AI Made It Too Easy to Start (Now I Can't Finish Anything)

00:00 — Opening

Hey, welcome to Slow Builds.
Lately I've been building a lot. Probably more than I ever have in my life.
Like — sitting down and starting a new project? I'm doing it all the time.
And that's actually new for me. I went through a phase where I was burnt out, or I couldn't really come up with ideas. I was genuinely worried for a while that my mind was kind of out of gas — out of creative fuel.
But I'm back. My mind's constantly trying to put pieces together, come up with new ideas.
It's not always in a clean way…
But things are moving.
And that's actually the problem.
Because nothing's really getting finished.
Nothing feels done.
And I think AI is a big part of why.

01:15 — What It Looks Like From the Outside

It's strange. Because if you looked from the outside, it would probably look like a lot is happening.
I have a trading app live — people can use it. I have this YouTube channel. I've been posting consistently, more than I planned. There's a book I've wanted to write for years that's probably 70% done.
So things are happening.
But internally? It feels scattered.
Like everything is in motion… but nothing is landing.

02:45 — The List (Let Me Actually Walk Through It)

So I started listing everything out.
There's the kid's trading app — that one's actually live. It's out there. They want to make some changes, but technically that one's done.
There's Steady — that one's personal. It's like a coaching app, SMS-based, interacts with AI. But it's very personal to the people who'd use it, so I don't want to put it out until I fully trust it. Privacy, auth — all of it has to be right. So that one's intentionally slow, and I'm okay with that.
There's a dividend tracking app — that one was going well, then AI took it down a different road. Feels like it's going nowhere right now.
There's GoChores — started as an app for my kid, they lost interest, but I converted it into kind of a template. A test bed I run everything through to make sure my agents and infrastructure are set up properly.
There's my kid's website — AI built the whole thing, we got the URL, connected to Railway. She's got a marketing company she's spinning up. That one's basically done, just waiting on her to approve the changes.
And then there's the book.
And this channel.
And then… I started getting deep into OpenClaw.
That's where things started to feel different.

04:30 — The OpenClaw Rabbit Hole

I set up a server at home. Got OpenClaw running — agents, schedules, automation, all of it.
And it's genuinely cool.
It can post to X, read responses, generate ideas, build things. It's connected to Linear so it tracks issues, bugs, testing. It deploys code. It does all kinds of things.
I have a Telegram channel where it gives me updates every morning.
I even set it up for this channel — I have an X account where it does nightly posts, responds, follows, all of that.
And I really started leaning into it. Like — what if I just let this run?
Let it come up with ideas. Let it build. Let it iterate.
And it does.
It just… never stops.

06:00 — The Treadmill Feeling

At first it feels like progress. Because something is always happening.
You check in — it made changes, it tried something, it improved something.
But after a while… it starts to feel like a treadmill.
It's moving. But not arriving anywhere.
Every morning I get my Telegram messages. I see what happened last night. But I'm not touching it. I don't actually know what's going on.
There's no definition of done in that system.
So it just keeps going.
And that's where that quicksand feeling comes from.

07:30 — The Bigger Realization

And I think it's bleeding into everything else.
Because AI makes it so easy to start something. To explore ideas. To keep iterating.
There's almost no cost to continuing.
So you don't have to decide anything. You don't have to say — this is enough.
Before, effort forced you to choose. You had to validate something before building it, because building cost time. Now? I don't even have to validate. I can just spin it up, see what it looks like, compare it to what exists.
I'm not doing the hard lifting anymore.
But the fast-building cuts both ways. Because now I can quickly throw something together — and then never finish it.

09:00 — Finishing Feels Different Now

And finishing isn't just finishing anymore.
It's saying: This is good enough for someone else to see.
This might break.
My name is attached to this.
That back-developer part of me, that engineer brain — it knows you can't just throw things out there. Especially if you're collecting money, collecting people's information. There have to be guardrails.
Which is actually part of why the OpenClaw infrastructure thing consumed me — I was trying to build those guardrails automatically into the system.
And I got so caught up in that… that there's no ever-done. There's just a never-done feeling.

10:30 — Everything Leads to Something Else

Here's a good example of how this works.
I built a little widget for my kid's website — so she could leave notes about what she wanted changed, or use AI to make quick edits herself. I could log in, see all the notes, which page, what she wanted, whether the AI edit worked or failed.
It was really cool.
And as soon as it worked, my brain immediately went: wait — maybe there's something here. Maybe this is a WordPress plugin. Maybe it's a website widget people would actually pay for.
That's the cycle.
You start touching something, AI makes it easy to jump onto something else. Another hole. Another fork in the road. And you go, well, maybe this is a viable product.
I'm not trying to build the next Facebook. I just want small apps that run. Things I want to use. Things other people might want and would pay a few dollars a month for. Just enough to keep the lights on — cover the hosting, the tokens.
But every idea spawns three more.

12:00 — What I'm Actually Starting to See

I don't think the problem is that I have too many things.
The problem is I'm not deciding what gets finished.
Everything just stays in progress. And AI makes that very easy to do.
You can always tweak one more thing. Add one more feature. Improve one more flow.
There's no natural stopping point anymore.
The projects that did finish? They had someone else's expectation attached to them. My kid's trading app — there was a real reason it needed to ship, a real timeline. The Glove Box GPS app — I knew what it needed to do. Done is clearer when there's someone on the other end.

13:30 — Where I'm At

So I think what I need to do is simpler than I've been making it.
Not build more.
Not organize more.
Just pick one thing… and actually finish it. Even if it's rough. Even if it's alpha. Let it exist.
Glove Box GPS is basically there. It's on my phone, my wife's, going on my kid's. We're testing it. And I'm going to let that be good enough for now.
That's the shift I'm trying to make. From done to done enough.

14:30 — Closing

Right now everything is moving. But nothing is landing.
And I think the real shift isn't about building anymore.
It's about being willing to look at something and say: this is done enough. And letting it go.
If you're out there building with AI — and you don't even have to be a developer, because AI makes it so easy for anyone — if you're creative and you're using AI to start things… books, channels, apps, agencies, whatever it is…
Are you finding that starting is easy, but there's never an ending?
Let me know. I'd genuinely like to know.
Thanks for watching.