What type of hardware substrate do you envision for the first AGI?
Many scenarios seem to envision AGI emerging in a particular computation
system, and/or perhaps in a robot or robots. Many scenarios involve the
system being connected to the Internet, as a crucial aspect.
What I tend to envision
On Sat, Jun 16, 2018 at 9:54 PM, Nanograte Knowledge Technologies via AGI <
agi@agi.topicbox.com> wrote:
> Stepping back for a moment to be able to see better what is being said.
> It's a jumble out there.
>
> Any computational solution would require a technology stack for every
> purpose to be ac
Stepping back for a moment to be able to see better what is being said. It's a
jumble out there.
Any computational solution would require a technology stack for every purpose
to be achieved. Arguing over which technology fits best into an AGI stack,
seems unproductive at this stage. Technology
On Sat, Jun 16, 2018 at 8:12 PM, Ben Goertzel wrote:
> On Sun, Jun 17, 2018 at 2:41 AM, Matt Mahoney via AGI
> wrote:
> > No, Stefan Reich and Mark Nuzz are right. Blockchain has nothing to do
> > with AGI and doesn't help.
> >
> > Blockchain is a terrible way to implement a distributed data sto
On Sun, Jun 17, 2018 at 2:41 AM, Matt Mahoney via AGI
wrote:
> No, Stefan Reich and Mark Nuzz are right. Blockchain has nothing to do
> with AGI and doesn't help.
>
> Blockchain is a terrible way to implement a distributed data store. It
> is not scalable. Every node stores a copy of every transac
CMR is built on top of a transport layer, which can be HTTP, SMTP, or even
carrier pigeon. The graph is a hierarchical ontology with redundant links,
so message loss is irrelevant. It is designed to work even with malicious
nodes injecting spam or forged messages. It requires nodes to be
intelligen
A voice of reason.
Matt Mahoney via AGI schrieb am Sa., 16. Juni 2018
20:42:
> No, Stefan Reich and Mark Nuzz are right. Blockchain has nothing to do
> with AGI and doesn't help.
>
> Blockchain is a terrible way to implement a distributed data store. It
> is not scalable. Every node stores a co
OK CMR - unless your graph is merely a tree or a giant DAG you need to ack the
ack's or have session state or the message flow will be mainly unidirectional
when server nodes are busy/congested.
And/or re-transmits or your'e constantly going to be having lost messages and
cleaning up orphan edg
OK forget the word "blockchain" I see it upsets people who didn't buy in early
:)
Use hashgraph or .. tangle possibly (doesn't look good) or some new DLT graph
which gets invented here :)
I'm going to re-read Matt's CMR paper again, to see if it may work for
something like this.
I don't like
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA256
On 06/14/18 12:13, MP via AGI wrote:
> They’ve done demos for Intel in the past IIRC. But the secrecy (and
> yes, I’m aware, as irritating as it is) resides in that they
> haven’t patented it yet, and are afraid of secrets being stolen.
that's a l
It's just math. https://en.wikipedia.org/wiki/Central_limit_theorem
When you add random variables, you add their means and variances and
the sum tends to a Gaussian curve. When you multiply instead of add,
the same thing happens when you take the log of the distributions.
On Fri, Jun 15, 2018 at 1
No, Stefan Reich and Mark Nuzz are right. Blockchain has nothing to do
with AGI and doesn't help.
Blockchain is a terrible way to implement a distributed data store. It
is not scalable. Every node stores a copy of every transaction. It
takes O(n^2) space. My AGI proposal ( http://mattmahoney.net/a
Looking into this further, probably what is needed is an open source hashgraph
since it would require heavy DAG modification. Hedera's hasgraph is proprietary
:(
Hedera claims hundreds of thousands of TPS...
John
--
Artificial General Intelligence List:
Stefan,
Wrong! It's a paradigmatic shift. It adds something missing from AI/AGI
yearned for all these years - its own economy.
And on the non-economic side, three main things - distributed decentralized
autonomous as well as data security.
It is revolutionary in terms of taking the power away
I don't think anyone in the world still uses Internet Explorer.
On 16 June 2018 at 14:44, A.T. Murray via AGI wrote:
> The tutorial AI Mind in JavaScript for MSIE at
>
> http://ai.neocities.org/AiMind.html
>
--
Artificial General Intelligence List: AGI
Pe
Please, no blockchain in AI... it really doesn't help anything.
Cheers,
Stefan
On 16 June 2018 at 14:45, wrote:
> Distributed decentralized autonomous AGI, temporal memory, lossy =
> forgetting even just dropping blocks, some memory on-chain some off.
> On-chain slower more secure, off-chain mo
Distributed decentralized autonomous AGI, temporal memory, lossy = forgetting
even just dropping blocks, some memory on-chain some off. On-chain slower more
secure, off-chain more local, sensory, less-processed. Blockchain meaning
blockchain-like, hashgraph-like, or whatever.
There would be som
The tutorial AI Mind in JavaScript for MSIE at
http://ai.neocities.org/AiMind.html
has an InFerence module which engages in automated reasoning.The human user
may enter a first statement as a premise, such as "boys play games". Later
on, the user may enter an Is-A statement such as "john is a boy
18 matches
Mail list logo
