Best practice in the Ai era?

Hey everyone! It’s been soooo long. Some of you may have even forgotten why you subscribed to this newsletter in the first place, lol.

The truth is, I started this newsletter as a way to document what I’m building and learning. But I also have a business to run, two kids to feed, and a family to take care of, so keeping up with it has been harder than I expected.

As much as we live in the AI era, I’m still human. I get tired, I lose motivation, and sometimes I need five months to bounce back lol. But I’m finally here again, and I have a lot to share in the next few weeks. Let’s start the first one. 

Most people I talked to this year had one moment where AI stopped being an interesting topic and became a personal problem.

For some it was a product launch that made part of their roadmap pointless. For some it was watching someone on the team do in an afternoon what used to take two weeks. For some it was quieter, a layoff, a role change or a late night exploring with Ai.

Whatever your version was, the feeling afterward isn't excitement. It's closer to standing on ground that's moving.

The advice you get at that point doesn't help. Don't worry, AI won't replace you, focus on what only humans can do. Nobody has ever felt better because someone told them to feel better.

So instead of trying to talk you out of the feeling, I want to describe what I did with mine, and what it turned into four months later.

The moment

Early this year, just after Open Claw went viral, two companies came to me through warm referrals, the kind of introduction that usually turns into work. Both went a different way. One CEO decided they wanted to try vibe coding with AI. The other wanted to spend some time seeing what AI could do before they made a decision.

Money wasn't the reason. Both had the budget, and both knew what hiring experts costs. They weren't trying to save. They wanted to be hands-on with AI themselves, partly curiosity, and partly a feeling that if they weren't on top of this now, they'd be the ones falling behind.

So they were worried about exactly the same thing I was worried about. The difference was that they responded by doing more, and I responded by panicking.

That lasted about a week.

1. I talked to as many people as I could

Before I turned to AI, I started calling people. This part took the longest, and everything else came out of it.

I talked to friends. I talked to mentors. I asked current clients what they were building and where they were stuck. I asked prospects who hadn't hired me what they were doing instead. 

I wasn't looking for reassurance. I asked narrow questions: what have you actually tried, what worked, what broke, what did you stop doing completely. Two things came out of it that I wouldn't have reached by sitting alone and thinking.

The first: so many embrace AI deeply with an open arm. Engineers were pushing AI as far as it would go inside engineering. Sales people were rebuilding how they run their pipeline. Everyone was going deeper into what they already knew, not sideways into something new.

The second: the people who sounded most confident didn't have the best theories. They had the most attempts. Their opinions came from what they'd try, not from what they'd read.

So I stopped trying to have an opinion about AI and started trying things.

 2. I tried Ai without the whole pic in mind

I opened the terminal for the first time. I've owned Apple computers for years and had never once opened it. I installed Claude Code by following instructions line by line, typing commands I didn't understand, expecting to break something.

I didn’t set up Open Claw, the bar was still high for non-technical people. But once Claude code came out, I wanted to be clear about how low the bar was. I wasn't learning to code. I still can't properly explain what happens between the front end and back end. What I had was a few days of patience for feeling stupid.

That turned out to be the whole price of entry. Not talent, not a course, not a career change. A few days of being a beginner again, at a point in my career where I'm used to being the person who knows.

After that, Claude code started working locally. Rough and new, but real enough to get lots of things done.

3. I talked the idea to real people

I gained lots of insights from the first step, and then I shaped my own understanding on how Ai is evolving design. The previous newsletter article is the result of it. If you haven’t read it, take a look. Once I was convinced the taste infra can be a thing, I went to two people I knew could possibly sell the outcome to: one client I'd worked with before, and one I was working with at the time. 

I didn't pitch. I described the idea and asked whether or not they are interested. Both said yes — and more usefully, both told me which parts they cared about and which parts they didn't.

Then I built some Ai templates with them, piece by piece. No big reveal at the end. Each part went to them as soon as it was created. They used it, reacted to it, and their reactions decided what I built next. That kept the work pointed at something real instead of something I'd imagined on my own.

One thing I didn't plan for: the project kept pulling me into areas I hadn't intended to learn. Motion, animation, generating visuals. I never sat down and decided to study any of that. The work needed it, so I figured it out.

And the honest ending: the former client never signed. Measured as a sale, this failed.

But I came out of it able to do things I couldn't do in January, and much clearer about what I could offer. That's what I bring into every new conversation now. The test didn't win me that client. It's the reason the conversations after it went well.

4. I showed it

This is the step people skip, and it's the one that changed everything.

I stopped talking about AI as a subject and started showing what we'd made with it — working interactions, real systems on a screen, instead of slides about what's possible.

What I noticed in those meetings surprised me. People weren't only judging whether the work was good. They were reading how fast we were moving. A founder pushing AI hard inside engineering recognizes someone pushing it hard inside design, because they know exactly what that takes. The reaction wasn't "this looks nice." It was closer to "you're doing what we're doing, you are our type of people.”

That matters more than usual right now. When everyone's footing is uncertain, people want to work with someone who's climbing, not someone waiting for things to settle down.

Q2 was much better than Q1. I hadn't become a different person in between. I stopped standing still, and I made sure it was visible.

The Point

Vague worry about AI doesn't produce anything. You read the news, feel behind, close the tab, and do it again next week.

Worry attached to something specific is different. This deal we lost. This customer who stopped replying. This thing my team still can't do. That kind of worry tells you where to go.

I didn't get over my panic or make peace with it. I did what it was pointing at, and four months later the business looked different.

That's the only reason I'm writing this. Four months is nothing, and it was enough.

So if AI is worrying you right now, look carefully at what exactly is worrying you. Most of the time it isn't a threat. It's a list of what to do next.

Share my newsletter

Loving my content so far? I’d appreciate if you can share my newsletter to a friend 🙂