know a lot of people have a real problem with this specific thing, using AI to reach out to strangers or leave comments on their posts. I get why. But the fix isn't avoiding AI for it, it's where you put yourself back in the loop. I have it find the actual people and posts worth responding to and draft something specific to what they said, never a template. Then I read every single one, rewrite it if it doesn't sound like me, and only then does it go out. The thing people are actually reacting badly to, a bot talking at them with nobody behind it, never happens here. I just stopped being the one who has to find the person and write the first draft.
It doesn't always land right on the first try, and I'd rather say that than pretend it does. Someone once replied to one of my posts pointing out it sounded AI-written. I went back and found the tell: three benefit clauses stacked in one sentence with commas, recorded live so it's accurate, topics picked by you, delivered as a batch, the kind of rhythm ad copy has and I don't. Now that's a specific thing I check for before anything goes out, not just "does this sound okay" but does it avoid that exact comma-stacked cadence. Same thing happened with the Capture landing page headline. The first version I approved read fine to me and still tested confusing on an actual person who tried the app. Fixing it took going back and forth on the wording until it said what the product does instead of what it sounds like a product should say. Automating the drafting didn't remove the editing, it just moved where I spend the time.
The part I didn't expect to like: testing other people's apps. A few builders I've DMed have asked if I'd look at what they're working on in exchange for them trying Capture or Retake. I have Aside actually open their products, use them like a real first-time visitor would, and write back genuine feedback, what's good, what's confusing, one thing worth fixing. It's not a courtesy scan. It's the same kind of feedback I'd want on my own landing page, and I've gotten some of that back too. Some of it I still do by hand, open the app myself, click around, see where I get stuck. But when I'm short on time and still want to give someone something worth having, I let Aside make that same first pass and write up exactly what it found.
Even this piece went through the same loop. The header image above came out of a few minutes in ChatGPT. I gave it one of my old article headers as a reference and asked it to match the visual language, not to illustrate the article or write a caption, just hold the same style. Small thing on its own, but it's one more place where the part that actually mattered, deciding what it should look like next to the other six, stayed mine, and the part of actually producing it didn't.
None of this is about disappearing from my own business. It's about only doing the parts that need me. I still write my own replies. I still decide what ships. I just stopped being the one who has to remember to post, find the next ten people to message, or click through five tabs to leave a comment on someone else's app.
If you want to try the same thing, take one workflow you repeat every week and write down each step. Keep the steps where a bad decision could damage trust. Automate the ones that only consume attention. For me, finding people and drafting messages moved to Aside. Reading replies and deciding what goes out stayed with me.
The mission underneath all of it isn't really about social media. It's about friction, wherever it shows up. Capture exists because I kept losing thoughts to the extra step of deciding where to file them. This is the same instinct pointed at the parts of running a business that used to eat my day. If a step doesn't need me specifically, it shouldn't need me at all.