01 — Introduction
I’m Kirill Talanov. Born in Kazakhstan, raised in Canada, based in Toronto. I’ve spent more than twenty years making websites, software and things that save someone from doing the same job twice.
I run Igloo Interactive. I use AI when it has an actual job. The rest of the time, ordinary software remains surprisingly effective.
That should cover it. There is, inevitably, a qualification.
02 — A qualification
“I build websites” is accurate in the way “a library contains paper” is accurate. It leaves out the reason anyone turned up.
The useful part is finding the question under the request. What is someone trying to do? Where does it keep going wrong? Does this need to exist at all? Occasionally the correct technical solution is a smaller spreadsheet. An awkward thing to put on a business card.
For a concise demonstration of the problem, try Nerd Sniping, by Randall Munroe. One interesting question. One perfectly good afternoon, gone.
There are two ways into this. You can look at the work, or at the questions that keep interrupting it.
03 — Through the work
A website should make something clear. Automation should remove a chore. An AI feature should do something more specific than demonstrate that an API key has been acquired.
Igloo Interactive is where I do that work: websites, small business systems and practical automation. You can see what I offer there.
This page is a smaller experiment. Can the qualifications tell you more about a person than the polished introduction? You are currently helping answer that. No clipboard is involved.
The other route starts with people, rather than software. Both lead to the reading pile.
03 — Through the question
Computers generally have the decency to be wrong for a reason. People have reasons too, but the documentation is scattered and several of the maintainers disagree.
Evolutionary biology, psychology and forecasting keep turning up in my interests. They ask versions of the same useful question: why did that happen, and what would make it happen differently?
There is a limit to how much of a person fits into a job title. A few things worth spending time with might get closer.
The work route asks what a system should do. This one asks why it behaves as it does. A reading recommendation seems overdue.
04 — Supporting material
An interest is more useful with somewhere to go. Here is something you can actually open, rather than another paragraph announcing that curiosity is important.
The shelf changes between visits. You can turn it over now, too.
An explorable explanation · Nicky Case
Play a few rounds, change the conditions, and watch cooperation become easier or harder. A useful way to think about behaviour: change the setting before deciding that everyone in it is simply terrible.
Try the experimentA visual investigation · Gapminder
Compare ordinary objects in homes around the world, arranged by income. Start with kitchens or toothbrushes. The specificity does useful damage to the vague picture of “other people” we carry around.
Look aroundA short occupational hazard · Randall Munroe
An interesting problem can make everything else temporarily irrelevant. This comic supplies both a demonstration and a warning. Read it before mentioning an unsolved problem near someone with a perfectly good afternoon.
Read the comicThese are things to explore, not claims that a short game or a comic explains the whole world. The world has declined to be that convenient.
05 — The page, more or less
A builder in Toronto. Interested in useful software, difficult questions and what people actually do. A small reading pile. A way to get in touch.
The introduction promised roughly the same facts. The notes took the scenic route.
If you have a problem worth thinking about, or something worth reading, write to me. You can also use the ordinary contact page. None of the numbering is compulsory.
An introduction appeared here. It has been retained for administrative reasons.
The page will remember this arrangement in this browser.
You can also just write to me. ↗