Pavel @spavel.bsky.social · 20/09/2026Every chatbot prompt box says "check answers for accuracy." But increasingly, "humans in the loop" refuse to do that. Instead, they become complicit in the AI industry's project of destroying the provenance of information. But you are person, and it is an app. It can't take responsibility for you.productpicnic.beehiiv.comAssigning agency to AI means surrendering your ownClaude is not a colleague, collaborator, or sentient being. It is a SaaS product.8:54 PM · Sep 20, 2026 825485
Dan Turner @ddt.bsky.social · 20/09/2026↪ Replying to @spavel.bsky.socialI'd go even farther. The centering on answers is a bit of a decoy; what's key is checking sources. 030
Pavel @spavel.bsky.social · 20/09/2026Thanks to @jonnelledge.bsky.social @kjcharleswriter.com @emilymbender.bsky.social @doctorow.pluralistic.net @wolven.blacksky.app @axbom.com @sarahoconnorft.ft.com @ethanmarcotte.com @ali-alkhatib.com - and special thanks to @eryk.bsky.social whom I end up citing 3 separate times in this week's essay 4241
Alexander Hay @alexanderhay.myatproto.social · 20/09/2026↪ Replying to @spavel.bsky.socialRather than check the results, why not just look it up yourself? AI's USP is the marketisation of laziness and complacency. 210
Lloyd Sullivan @lloydgonkillya.bsky.social · 20/09/2026↪ Replying to @spavel.bsky.socialReminds me of Cory Doctorow's idea of the reverse-centaur. 220
Expat Vet of DOOOOOOM @expatvetofdooooom.bsky.social · 20/09/2026↪ Replying to @spavel.bsky.socialPeople don't want to have to do this. That's why it's so pernicious 000
clothilde @cynar.bsky.social · 20/09/2026↪ Replying to @spavel.bsky.socialthe black box invocation interests me because increasingly law enforcement & courts outsource evidence gathering & analysis to proprietary 3rd party black boxes. Essentially we are on the brink of reality wars where the custodians are product managers & requirements engineers. 030
Sander de Regt @shadowmakersdr.bsky.social · 21/09/2026↪ Replying to @spavel.bsky.socialIt does become increasingly more difficult to check an answer for accuracy if all the sources you might want to use to check become filled with AI generated 'info'. 050
Robert - DevOpsBob @ruzitschka.at · 21/09/2026↪ Replying to @spavel.bsky.socialAnd ultimately people also can’t take responsibility for the GenAI output. Human in the loop is a fallacy. robertruzitschka.substack.com/p/the-accoun...robertruzitschka.substack.comThe Accountability ConundrumThoughts on accountability, humans-in-the-loop, reviews and power play. 000