jonny (good kind) on Nostr: The language models are a protection racket. The business decision was to set the ...

The language models are a protection racket. The business decision was to set the internet on fire and charge for water.

https://www.theverge.com/2024/9/24/24253452/microsoft-correction-ai-safety-tool-fix-errors

Paying extra to correct dogshit model output is the same strategy as flooding the internet with spam and selling tailored chatbots as the cure. The long range consequences after all the social information systems have been hollowed out to be replaced by various shades of language models will be catastrophic. The critical part of this is not that it is "AI" correcting "AI," but that the trade is surrendering to total life surveillance to condition the model output. We are being forced to trade the (however problematic) collaborative web of human beings helping humans for the private surveillance oracle.

https://jon-e.net/surveillance-graphs/#the-enclosure-of-information-access-as-a-private-exchange-with-a

#SurveillanceGraphs