A collection of views

Artificial Intelligence.

I use AI every day and I do not trust it blindly. Here I note where it genuinely saves work, where the result has to be checked, and which promises do not survive the first enthusiasm. Experiments, not predictions.

Key questions

How do I use AI without handing over my judgement?

I write down what I think first: the plan, the text or the decision. Only then do I ask the tool to criticise it. The questions I ask most often are: Which assumptions are in this proposal that I haven't written down? Who would disagree with this, and why? Which three things in this text would a reader refute most easily? Each of them asks the tool to show me something, not to do something instead of me. More in I don't want an AI that does everything for me.

Read the article

Where does AI really save me time in development, and where don't I trust it?

It saves a lot of time, but speed is not correctness. A suggestion is not a verified result, and a tool's explanation of its own code is not independent proof. A concrete case from building this site: a problem with access to public images only showed up after the upload, and checking the actual address found it, not the report of completed steps. My thinking is in AI saves me a lot of time in development. I still don't trust it blindly.

Read the article

Is a good prompt enough for a good result?

A prompt solves a moment; a specification solves a project. For anything that will outlive one conversation I write a short brief, the project rules and the acceptance criteria before I ask for the work. More in A good prompt is not a good specification.

Read the article

What matters on this topic

"Do everything for me" is the worst way to use AI, I think, because I'd still have to defend the decision. I treat an AI answer as a map for exploring, not as a source of facts, and I check comparison tables and ratings by hand. A tool likes to praise unless I steer it elsewhere, so I ask it for blind spots, not for approval.

Critique of my own work is the best use I know. I don't hand over writing and planning completely: a draft, yes; my voice, no.

Google's guidance on helpful content (updated 1 October 2026) asks, among its questions about how content was produced, whether the use of automation, including AI, is evident to visitors. So when I write about AI I separate what is a verified fact, what is opinion and what is a suggestion, and I say what I tested myself.

In practice

The topics of AI and development overlap: articles about working with AI while programming are under Software development, questions about search and visibility in AI answers under SEO and AI search. Here I collect the general questions: where AI helps with research, ideas, planning, texts and decisions, and where I stay responsible.

Articles on this topic.