Why attempts?

The friction

Much of the internet's advice was never tested by the person publishing it, and AI can now produce that kind of advice without limit. The tell is everywhere: the most upvoted reply to any viral trick is some version of "has anyone actually tried this?" People are not short of instructions. They are short of outcomes.

The shift

When generating advice is free, a human attempt becomes the scarce unit - even a self-reported one, because advice rarely includes one at all. And attempts are cheap to make honestly: you were going to try the recipe, the fix, the routine anyway. Logging what happened turns everyday trial into shared evidence.

The concept

An attempt log for the web. Not reviews (opinions about products), not comments (arguments under content), not editorial testing (one lab's verdict). The atomic post is a structured report: this claim, these conditions, this outcome. Claims accumulate attempts the way papers accumulate citations. The first thousand users would be the people who already reply "I tried it" in comment sections - the DIY forums, the recipe tweakers, the productivity-method testers - given a place where that reply is the whole point.

What this demo can and cannot claim

This demo stores attempts in your own browser. It cannot verify that any attempt really happened - anyone can write a report, true or not, and AI can too. A claim with many attempts is a track record of self-reports, not proof. The full concept would need hosted accounts and reputation to give attempts weight - deliberately out of scope here. Prior art is real: Snopes and blogs test hacks editorially, Allrecipes has "I made it" photos for recipes, and science runs replication registries (FORRT's FLoRA). What we did not find among checked players: a general consumer attempt log where the attempt is the content and every claim carries its outcome history.