Mx. Chara Aznable (they/them) of Pnictogen on Nostr: Abstract reasoning implies the ability to make inferences about incoming data, ...

Abstract reasoning implies the ability to make inferences about incoming data, *judgments* about whether the data is good or consistent, and that is NOT what the techlords actually want LLMs to do. They want a device that regurgitates anything that's put into it, because one of the (unspoken) use-cases for genAI is the distribution of political propaganda. The techlords are fascısts; they want a machine to ratify their fascısm.

It's necessary for the survival of human civilization—yes I write hyperbolically but, I think, truthfully—for humanity in general, and thus the U.S. press establishment in particular (for they have an enormous cumulative effect on popular opinion) to understand that the output of a machine is not automatically trustworthy. But that's precisely what the #tech sector WANTS people to believe, and it's the reason they're hyping up the purported genius intelligence of their non-intelligent LLMs.

(cont'd)