Why I built this

The argument was at a family dinner, and the topic doesn’t matter. What matters is the moment someone said something that sounded completely reasonable, and I could see exactly where it broke.

So I explained it. I named the fallacy, walked through the jump, gave an example. And I watched it bounce off. Not because anyone at that table was slow. They’re smart people. But a fallacy explained in the middle of an argument is just one more thing to argue about.

It wasn’t the first time. I’ve been arguing about religion, philosophy and politics since high school, about ten years now, with friends, with family, with strangers online. Somewhere in all those arguments I picked up the ability to hear the jump in an argument almost as it happens. It took me a long time to understand that this isn’t a talent. It’s a reflex, and I only have it because I’ve practiced it thousands of times.

Meanwhile, everyone else gets a different kind of practice. Ads, headlines, viral videos and confident people rehearse slick arguments on all of us, all day. Most of them aren’t exactly lies. They’re built to sound right for the ten seconds it takes to nod along, and then move on before anyone checks.

That’s when it clicked. The people I was arguing with didn’t need me to explain fallacies better. They needed what I’d had: practice. Low-stakes practice, on arguments that sound like real life, with nobody’s pride on the line. You can’t learn to see a strawman while you’re defending something you care about. You learn it beforehand, on a hundred harmless ones, until it jumps out on its own.

I went looking for that kind of practice and found lists. Long, accurate, forgettable lists of Latin names. A few quiz apps that check whether you remember a definition, some with examples that read like no person has ever said them. Nothing that trained the thing I actually use.

I’ve been building software for a few years, mostly passion projects I cared about too much to leave alone. This became the biggest one.

Steelman is the practice I wish I could have handed to that dinner table. It covers 74 fallacies, from the everyday classics to the stealthy ones that hide inside reasonable-sounding talk, each researched against sources like the Stanford Encyclopedia of Philosophy. Every example is written the way someone would actually say it. In the debates, you read a conversation line by line and decide what’s a fallacy and what’s a fair point. Real arguments are mostly fair points, with one sneaky jump somewhere in the middle, and that’s the jump you’re training to catch.

There’s a trap on the other side too. The person who shouts “fallacy!” at everything isn’t thinking clearly either. So in Steelman, flagging a fair point counts as a mistake, a false alarm. Every lesson shows when an argument that looks like a fallacy is actually fine. The goal isn’t to win more arguments. It’s to see them clearly, including when the other person is right.

That’s where the name comes from. A strawman is the flimsy version of someone’s argument, built to be knocked down. A steelman is the opposite: the strongest, fairest version of it. Steelman trains you to catch the first and respect the second.

It’s for anyone who knows the moment I mean, at a dinner table, in a group chat or in a debate round: you know the argument is wrong, and you can’t say why fast enough. It’s for the debater who freezes in cross-examination, the student who wants to take an argument apart properly, the LSAT taker drilling flaw questions on the train, and anyone tired of nodding along to things that only sound right.

Here’s what I’ll keep doing. Every fallacy will be checked against real sources, and every example will sound like something a person would say. The app will teach you when an argument holds up as carefully as when it falls apart. And it will never promise that you’ll win every argument or become impossible to fool, because nobody is, me included.

It’s a reflex. You train it. I’ll keep building the place to do that.