I finally wrote down what I actually want:

To be the person whose “this works” is worth more than the vendor’s demo.

It took me an embarrassingly long time to get to one sentence. This post is how I got there, partly so I remember, partly because I think other technical founders get lost in the same spot.

The Jobs and Woz problem

For months I felt stuck between two people.

I admire Steve Jobs’ philosophy: taste, focus, saying no, refusing to ship what is not right. But temperamentally I am Steve Wozniak. I love the technology itself. I would rather read your code than have lunch with you. I do not work rooms. I can talk for hours about a thing I built, but “networking” as an activity drains me in minutes.

So the aspiration and the temperament looked incompatible, and that mismatch read as “I am doing founder wrong.”

Here is what unstuck me: Woz did not lack Jobs’ philosophy. He lacked Jobs’ appetite for leverage over people. Those are different things. Jobs’ actual function inside Apple was editing. Holding a standard, killing what did not meet it. That function needs zero stage presence. The keynotes were distribution for the philosophy, not the philosophy itself.

And I already have a distribution channel that fits my temperament: I write. Two blogs, public repos, tools people can use in ten seconds. Carmack did it through .plan files. Plenty of deeply technical people have Jobs-scale influence in their niche, built entirely through artifacts and text. That lane exists. I was already standing in it.

The pattern I could not see from inside

The sentence did not come from introspection. It came from looking at my own shipping history like an outsider and asking what one thing everything points at.

  • GattyWorks, the studio I founded, sells software audits: independent verification for buyers who cannot see inside what they paid for.
  • The generic medicine finder exists because a family member overpaid at a pharmacy and the pricing system was opaque enough that nobody noticed.
  • The hidden cost of free hiring platforms is an exposé of an opaque system.
  • At CIDECODE I mentored police-adjacent teams dissecting malicious APKs. Making hostile software legible.
  • The studio’s DPDPA compliance skill is public, so the checks themselves can be checked.
  • Even the Cloudflare account audit we published: 61 checks against our own infrastructure, findings in writing, before anyone asked.

One sentence covers all of it. I make opaque systems legible so people do not get cheated by them. Nobody assigned me any of those projects. That is just where my hands go when nobody is steering.

My site’s tagline has said “your Human in Tech” for years. I wrote it before I understood it. It means the human who checks the machines.

Why this is the right decade for it

AI is about to generate more software than anyone can vouch for. I run agent swarms daily; I write the spec, hand the diff to the agents, and decide what ships based on evals, not vibes. The generation side is getting cheap fast.

Which means the scarce thing flips. When anyone can produce software, the bottleneck becomes: who can look at it and say “this is real, this holds, this will not fall over,” and be believed? Verification becomes the product. Trust becomes the margin.

That is a decades-long aspiration with a daily practice I already love: reading systems, building evals, catching a model when it is confidently wrong. The Woz part of me gets to do the fun work every day. The Jobs part is the standard applied to it.

The introvert part flips

Here is the reframe that took the sting out of feeling lost.

An auditor should be the person who would rather read your code than have lunch with you. Independence is the credibility. If I were great at schmoozing vendors, my “this works” would be worth less.

I had also, without noticing, already designed my company around my temperament. Every GattyWorks brief gets a written reply. The lowest-friction offer on the site is a 20 minute call with no slides. Written findings, fixed prices, no discovery theatre. I thought I was avoiding people. I was actually productizing the interaction formats I can sustain.

The honest tradeoffs

The 24 hour websites and 48 hour MVPs the studio ships are the cash engine, not the mission. I am keeping them, but I now have a test for what deserves my week: does this make my “this works” worth more? Engine work that never compounds the verification franchise is how I ended up feeling lost in the first place.

The Woz failure mode is real too: pure building joy with no editor produces a pile of cool artifacts and no direction. My guardrail is that I encoded the editor as process instead of personality. Evals over vibes. A content gate where a denial is total. A public guarantee with a refund attached. The standard does not depend on my mood.

And some human contact is irreducible. Clients are people. I am not pretending otherwise, just budgeting for it honestly.

What I am doing with the sentence

Publishing it here, pinned, so I collide with it every time I open my own site. I process by publishing; the mission goes through the same pipeline as everything else.

If you are a technical founder feeling the same mismatch, the exercise that worked: ignore what you think you want, list what you actually built when nobody assigned it, and find the one sentence that covers the list. Your hands already voted.