AI
Why I came to AI, what I build, and what I have learned — the personal side. From 2023.
I believe AI really does save time — that is a fact, not marketing. But it is not only about speed: if you train the system properly and give it the right data, it can stand in for an adviser in a given task. That is no longer just a tool, but a genuinely capable assistant.
2023
Where I started
The idea did not come from books, but from my own business. In HOBOT robots I had long been fascinated by how a device without eyes or cameras — just algorithms — builds a map of the room and a route in its “head”: where it has already cleaned, where it still needs to go. That was pure programming, and that is exactly what I know how to do. That is when I thought: if algorithms can build a map of a room, what can they build in a business?
That is how Evolution AI was born — deep technical writing lives there; here is the personal view.
How I use AI every day
I use AI constantly myself — first of all in development. Code used to take a day or two, plus another day for testing. Now it is 1–2 hours to write and an hour for testing and fixes. My philosophy: AI does not replace a person, it greatly speeds up the process. A person is still needed — but as control, and as the one who sets the tasks.
A telling example — I have several servers where agents run for clients. A personal technical agent logs into the server itself, updates the system, syncs the data — just like a person would. This is not a script that blindly follows steps one-two-three: it has a goal, and if something goes wrong at step two, it tries again, looks for another path — and sees it through. At the end it simply sends a report.
I also use AI for content — searching for ideas, wording text — especially where a lot of similar work has to be done regularly.
What I think about the fears around AI
There are many myths around AI, and I hear some of them all the time. That AI will take people’s jobs — not quite: rather, the requirements of work change, and automation frees time for what needs human interaction, creative thinking, and empathy. That AI will become conscious — in practice it is still a data-processing tool, able to learn and adapt, but without self-awareness and emotions; real machine consciousness is still a long way off. That AI does not make mistakes — it does, and how: like any technology, it is subject to bias in the data and to the limits of algorithms, so it is important to build systems that can learn from their own mistakes. That AI pays off instantly — not that either: first you need investment, time to integrate and tune, and the return becomes noticeable only when the systems are truly built into the processes of the business.
My personal conclusion is simple: AI is not a magic button and not a threat, but a serious tool that needs the same sober approach as any other in business.
Coming soon
Personal stories and reflections on AI agents — blog pieces will land here too.