The Future of Everything is Lies

Link: https://aphyr.com/data/posts/411/the-future-of-everything-is-lies.pdf

Author: Kyle Kingsbury (Aphyr), 2026

Kyle Kingsbury — best known for Jepsen, the distributed systems testing framework that has humbled dozens of databases — turns his diagnostic lens from distributed systems to the epistemic crisis created by AI-generated content. The argument is characteristic Kingsbury: precise, unsentimental, and grounded in the mechanics of how systems actually fail.

The core thesis is that we are building an information ecosystem where the cost of generating plausible content has collapsed to near-zero, while the cost of verifying it remains high and is rising. This asymmetry is not a temporary problem that better models will fix — it is a structural property of the system. When generation is cheap and verification is expensive, the equilibrium is more lies, not fewer. The volume of synthetic content will exceed any human or automated verification capacity.

Kingsbury draws the parallel to distributed systems explicitly: in both cases, you cannot trust the nodes to report their own state correctly. You need independent verification, cross-checking, and proofs. But the web doesn't have consensus protocols. It has hyperlinks and social signals, both of which are gameable at scale.

The piece is worth reading for anyone thinking about how AI intersects with truth, trust, and information architecture — not because it offers solutions (it largely doesn't), but because it frames the problem with the same rigor that Jepsen brings to database correctness: here is exactly how this fails, and here is why the obvious fixes don't work.