Why this exists

I built this because confident, specific, repeatedly-wrong predictions about AI shouldn’t evaporate the moment their deadline passes. People made plans — career decisions, real anxiety — off forecasts delivered with total certainty. The least anyone owes them is a scoreboard.

What stands out to me, reading them in order, is the pattern: an imminent date stated with high confidence, and when it passes, a new date rather than a correction. You can judge that for yourself — the moving-target timeline on the front page just lays the dates end to end.

This isn’t personal and it isn’t a takedown. It’s a record. Anyone who makes specific, dated public forecasts — in any field — should expect those forecasts to be checked against what actually happened. That’s how a healthy information ecosystem keeps its commentators honest, and it cuts both ways: where a call lands, the ledger says so plainly.

So that’s the rule I held myself to: no claim on the ledger without a primary source, no verdict that isn’t checkable, and a corrections policy that means it. Disagree with my opinion all you like — the receipts don’t need it.