I’ve built a lot of things. This is the one that has to matter.

Most drugs are still found, not made

They are found the way you find one key in a drawer of ten thousand keys. Make an enormous number of candidates, test them all, hope. It is slow because it is a hunt, not a design.

What changes that is being able to say what you need, and have it built.

That is the long ambition, and I don’t think it is as far off as it sounds. But nobody gets to start there.

So I started where life starts

Almost everything that happens inside you is done by a protein. They cut, carry, build, hold a metal in place, take a thing apart. Nothing further up the chain gets made until somebody can make one of those to order.

They are also the part AI has genuinely become good at, in the last two or three years, in a way that is not hype. So that is where I started — the way I have started everything I have built: at the smallest real thing, done properly, rather than at the headline.

It is where the name comes from. Vida — life.

Twenty years, pointed at one thing

My degree is computer science with bioinformatics. So the biology here is not something I picked up from a paper. It is what I studied.

Then ten years writing software. Then ten years running products — at Wix and before it — taking things from a first idea through to millions of people using them, and building them with the AI methods as they arrived rather than years after.

And through the last ten of those, my own things on the side: shipped on my own, run on my own, used by thousands of people. One of them, a music app, has been in the stores since 2021.

Four separate lives. Vidika is the first thing that needs all of them at once.

And the thing I cannot do

I am a computational person, not a bench biologist. I have never run the experiment that decides whether one of these designs actually works, and I am not going to pretend otherwise.

That limit is the reason the whole thing is built the way it is. I could not lean on my own judgement about a molecule, so I built something that is not allowed to lean on its own either. Every number Vidika shows you came out of an instrument that produced it. Nothing is recalled from memory, nothing is filled in to make a page look finished, and a design that cannot be traced back to the machine that made it is thrown away rather than shown to you.

A number that was invented is worse than no number, because it gets believed.

Those are not good intentions written in a document. They are checks in the code, and a request that breaks one is refused before it costs anybody anything.

One person, which is not what it used to mean

Five years ago this would have needed a floor of a building. It does not any more: the instruments are rented by the hour, the models do the work that used to be done by hand, and the ground has moved under what a very small group can cover.

A team of one or two can now do what needed a team of fifty.

And not alone in the way that phrase usually means. I have partners and advisors from the genomics and biotech world who look at the work, argue with it, and tell me when something is wrong. The judgement I do not have, I borrow from people who do.

What is left is small on purpose. No queue and no support tier. I build this at a desk at home, mostly at night, and it is early — invitation-only, running real work for a design partner, and every campaign still ends at somebody’s bench, because Vidika designs molecules and does not make them. Anything only a bench can settle is labelled as exactly that, right beside everything that was proved.