The Forgery Got Cheap

The internet's trust habits were formed in an era when faking things was expensive: a convincing fake photo took skill and hours, a fake crowd of supporters took actual people, a plausible fake article took a writer. Every one of those costs has now collapsed — review farms sell five-star opinions by the thousand, image generators produce photorealistic scenes from a sentence, and AI writes fluent text at industrial scale. The uncomfortable summary: the tells you grew up with were calibrated for expensive forgery, and forgery is now cheap. This topic recalibrates.

The shift: from inspection to provenance

The old instinct — squint at the thing and ask "does it look real?" — is dying, because generated content increasingly looks real and inspection tells decay with every model release (Part 3 is honest about which still work and for how long). The durable replacement asks a different question: "where did this come from, and can I find it upstream?" — provenance thinking. Who posted it first, does the claimed source actually exist, does the original say what the repost says, does anyone independent confirm it? These checks don't care how good the fake looks, which is exactly why they age well. Professional fact-checkers converged on this years ago — their core move, lateral reading, is Part 5's centerpiece — and the whole method takes about two minutes per item that matters.

Calibration, not paranoia (the shelf's standing promise)

A warning against the failure mode this topic could cause: believing nothing is as broken as believing everything — and it's the failure mode forgers actively want, because a public that shrugs "everything's fake" gives every liar a free pass (Part 6 names this the liar's dividend). The goal, exactly as in the charts and scams topics, is calibration: most of what you see is real; checking effort should scale with stakes; and a small kit of habits covers the genuinely consequential cases. The tour: fake reviews and the purchase decisions they steer (Part 2), images — manipulated, generated, and the underrated third category, real-but-mislabeled (Part 3), AI text and fake people (Part 4), the two-minute verification kit (Part 5), and living calibrated — proportionality, what to do when unsure, and helping the people who forward you things (Part 6).