Article Examines AGI Sample Efficiency In Human Brain Versus LLMs

my new article on AGI and sample-efficiency (and some spicy takes)

https://open.substack.com/pub/scaling01/p/on-agi-and-sample-efficiency

vertical take-off on the substack !11!11

I will post some longer form articles there from time to time

https://scaling01.substack.com/

please tell me this shit works now 😂

what do you mean it may take a few minutes

what's so hard about adding 19 database entries lmao

@scaling01 It won't open

@iamgroguu does it work now?

@scaling01 sample efficiency is the real bottleneck

more compute without better inference paths just scales the noise

@roarkittys lmao

thanks for letting me know

@scaling01 Yezzir

@scaling01 Doesn’t open or let me type email in substack

@scaling01 Hell yeahhhh

@scaling01 so good! this is sorely needed.

only thing i would push back on is calling awareness either way. agree likelihood is very very low but it cannot be ruled out because we don't know what it is. this is going to be very problematic [forever].

@scaling01 so, AGI would be not a model , but a company?🙂

@scaling01 solid, thanks!

@scaling01 Yesss Lisan al Gaib on the stack

@scaling01 sample efficiency is the real bottleneck

@scaling01 “From that it’s clear that any comparison between these two systems regarding their sample-efficiency will be wrong in some way, because the purpose of the brain, which is a general learning system is to continually train a model, while an LLM is just the model.”

@scaling01 great weekend read

@iamgroguu let's go, first try

@scaling01