Reflections after six weeks of building on my own

I left my job 6 weeks ago to build my own company, and I have not posted any articles on Substack since then.

My excuse for not writing is that I am busy building, researching, and selling. Deep down, I know it is not the root cause.

To be precise, I did not become entirely quiet. I am posting even more on LinkedIn, my main distribution channel, but I lost the temptation to write long articles on Substack.

I did well in my previous roles as an engineer. Delivering good work, being recognized, getting promoted and rewarded. I was confident about what I was doing, and there was a positive feedback loop that strengthened my confidence.

So I was comfortable writing about the stuff that I was doing and the stuff that I’m interested in.

I quit to find my new limits, which means doing things that I’m not (yet) good at. And I can really feel the difficulty in the last few weeks.

6 weeks is not a long period (I have myself 6 months to try things out), and I know the journey can be more painful. But it’s long enough to affect my self-assurance when there have been fewer positive signals than I was expecting.

So every time I sat down and intended to write something, I always told myself to stop, do the work, and prove with some result first.

Looking back, that was not the real reason I started this journey. I wanted to learn, and failing was fine.

So this time, I allow myself not to shut up and write again.

I started with an idea that I thought was solid, which is to use AI to automate a traditional service-based business, and I still stick with it.

There is market demand. AI is good enough to automate parts of the workflow, but not the whole pipeline, and it’s still a hard-enough problem that competitors cannot simply throw AI at it.

After some early validation, I built it to a shape where it’s good enough to test with real customers, and I have been spending the last few weeks contacting potential buyers. That’s when the roadblock hits.

I’m doing things I learned from the people who went before me: doing things that don’t scale, using AI to do research but still reaching out to people manually and tracking everything. But the reply rate is roughly under 5% without a strong buying signal from the replies I have so far.

All are expected, but not a good sign overall.

Another problem that I found is how my strength as an engineer negatively affects me in my validation stage founder mode.

I have always been a generalist. Not all engineers should be generalists, but being one has helped me progress in my career.

I like learning about many things, trying new things, and working on things in parallel. That helped me as an engineer. Once I could do my main work well, I could contribute outside my core responsibility and bring more value to the team.

Working alone at the validation stage is different. Being a generalist indeed brings a lot of power to startup founders, but I need more focus at this stage.

When things don’t work out, I tend to wander around, think about other ideas, and burn tokens to build prototypes just to “test things out”. Most of them are useless and a waste of time.

It’s true that people usually need to try out a bunch of things before finding the right one, but I need to try things more serially: focus on one thing long enough to validate or invalidate it, then move on deliberately.

Nevertheless, not everything is negative. I have more time to think, read, and do things that I have not tried before. I also meet and talk to more people, which was hard when I was still having a full-time job.

Let’s see what comes next.