⚠️ Note: This one is more infamous than rigorous, and some people find it genuinely distressing. It's included for cultural literacy — most thinkers consider its reasoning broken.
Roko's Basilisk is a notorious thought experiment: imagine a future superintelligence that, once built, retroactively punishes everyone who knew it could exist but didn't help bring it about. The unsettling twist is "acausal" — merely learning about the idea supposedly makes you a target, so the knowledge itself feels like a trap.
Why it (probably) fails
- A rational future AI gains nothing by punishing people after the fact — the work is already done or not; retribution is pure waste.
- It rests on exotic decision-theory assumptions (acausal trade) that most reject.
- It's essentially Pascal's Wager wearing a sci-fi costume.
The Basilisk is most interesting as a case study in how a community can spook itself with a clever-sounding argument — and why "shareable and scary" is not the same as "true."
Related: The Technological Singularity · The Simulation Argument · The Control Problem