The cease and desist I was secretly proud of

Version 1.0 · Published 2026-09-09

The letter arrived from a large streaming company’s lawyers. It was polite, clearly written, and asked us to please stop doing whatever it was we were doing.

The embarrassing part is that we kind of deserved it. The other embarrassing part is that I was secretly proud of it.

What we were actually doing

This was the very first version of Unless, years before conversational AI was a thing anyone talked about. The idea was already there in spirit, even if the technology was nothing like what we run now.

We wanted to make websites feel personal. Back then that meant machine learning and a fair amount of client-side browser work, changing what someone saw based on who they were. Crude by today’s standards. The intent was not crude at all.

The real target was never entertainment. It was the complex legacy interfaces you find in finance, insurance and administration, the ones that are genuinely hard to understand and that people have to use anyway. Our bet was that if the content adapted to you, it would feel more familiar, and familiar things are easier to trust and easier to get through.

Why we were on their website and not on an insurance portal

Here is the problem with demoing that idea on the thing it was built for. Nobody can see it.

Show someone a pension portal that has been quietly reshaped around them and they will nod politely, because they have no memory of what it looked like before. The change that matters most is exactly the change that is hardest to notice. So for demos we used websites everyone already knew by heart.

The changes were not exactly subtle. What struck me, and what still strikes me, is how instantly people accepted them. Nobody said “you have altered this page”. They behaved as though it had always looked that way, and then moved on to asking what else it could do. That reaction was the actual product insight, and we got it from a demo we had no business running.

Caught by a machine, not a person

Their security system noticed us fiddling around long before any human did. Full credit to them for that; it was a better catch than we deserved.

A while later the letter came. No drama, no threats, no numbers with lots of zeros. Just a clear and friendly request to stop. Classy, genuinely, and I have thought about that tone since whenever we have had to write something uncomfortable to somebody else.

The sensible reaction would have been mild panic. We stopped immediately and moved our demos somewhere safer, which took about an afternoon and should have taken us about five minutes of thought a few months earlier.

Why it felt like good news

But underneath the sensible reaction, a small part of me was thrilled.

A real company, with real lawyers, had noticed us. Not a friendly investor being encouraging, not a conference audience being polite. An organization with no interest in our feelings had run into something we built and had to respond to it. For a tiny startup, that letter felt like proof of life. We had done something real enough to set off an alarm somewhere.

I am not recommending this as a strategy. It is a bad strategy. It is just true that of all the feedback we got that year, the only piece that came with no incentive to be nice was the one from the lawyers.

What survived

Looking back, it is one of my favorite early memories, and not only for the story. We were clumsy and a little reckless, and the core belief was right.

Interfaces should adapt to the person, not the other way around. That belief runs through everything we build now. What changed is the technology and whose website it runs on. Instead of fiddling with someone else’s site, we give a company a Customer Agent that understands the people it is serving and adapts the experience to them. Inside their own product, with their permission, and with a record of every decision it made.

Same dream, far better tools, nobody’s lawyers involved.

That is startup life, more or less. You stumble toward the right idea, and occasionally you set off somebody else’s alarm on the way there. The early, clumsy version of an idea is often right at the core. That letter was not a failure. It was the first sign the idea mattered.