←back to thread

273 points geox | 2 comments | | HN request time: 0.001s | source
Show context
gcanyon ◴[] No.40712874[source]
You have to think that there were breakthroughs in communication technology — not just language in general but possibly also one individual who happened to be good at explaining things, either before or after language, who both taught more people, but also taught them how to teach — that led to step changes in technology.
replies(8): >>40713012 #>>40713840 #>>40713885 #>>40714141 #>>40714994 #>>40716449 #>>40717648 #>>40718490 #
dboreham ◴[] No.40713012[source]
Theory: there are no humans without language. Consider: what language do you think in?
replies(7): >>40713064 #>>40713200 #>>40713207 #>>40713659 #>>40713766 #>>40713849 #>>40714603 #
BenFranklin100 ◴[] No.40713207[source]
We think in something Steven Pinker termed “mentalese”. It is distinct from human language. There are many examples of people who are cognitively normal, but either lack or have severely impaired language for various reasons.

The distinction between reason and language is not widely appreciated, and is a main if not the primary reason people overestimate the abilities of LLMs.

replies(2): >>40713256 #>>40713854 #
1. davidmnoll ◴[] No.40713854[source]
Stating the mentalese hypothesis as fact is a bit tenuous
replies(1): >>40717916 #
2. BenFranklin100 ◴[] No.40717916[source]
Sure, like most areas involving intelligence, much work is yet to be done. However, mentalese or the Language of Thought Hypothesis, occupies a much stronger position than the hypothesis that human language itself is our fundamental engine of reason, which is almost assuredly not true. This last fact has serious implications for the valuations of the majority of AI companies.