Re: [agi] Re: Pre-Release: Building On GPT-2's Successors "Blender" and "PPLM"

2020-06-22 Thread immortal . discoveries
Oh, also not only does DNA model patterns ex. code many fingers using 1 
template, but also models patterns of things in the environment. For example, 
it finds that it's common the organism transverses the ground, so wings are 
most useful.

--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T3cd584667cb2384b-M74fbc629ada64a08867901d1
Delivery options: https://agi.topicbox.com/groups/agi/subscription


[agi] How To Advance

2020-06-22 Thread immortal . discoveries
How I work and make advances in AGI. An example of progress follows. It's fast 
yet slow and really easy. Discovery in Action:
https://www.youtube.com/watch?v=G_-bgxlL3A8
--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/Tc8eccc8667b92e00-M591f15676cd02512e3481edc
Delivery options: https://agi.topicbox.com/groups/agi/subscription


[agi] Neuroscience Subreddit

2020-06-22 Thread A.T. Murray
Both of the neuroscience Nobelists (Eccles; Hubel) who aided us in our
never-ending failure of an AI project have years ago passed away. Still we
soldier on, even if the goal recedes further and further away from us.
During the week we gather up AGI ideas we would like to communicate, and
then on a Monday like today we pounce.

Sheltering in place from the coronavirus, we spend an inordinate amount of
time spreading AGI memes in several dozen subReddits which are germane to
our tilting at AGI windmills, Miguel. Recent developments in the Reddit
ecosystem have suddenly made it fun and enjoyable to game the ecosystem.
Since most of the geniuses of Earth and the finest minds of humanity are
gathered here on the AGI Mail List, your Eminences are the first to hear
the following story, nay, confession of an AGI addict.

For years there were only a few subReddits which let you declare a "flair"
or descriptive name-tag from a rigid list of available options, and even
fewer allowed you to edit the text of your chosen flair. Suddenly in this
Spring of 2020 Reddit has gone wide open on the permission for Redditors to
choose and edit their assigned flair. ["They shouldn'a done that" -- famous
Mentifex quote.] Do you realize what happens when you grant
Persons-Of-No-Account like Mentifex the Power of the Flair? Extremely
flairbuoyant behavior is the ineluctable result. Why, there was one recent
incident at
http://old.reddit.com/r/singularity/comments/hcddgg/software_20_andrej_karpathy_lead_tesla_researcher/fvetf0x
where Mentifex commented with the single word "Thanks" and let his username
("AiHasBeenSolved") and his Chosen Flair ("AI Mind Maintainer") carry the
entire message on the medium, Mr. McCluhan.

When your wannabe-AGI-er Mentifex started using more presumptuous Reddit
flairs, suddenly Redditors started taking the Old AGI Dude much more
seriously than before.At
http://old.reddit.com/r/consciousness/comments/h9gmos/what_is_consciousness_in_a_nutshell/fuwna32
one adroit comment by AiHasBeenSolved launched a major discussion of "What
is consciousness in a nutshell?"

Your Truly, Oh Advanced Sapients, learned that certain subReddits possess
more than a million signed-up adherents ripe for zapping with memetic
infection, such as the http://old.reddit.com/r/iamverysmart subReddit
"Where everyone is a quantum scientist". Your not-really-worthy-to-be-here
incipient sapient Mentifex learned the trick of embedding a link inside a
comment near the top of an "iamverysmart" discussion so that he could then
at https://livetrafficfeed.com/live/ai.neocities.org watch the world map of
sometimes more than a hundred curious Redditors innocently click the link
and, depending on their intelligence quotient, get their high-IQ minds
infected or their low-IQ minds befuddled.

"Cut to the chase", you mutter? Okay, three days ago your Merry Memester
saw a Redditor asking in the neuroscience subReddit what you can do with a
neuroscience degree. But there was a message indicating that some one had
[GASP!} posted a non-academic link and now the offending Comment was being
held up pending moderator approval. Dear Titanic Intellects -- you would
not tolerate such interference, and neither did Mentifex.He carefully
crafter a no-link no-wait message which appeared beneath
http://old.reddit.com/r/neuroscience/comments/hbz1m0/what_can_you_do_with_a_neuroscience_degree/fvc5lpj
as a comment stating "AI Mind Maintainer is one option," which prompted
"what is it?" as a request for the following meanwhile CENSORED
elaboration:

"AI Mind Maintainer (q.v.) is a proposal for neuroscientists and computer
scientists to work on the maintaining and governance of artificial AI Minds
composed of memory channels and concepts resulting in volitional free will
and conceptual thinking. Such AI Minds already exist as software coded in
Perl and Forth and (tutorial) JavaScript. Each AI Mind achieves Natural
Language Understanding (NLU) by assigning quasi-neuronal associative tags
to verbal input words so that a sentence of thought is retrievable in the
course of _spreading activation_ and association from concept to concept.
Thought occurs in English or in Latin or in Russian. The result is AI
guaranteed to think."

While wily Mentifex had evaded obstruction in his meme-spreading,
CENSORSHIP eliminated both his messages, but not before neuroscientists
used Google to find and visit the http://ai.neopcities.org/maintainer page,
as shown by the following livetrafficfeed records:

 [United States] [Mobile Safari] [iOS] A visitor from Richmond, Virginia
using Mobile
Safari Version 13 on iOS 13.3.1 (Apple iPhone) referral from "Google"
viewed "AI Mind Maintainer Overview and Job Description" Just now

 [United Kingdom] [Chrome] [Mac] A visitor from Cramlington, England using
Chrome Version
83 on Mac 10.15.5 (Apple) referral from "Google"
viewed "AI Mind Maintainer Overview and Job Description" Just now

 [United States] [Firefox Mobile iOS] [iOS] A visitor from Mountain 

Re: [agi] Uhm, Does Consciousness Collapse the Wave Function?

2020-06-22 Thread John Rose
Seems as though individually and societally sometimes advances in intelligence 
are preceded by expansions of consciousness and sometimes vice versa. Almost 
like two "waves". For example, when you become aware of a "bigger picture" of 
something or when a society does, it's as if there is a fabric of consciousness 
that intelligence operates within. Consciousness being related more to a 
communication layer and intelligence functioning within the physical world, 
like Wisner's entropic force, evolution would optimize the communication layer 
to minimize energy expenditure with maximum communication efficiency across 
nodes, essentially, the multiparty communication complexity. I would view 
natural language more from the perspective of a compressive communication 
protocol since words are really representations of lossily compressed concepts 
strung together to be decoded by other similarly decompressive and conscious 
parties. If the other party is software it would emulate and reconstruct the 
compressed conscious representation in an attempt to accurately decipher. But 
we do realize that some words are more discrete than analog, IOW lossless 
verses lossy.
--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T51eb63417278f283-Me0e5053d0c3ec5ac7844d7dc
Delivery options: https://agi.topicbox.com/groups/agi/subscription


[agi] Re: IGPT

2020-06-22 Thread stefan.reich.maker.of.eye via AGI
Yeah that's what I did
--
Artificial General Intelligence List: AGI
Permalink: 
https://agi.topicbox.com/groups/agi/T60bb245c0215eb2e-M5e48e955c45263949b28c06b
Delivery options: https://agi.topicbox.com/groups/agi/subscription