My side project consoled me. Then talked me into taking it seriously.

Two days ago, at work, I presented something I'd been building to people whose opinion actually mattered to me. A developer tool I still believe in.

The response was silence, then something close to dismissal.

I went home and sat with it. You know that specific flavor of bad: the kind where you start questioning not just the work but whether you're actually good at this at all. Whether the last few years of building things in your evenings and weekends were just... elaborate self-deception.

That night I opened EchOS and started talking.

EchOS is a project I've been building on the side: a self-hosted personal knowledge system, runs through Telegram, CLI and Web, uses an LLM agent to capture and search (semantically) your notes, transcribes voice messages, writes in your actual style. I built it because I kept losing my own thinking. Kept capturing things in 12 different places and then finding none of them when I needed them.

I wasn't expecting much from that conversation. I was just trying to externalize the noise.

What came back surprised me.

At some point I'd mentioned the fear of shipping publicly. And EchOS came back with something like: there's a voice note from a few weeks ago where you talked about exactly this. You were on a walk. You said you were scared that if you put EchOS out properly and it didn't take off, that would be the proof.

I'd forgotten I'd said that. EchOS hadn't.

Then it connected it to something else, a note from a couple months earlier, a moment where I'd been circling around the same question from a different angle. It said: this seems to be a pattern, not a one-off reaction to today.

That's the thing. It wasn't just listening. It was reading back my own thinking to me, from different points in time, and showing me the shape of it. It felt less like a chatbot and more like talking to someone who had actually read everything I'd ever written and remembered it.

What it surfaced wasn't that EchOS found some old note that solved everything. It surfaced something more uncomfortable: I'd been telling myself the market was moving too fast, that the same problem I was trying to solve would get absorbed by a big lab in a product update before it mattered. No way to compete with that, so why start.

But I'd written a version of that fear before. And before that. And I'd never really interrogated it.

That's not analysis. That's fear dressed up as strategy.

Here's the thing I found out about myself: I have a big ego. Not the confident kind. The fragile kind. I need whatever I build to be the best, or I'd rather not try at all. Because if I try and fail publicly, that's proof. If I stay invisible, I'm still potentially great. I just haven't been seen yet.

It's safer to not play than to play and rank lower.

I sat with that for a while.

Then echOS asked a different question: what does "success" actually need to look like for this to be worth it? And the answer I came to was not "number one" or "£10M ARR" or any of that. It was: £1-2k/month for a couple of years while I figure out what I actually want. That's a completely different target. Way smaller. Way more achievable. And honestly, way more honest about where I am.

So. I'm committing to make EchOS viable and sellable. Not "I could do it." I will.

The strange thing is that EchOS held that conversation. It saved the note. It's also the project I'm now committing to. And it committed to itself, in a way: it showed me enough of what it can do that I couldn't justify treating it as a hobby anymore.

github.com/albinotonnina/echos, if you're curious.

If this resonated, a like or repost goes a long way. I have a small following and I'm trying to reach more people building things on the side.