Post #4503670

>>914754931 idk an LLM in a fundamentally language mediated action-space sounds like at some point that class of model could pull off some shit I understand dismissiveness is cool, but I've never really been a cool guy so I've always felt the urge to push back a little against the cool "autocomplete" view on these machines. I mean autocomplete is fine as long as we place humans acting in language space on the same tier, our souls may be infinite, but bound to a linguistic ontology we're constrained to the same kind of autocomplete actions unfortunately, and why a lot of acid heads and mystics break down at that barrier and write koans or other confusing stuff. IDK I'm on my 3rd brewski, but cool reply maybe idk anything and living in a speculative fiction retarded reality :nickland: