On Sun, Feb 17, 2019 at 2:10 PM justcamel wrote:
>
> So you are a supporter of the idea that AGI will emerge from defining 10k
> nonsensical public variables? Why are you giving this full-time troll a stage?
Dude... this is the Trump Era ... and you're asking why give trolls a
stage? I though
Ben, the world is experiencing never-seen-before levels of suffering,
exploitation, destruction, chaos and insanity. The number of people who
understand the nature of the process which is causing a whole
civilization (or rather their quality of consciousness) to degenerate is
so low ... it's ju
Probably most people here haven't had time to look at 15k lines of code and
form an evaluation of it. Probably most people here share your serious
concerns, we're just not sure that banning someone from a mailing list is a
very important part of minimizing pain/suffering and maximizing
goodness/chi
Taking 30 seconds to look at the file should be enough ... and I said
implement regex/filters that block postings containing certain strings
... I am the last person to ban anybody from anywhere. It's just insane
to have 75% of the list's postings to be about the "work" of a
passionate troll.
Steve,
OK, thanks, I think I kind of see what you mean.
My approach to AI is to simulate human dialog. All our lives, we try to
figure out complex problems in dialog - and are successful -, so when the
AI joins this dialog, it can participate in the truth-finding. Sure, we're
never 100% in our un
Well then let's just up the volume of the non-troll posts :)
On Sun, 17 Feb 2019 at 12:34, justcamel wrote:
> Taking 30 seconds to look at the file should be enough ... and I said
> implement regex/filters that block postings containing certain strings
> ... I am the last person to ban anybody f
I'm not sure how one would go the next step from a random-speech-generating
network like that.
We do want the speech to mean something.
My new approach is to incorporate semantics into a rule engine right from
the start.
On Sun, 17 Feb 2019 at 02:09, Ben Goertzel wrote:
> Rob,
>
> These deep N
> You can't upload yourself. You exist a layer above this physical realm.
THIS is a great statement.
On Sun, 17 Feb 2019 at 08:47, justcamel wrote:
> You can't upload yourself. You exist a layer above this physical realm.
>
> Super Mario can't upload the controlling player "into" Luigi. It does
One can see the next steps from the analogy of deep NNs for computer vision
First they did straightforward visual analytics, then they started
worrying more about the internal representations, and now in the last
6 months or so there is finally a little progress in getting sensible
internal repre
I think an ad-hoc logic would be something like the logic of all computer
programs. No question that a computer program has something to do with
logic, but a generalization of the logic of all programs would involve
something more of an informal logic. You can begin to develop a
systems-analysis ap
These days a symbolic system is usually seen in the form of a network - as
almost everyone in this groups know. The idea that a symbolic network will
need deep NNs is seems like it is a little obscure except as an immediate
practical matter.
Jim Bromer
On Sun, Feb 17, 2019 at 8:27 AM Ben Goertzel
I mean, the results are impressive for sure. Almost on the level of the
usual text-farm pseudo-content made through cheap labor.
Incidentally, Ars Technica's commenters feel that the world just ended
because of this software. (The comment section there is the weirdest ever
IMHO...)
https://arstec
Is that an anti-NN argument? Not exactly sure what you're saying there.
On Sun, 17 Feb 2019 at 15:42, Jim Bromer wrote:
> These days a symbolic system is usually seen in the form of a network - as
> almost everyone in this groups know. The idea that a symbolic network will
> need deep NNs is see
Steve Richfield asks:
> 1. THEORY: In broad computer science terms, how does your system work?
> From what I can tell, it is an ad hoc text manipulation program capable
> of gathering information and answering simple questions within the
limited
> subject domains that have been programmed. Right?
The most significant advancements seem to be made by using NNs with
categorical feature detection or by using discrete systems in a network of
some kind. The networks may not be explicit in discrete methods in but even
in the earliest developments they were intrinsic to the case detection.
Discrete
The argument of discrete vs weighted systems and GOFAI vs neural networks
seem pretty dated to me. So I guess I should be careful on the terms that I
use. A symbolic network would have to operate on different levels. One
thing I have learned is that you do not want to translate information in
terms
On Mon, Feb 18, 2019 at 2:27 AM Ben Goertzel wrote:
> ...
> Don't get me wrong tho, I don't think this is the golden path to AGI
> or anything
By contrast, if anyone does want to look for the golden path to AGI, rather
than the next step down a dead end as usual! In case someone here is
loo
Nothing wrong with pushing your own results if you consider them
worthwhile...
On Sun, 17 Feb 2019 at 21:46, Rob Freeman
wrote:
> On Mon, Feb 18, 2019 at 2:27 AM Ben Goertzel wrote:
>
>> ...
>> Don't get me wrong tho, I don't think this is the golden path to AGI
>> or anything
>
>
> By cont
On Mon, Feb 18, 2019 at 10:05 AM Stefan Reich via AGI
wrote:
> Nothing wrong with pushing your own results if you consider them
> worthwhile...
>
Well, I think on one level it's much the same as Pissanetzky.
Pissanetzky's is a meaningful way of relating elements which generates new
patterns. Yo
> demo.chaoticlanguage.com
It works with "I went to Brazil", but seems to break with "In Brazil,
people are friendly" (it creates "Brazil people" as a node). Any way to
give it feedback?
On Sun, 17 Feb 2019 at 22:48, Rob Freeman
wrote:
> On Mon, Feb 18, 2019 at 10:05 AM Stefan Reich via AGI <
>
Feedback? To me?
Any number of ways to break it. It's old now. 20 years back. And the data
set a few 10's of 1000's of words I scraped up from some websites back in
the day.
Just treat it as a proof of concept: you get (meaningful) hierarchy from
novel rearrangements of word vector "embeddings".
***
I was wrong to start with embedding vectors as the base representation
for patterns.
I now think the way to implement it is not with vectors, but directly
in a network of observed language sequences.
***
This is what we're doing in our OpenCog based language learning project
But we're using
justcamel wrote:
> If you dissolve the identification with form and enjoy true freedom
> then why would you go back to identifying with yet another embodiment?
Normal humans are hard-locked to their existing form. Transmunanists
wish to use technology to change that form to some degree or other,
Spiritual enlightenment is the realization that there never was any
"binding". You being _something_ is just an idea. The 12 year old child
playing Super Mario never really is bound to Super Mario ... it's just
an assumed role ... an identification with the form of Mario ...
You are consciousn
On Mon, Feb 18, 2019 at 4:01 PM Ben Goertzel wrote:
> ***
> ...
> And likely the way to do this is to set the network oscillating, and
> vary inhibition to get the resolution of "invariants" you want.
> ***
>
> But we are not doing that. Interesting...
Cool. Maybe there could be a match. I wan
25 matches
Mail list logo
