• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Super Artificial Intelligence, a naive approach

MikeG said:
Maybe you could get your 3 year old nephew to carefully explain to you that if this is the figure for the lower bound of the human brain, then you shouldn't have used "10^16 to 10^18" in all your claims prior to this thread.

Maybe your three year old nephew could also explain to you that 10^18 is a thousand times higher a number than 10^15, and that therefore it is inconceivable that human brains vary between your upper and lower bounds. One or more of these figures is made up.

Your 3 year old relative might also help you understand that 10^15 is not "roughly" 10^16 as you claimed. The former is only 10% of the latter.
Click to expand...



(A)
WHY DID I WRITE 10^16 SOPS?

At 10 impulses per second, and at a guess of 10^15 synapses, we have a total of 10^16 synaptic operations per second.

So, 10^15 synapses corresponds to the source in original post, which yields 10^16 sops.



(B)
The lower bound 10^16 sops =10^15 synapses.
And so 2020 was computed using 10^15 synapses (aka 10^16 sops)



PS:
10^15 synapses was rough value. The actual figure was some small value x 10^15.


[IMGw=640]http://i.imgur.com/1jiSyAv.jpg[/IMGw]
 
Last edited:
ProgrammingGodJordan said:
.........At 10 impulses per second, and at a guess of 10^15 synapses, we have a total of 10^16 synaptic operations per second........
Click to expand...

Obviously 10 times 10^15 = 10^16.

Now, justify both figures. Where does that 10 SOPS figure come from? Where does 10^15 come from, and why were you using a different figure (10^16 to 10^18) previously?
 
Aridas said:
More specifically, the 10^16 to 10^18 refers to something completely different than the 10^15. Saying that the 10^15 is roughly the 10^16 is utterly fallacious from the get go because they're entirely different things, before the orders of magnitude issue can even come into play.
Click to expand...

Yeah, I've asked about the units once or twice. They're still not clearly laid out anywhere.
 
MikeG said:
Obviously 10 times 10^15 = 10^16.

Now, justify both figures. Where does that 10 SOPS figure come from? Where does 10^15 come from, and why were you using a different figure (10^16 to 10^18) previously?
Click to expand...

10^15 synapses is given in the original source. (Koch Christof, biophysics of computation 2004)

The 10^15 corresponds to the 2020 year computation, and it also corresponds with 10^16 sops.

So, 10^15 synapses = 10^16 sops, all in the original post.

In other words, I am not using "different figures".
 
Last edited:
ProgrammingGodJordan said:
10^15 synapses is given in the original source. (Koch Christof, biophysics of computation 2004)

The 10^15 corresponds to the 2020 year computation, and it also corresponds with 10^16 sops.

So, 10^15 synapses = 10^16 sops, all in the original post.

In other words, I am not using "different fugues".
Click to expand...

Where does the 10 SOPS/ second figure come from?
 
Where does the figure 10 come from? Have you just made it up? How do you know there are 10, rather than 8.8 or 127 or whatever? Is it even a valid way of measuring brain activity?
 
abaddon said:
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.
Click to expand...

Wrong.
10^10 neurons, not 10^10 synapses.

The 10^10 is closer to number of neurons in human brain. (See the same source provided in original post)
 
Last edited:
Aridas said:
You still have yet to give any reason to accept that 10^15 or 10^16 were valid to use in the first place, given that your 10^15 referred to something notably different than what you claimed and your 10^16 still appears to be pulled out of thin air, given that nothing in either of those links supports it at all.





Again, what is your lower bound actually based on at this point, if not you trying to cover up your mistakes? You've utterly failed to support it meaningfully.
Click to expand...

Ironically, 10^15 was not "notably different" than 10^16 synaptic ops per sec.

At 10 impulses per second, 10^15 synapses = 10^16 synaptic ops per sec. (Which was long mentioned in original post)
 
Last edited:
Aridas said:
ETA on this part: Incidentally, lying about how you actually said that in the original post rather than talking about it being at least 10^15 in computational size isn't a point in your favor.

.



You'd have much less of a hard time if you just admitted your error and the conversation just moved on from there. Even if your conclusion is correct for other reasons, when one invokes bad logic, one should be prepared to be called out on their bad logic. Simply running away in the face of being completely unable to defend your claim isn't really a good thing, regardless.
Click to expand...



The following response applies:


ProgrammingGodJordan said:
Ironically, 10^15 was not "notably different" than 10^16 synaptic ops per sec.

At 10 impulses per second, 10^15 synapses = 10^16 synaptic ops per sec.
Click to expand...
 
Porpoise of Life said:
So, where did you get 10 impulses per second from? I can't find it in any of your sources.
Click to expand...

I had 10^16 sops in original post, and used 10^15 synapses (which corresponds to 10^16 sops) to get year 2020, as minimum year for which 10^16 artificial sops (or 10^15 artificial synapses) shall probably be achieved.
 
Last edited:
abaddon said:
Nope. 1015 is well outside your claimed range of 1016 to 1018.
Furthermore, since you are suggesting an AI at least equal to human level, we must perforce use the maximum figure. Otherwise, your suggested 1015 figure is one thousandth what humans are capable of.

I intentionally quoted your numbers from the OP so that you could not dodge.
Click to expand...

The following better disregards your "10^15 is well outside of your claimed range of 10^16 etc" accusation:


ProgrammingGodJordan said:
Ironically, 10^15 was not "notably different" than 10^16 synaptic ops per sec.

At 10 impulses per second, 10^15 synapses = 10^16 synaptic ops per sec.
Click to expand...
 
ProgrammingGodJordan said:
Ironically, 10^15 was not "notably different" than 10^16 synaptic ops per sec.
Click to expand...

Please don't consider a career in any profession where missing by a factor of 10 is considered important. Heck, I'm scratching my head trying to think of any career in which 1 of something isn't "notably different" to 10 of something.

I think even suggesting this is an indication that what you're doing isn't working.
 
ProgrammingGodJordan said:
Small correction: Unity 3D was used in item iii.

For typical neural network code, see some other items, like items ii, iv or v:
Click to expand...

I had a look but they 404. The other repositories I looked at are equally trivial apart from a potentially interesting javascript learning library you forked from elsewhere. Yeah, some of us know git, java, etc. Nothing to see here folks, move along.
 
ProgrammingGodJordan said:
I had 10^16 sops in original post, and used 10^15 synapses (which corresponds to 10^16 sops) to get year 2020, as minimum year for which 10^16 artificial sops (or 10^15 artificial synapses) shall probably be achieved.
Click to expand...

So you derived 1016 by starting from 1015 and multiplying by 10 operations per second, and you got 10 operations per second because you started with 1016 and needed to derive 1015? This doesn't strike you as at all circular?
 
You must log in or register to reply here.

Back
Top Bottom