Super Artificial Intelligence, a naive approach

[hecd2]

You may have missed the orders of magnitude Wikipedia URL, which gives minimum of 10^15.

10^15 synapses not synaptic ops/sec

 

[hecd2]

Of the two wiki links you so ham-fistedly provided, only one mentions 10¹⁵, this one...

https://en.wikipedia.org/wiki/Orders_of_magnitude_(numbers)

And it's entry reads...


Well, duh.

Uh, no:

1015

(1000000000000000; 10005; short scale: one quadrillion; long scale: one thousand billion, or one billiard)
ISO: peta- (P)

• Biology – approximately 1015 synapses in the human brain [20]

 

[abaddon]

IBM's Dharmendra Modha would probably disagree.

His statement: "Before the end of 2020 we will be able to produce a brain in a box".
https://m.youtube.com/watch?v=UHPwIBiFVq0

Nope. Not playing a game of chase the youtube for your amusement.

If you have a point then you should be able to type it right here in your own words instead of trying to get people to insert both feet into one trouser leg trying to figure out what it might be that you might perhaps mean.

 

[Aridas]

You may have missed the orders of magnitude Wikipedia URL, which gives minimum of 10^15.

No, it doesn't. Bear in mind that synapses are very much not transistors. Similar in some ways, sure. Not the same, though, by a long shot.

 

[ProgrammingGodJordan]

I tire of this discourse.

Edited by Agatha: 

Edited to remove page-stretchy picture which was already posted 6 posts before this one. Do not spam or use disruptive formatting.

Take the words of IBM's Dharmemdra Modha, that deals with neuromophic chips.

His statement: "Before the end of 2020 we will be able to produce a brain in a box".
https://m.youtube.com/watch?v=UHPwIBiFVq0

 

[Aridas]

IBM's Dharmendra Modha would probably disagree.

His statement: "Before the end of 2020 we will be able to produce a brain in a box".
https://m.youtube.com/watch?v=UHPwIBiFVq0

Little problem. He looks like he's referring to "brain-inspired computing" rather than hitting the 10^18 area. It's like a brain, not that it can necessarily reach the same power as a human brain by 2020.

 

[Porpoise of Life]

Criticism is welcome/needed.

At least the 'welcome' part of that isn't true.

 

[ProgrammingGodJordan]

Little problem. He looks like he's referring to "brain-inspired computing" rather than hitting the 10^18 area. It's like a brain, not that it can necessarily reach the same power as a human brain by 2020.

Brain might be less than 10^18.

Keep in mind they publicized 10^14 a little while back.

Key thing here is they don't say consciousness by 2020, merely human brain power.

PS: brain inspired computing means a non trivial degree of human level brain power, according to modha.

 

[ProgrammingGodJordan]

At least the 'welcome' part of that isn't true.

Sensible criticism is welcome.

Anyway, I am literally tired, just left gym.

 

[abaddon]

I tire of this discourse.

Of course you do. Having appealed for criticism, you don't much like it when you get it. Which begs the question. Why did you invite it?

Take the words of IBM's Dharmemdra Modha, that deals with neuromophic chips.

His statement: "Before the end of 2020 we will be able to produce a brain in a box".
https://m.youtube.com/watch?v=UHPwIBiFVq0

Spamming the same link will not help your case. Cherry picking from that link will not help your case. Ignoring context will not help your case.

 

[Porpoise of Life]

Sensible criticism is welcome.

Anyway, I am literally tired, just left gym.

"number of synapses does not equal number of operations per second" and "Your new claim is smaller by factor 10, compared to the lowest limit of your earlier claims" are sensible criticism.

What kind of comments were you hoping for?

 

[ProgrammingGodJordan]

Of course you do. Having appealed for criticism, you don't much like it when you get it. Which begs the question. Why did you invite it?

Spamming the same link will not help your case. Cherry picking from that link will not help your case. Ignoring context will not help your case.

The context provided in the original post was human level brain power by at least 2020.

The context provided in dharmendra modha's video was, human level brain power by 2020.

So, the link does not "ignore the context".

 

[abaddon]

Uh, no:

I stand corrected.

 

[ProgrammingGodJordan]

"number of synapses does not equal number of operations per second" and "Your new claim is smaller by factor 10, compared to the lowest limit of your earlier claims" are sensible criticism.

What kind of comments were you hoping for?

(1)
See IBM data, or other data on web.

I have not the will power to repeat myself.

Either way, computational neuroscientist Dharmendra Modha from ibm expresses that human brain in a box (from ibm) will arrive by 2020.


(2)
PS:
I predicted sensible comments from a few beings here. (1-3 persons).
Otherwise I predicted garbage.
My predictions appear to be accurate thus far.

 

[Aridas]

Brain might be less than 10^18.

Sure. More specifically, it's fairly likely that the brain isn't perfectly efficient. It even may well be possible that present day computers, with sufficiently good code running, could outperform a human brain in pretty much every way that matters. With that said, though, you haven't backed up 10^15 or 10^16 validly. Nor did you actually address the point I had just made, which was that you seem to fairly certainly be engaging in a convenient misinterpretation of what Modha was saying and trying to twist it to seem like it supports the the points of yours that are actually in dispute, rather than points that aren't so much in dispute.

When it comes to the 10^18 number by 2020, regardless, I'm a little more surprised that you haven't been pointing towards China's project mentioned on the exascale wiki page which is apparently planned to be finished by 2020. That, at least, would be something something more solid for you to point towards, after all, rather than trying to grasp at straws like you've been doing.

Keep in mind they publicized 10^14 a little while back.

Key thing here is they don't say consciousness by 2020, just human brain power.

Which still begs the questions of which measurements are being employed and why you are even bringing up consciousness.

PS: brain inspired computing means a non trivial degree of human level brain power, according to modha.

It seemed much more like brain inspired computing had a heck of a lot more to do with "how" the system works, given that it seems to be modeled on how brains work, rather than actually saying much about the raw power. Raw power might be potentially be reachable with such a model, but is, at best, a secondary concern there for why it's brain inspired computing.

 

[abaddon]

The quote in context

And the ultimate vision is, which I believe will be possible, before 2020 ends, is that we will be able to produce a brain in a box, which was the original vision of the Synapse Project. Ten billion neurons into {unintelligible} box. This is no longer science fiction, this is actually happening

So first, this occurs at 19:38 in the video, but PGJ wont tell you that.

Second, it is an aspiration and nothing more.

Third, the number he aspires to is 10¹⁰.

 

[fagin]

(1)
See IBM data, or other data on web.

I have not the will power to repeat myself.

Either way, computational neuroscientist Dharmendra Modha from ibm expresses that human brain in a box (from ibm) will arrive by 2020.


(2)
PS:
I predicted sensible comments from a few beings here. (1-3 persons).
Otherwise I predicted garbage.
My predictions appear to be accurate thus far.

So you believe it will be one or the other?

 

[theprestige]

(A)
The calculation with minimum brain speed (instead of maximum) yields 3 years roughly.

Keep in mind the original post expressed AT LEAST 2020.
HINT: Replace 10^18 with 10^15 roughly.

10^18 is *three orders of magnitude* larger than 10^15. They are not at all "roughly" the same value. They are quite clearly two extremely different values, one of which is substantially smaller than the other. So much smaller that no useful calculations can be made by substituting the one for the other.

15 and 18 are roughly the same. Even 1500 and 1800 are roughly the same for calculations to within an order of magnitude. 10^15 and 10^18 are not roughly the same in the context of your inquiries.

 

[MikeG]

Ehh. No need to react like that, regardless. If you think that he's behaving badly, alright. He's done that already. Why descend to the level that you think that he's on, though? It gains you nothing and certainly doesn't help show that you are any better.

Fair comment. I'll try harder.

However, I warrant you haven't been dealing with this crap for the last couple of weeks. It does try one's patience eventually.

 

[ProgrammingGodJordan]

The quote in context



So first, this occurs at 19:38 in the video, but PGJ wont tell you that.

Second, it is an aspiration and nothing more.

Third, the number he aspires to is 10¹⁰.

That's odd.

I gathered that they had already achieved 10^14 in a relatively recent paper:

http://www.modha.org/blog/SC12/RJ10...122214894244438&cm_mc_sid_50200000=1489431888