2 ms·
Insightful comment, but I'm wondering if you would comment on why? You said you don't expect this to change. Do you think there's some limit inherent to any cyb
by xp84 4mo ago
Insightful comment, but I'm wondering if you would comment on why? You said you don't expect this to change. Do you think there's some limit inherent to any cybernetic 'mind' that simply cannot be overcome by any amount of advancement in the currently-known types of research? Is this specifically about LLMs?
For me, I simply don't know. I know that I have a soul, but I'm still not sure if my brain's output could or could not be perfectly replicated by a simulation, or, to propose a much easier experiment, if a simulation of a brain could be made and trained, if it would be distinguishable in its output from a real brain.
- PaulRobinson 4mo agoApologies for the late reply, 10 days on. I think the human mind is more complicated than regurgitation. I think repeating things we've seen is key to learning and reasoning, but there's a ton of other stuff we don't full understand that LLMs are not even attempting to solve for. To get to being able to simulate it, I am reminded of the Borges story On Exactitude in Science [0], where we end up just building, well, an actual brain. Can we do that within 3 orders of magnitude as efficiently as the sum of human evolution in our lifetimes? I doubt it. [0] https://en.wikipedia.org/wiki/On_Exactitude_in_Science https://en.wikipedia.org/wiki/On_Exactitude_in_Science