@doomslide IMO billions-LOC lean proof is same as numerical approximation of Gauss integral instead of finding the antiderivative itself. Proves something, but contributes nothing to understanding. The death of mathematics is overstated.
@grep_cat_ Bro, there is no way that's happening. If AI companies really end up causing massive social unrest, there will be public hangings of these people. This is human world, and we'll do whatever to keep it that way.
@JacekF90@BellPeSound A co mieli kurwa zrobić, taranować z ryzykiem uszkodzenia swojego samochodu i uniemożliwienia pościgu? A wtedy ten bydlak by zabijał dalej. Co więcej oddalali się w ten sposób od gęstego od przechodniów centrum. Ludzie co wy macie kurwa w głowach? To nie Hollywood
@the_aiju Yeah, feels like gaining knowledge lost its "wow" effect. It's more like stupid trivia that LLMs will do some brainfuck around instead of actual useful, applicable knowledge.
@Sentdex@ChenTessler when r u gonna test gpt-6 for ur robotics stuff? interested in ur opinion on robotics/3d/spatial reasoning advancements from this new model
@JustSayElvis Nobody cares. They will say something about "you can do more ambitious things" then vanish without a single comment on socio-economic repercussions.
@ronSzab9@tenobrus It's kind of different when you have entire structure of cooperation memorized and are being pressured to put into practice, contrary to subconscious incentives that create the structure ad hoc.
@ronSzab9@tenobrus Yes but we have many internal reward systems that drift us towards these behaviours, with language or without. It's just evolution-induced intuition to cooperate. I would argue that these LLMs behaviours are much less innate, rather a some form of crooked mirror.