@Matt

We are humans.
The internet is a team of humans (multi-agent ensemble).

Google is not AGI. AGI would be AGI, and a team of AGIs would be a multi agent 
ensemble of AGIs.

Google search uses BERT but this is an AI. Looking all of Google Search, still 
not AGI either.

So no, it's not going to reach AGI before we do...cuz it's not AI much, things 
like BERT is though, that's AI - something that intelligently searches its 
memory and predicts the future accurately on many diverse prompts. Oddly Google 
Search sounds like this, though it isn't really doing that now is it, the only 
thing in it that does is BERT or something very similar plus BERT. Google can't 
predict the next letter, BERT does the matching and predicts that as the 
suggestion, no extension/entail to the prompt like GPT-2 does. Yes in the 
search bar it has suggestions, but this is either BERT too or a simpler memory 
system just for some help.

Yes making a better processor in a lab is part of building AGI, none of us do 
that one yet either, we keep our hands clean, but like adding more data, it is 
useful to some extent and only part of it, there is still many more AI 
mechanisms to add that'd make it into ASI level, so, no, we are not done AI and 
need to only change or speed up the optimization/ code/ processor.

Alexa/ Siri are interesting tools but in the end you need millions of such to 
help us solve many big problems, these are more pet peeve problems they solve. 
They don't do what DALL-E can do; solve many prompts fed to it ex. give me an 
image of pikachu in black & white upside down in the shape of a bench, stacked 
on another such clone, with blue stripes on the legs.
------------------------------------------
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T7c7052974ce450f1-Mdd9f33b7bc2a37a081d24274
Delivery options: https://agi.topicbox.com/groups/agi/subscription

Reply via email to