Human versus AI; can we still compete?

[Deleted]

Synical much Virg?


So, as in the Matrix, do you think the human race has peaked?

[bimetalaupt]

Jul 26, 2011 7:37:53 GMT -5 ArchietheDragon said:

Synical much Virg?


So, as in the Matrix, do you think the human race has peaked?

Archie,
With all the Black swans and risk on /risk off trading we have seen a huge reduction in the return from the "Expert system" from over 20% return per year (after cost) to about 8% (7.89577%) Return..
also the correlation numbers from 87% down to 27.67373101%...

Gold to m2 is the only correlation holding up..95.439278400568114%
Just the numbers,
Bruce

Attachments:

[Virgil Showlion]

So, as in the Matrix, do you think the human race has peaked?

In a limited number of respects, yes. In many, no. A lot will depend on the next decade.

I don't subscribe to Kurzweil's vision that humans will merge wholly with machines to become pure energy beings within the next 20 years. Frankly, the prediction is insane, and the timeframe is motivated more by Kurzweil's impending mortality than by any of the research he catalogs on his site. And I take exception to futurists throwing out "exponential growth" as an explanation for nonsensical predictions, since (as I say) the qualifier "exponential" is applied to a different concept every three months.

Gentlemen like Kurzweil often hurt researchers in the fields of computational intelligence, AI, etc., by critically oversimplifying the hurdles in getting even "simple", highly-task-specific intelligences to work properly, and by giving people grossly inflated expectations that undermine the modest advances that researchers are able to claw out one step at a time.

[yclept]

Virgil,
No doubt Dr. Kurzweil (and I grant that all the PhD degrees are honorary) does have some fringe ideas (I understand he's signed up for cryogenic preservation, and he's proposed sending nano-bots down into his father's grave to retrieve some DNA!), and he does take a lot of pills and such. Still, someone has to look ahead, and in doing so help to prepare the world for what's coming on whatever time-frame. I think his implementations of computer devices to aid handicapped people are worthy of praise. He missed by a few years on some predictions (I understand the defense department will have a 20 petaflop computer this year rather than 2009 as Kurzweil predicted). But I think it's wrong to ascribe all of his work to a desire for his personal immortality.

This month's National Geographic had an interesting article on some of the current robots being made with emphasis primarily on the physical bots rather than AI.

I still contend that the first AI that can write a better version of itself will let the cat out of the bag, and at that point improvement will be exponential.

If you had told the builders of the first transcontinental railroad in 1869 that 100 years later a man would walk on the moon, I'm sure they would have carted you off to the asylum and started trepanning, as they would surely have seen you as an otherwise lost cause. Technological advances in the 20th century were certainly exponential with more in the last 10 or 20 years than in the rest of the century.

I think it wiser to at least consider potential progress we are likely to see and to not expect it to come at the always glacial rate with which we are familiar. Rather that, than stand around in Ned Ludd's bare feet (the clogs anecdotally having been thrown into the loom).

[Trongersoll]

I doubt that you will ever see (in your lifetime) an AI that can write a better version of itself. Heck, I'd be amazed to see an AI "create" anything. There is a difference between learning and doing.

[yclept]

A problem solving robot that can query other robots through the internet to learn new tasks:
www.kurzweilai.net/are-you-ready-for-a-robot-that-learns-on-the-internet

[yclept]

I think I see more swan dives off the roof:

news.xinhuanet.com/english2010/china/2011-07/30/c_131018764.htm

Maybe Apple should think about moving some of this manufacture back home. They have plenty of money to buy any robots they would need. I read the other day they have more cash on hand than the US Treasury did.

[yclept]

See you all in class! Warning, the textbook is over a hundred bucks.

spectrum.ieee.org/automaton/robotics/artificial-intelligence/you-you-can-take-stanfords-intro-to-ai-course-next-quarter-for-free

[yclept]

This looks to me to be a massive step forward (and by IBM, which is an entity I don't think anyone would classify as kooky!). Here's a chip that won't be programmed, but rather will learn from experience and rearrange it's connections accordingly. No programmer desired or needed!

www.kurzweilai.net/ibm-unveils-cognitive-computing-chips-combining-digital-neurons-and-synapses?utm_source=KurzweilAI+Daily+Newsletter&utm_campaign=966ad22468-UA-946742-1&utm_medium=email

[yclept]

www.computerworlduk.com/news/networking/3302464/algorithmic-stock-trading-rapidly-replacing-humans-warns-government-paper/

[lovetobike]

I hope this means you are coming back to the boards YClept! You might enjoy participating in this blog: www.edstalkingstock.wordpress.com. He is just a guy who is sharing his investing strategies.

[yclept]

www.telegraph.co.uk/technology/news/8753784/The-300m-cable-that-will-save-traders-milliseconds.html

I'm not planning to post in any significant way. I wasted far too much time researching and posting, only to discover that work could be thrown away at another's whim. Assertions that deleted posts were restored were not true, and merely emphasized to me that these boards are a classic example of a cultural hegemony and thus not a societal organization I should voluntarily contribute to in good conscience.

[yclept]

www.bis.gov.uk/assets/bispartners/foresight/docs/computer-trading/11-1276-the-future-of-computer-trading-in-financial-markets.pdf

The government (UK) paper that was the basis of the story linked in post 70 above.

[The Virginian]

Personally I feel that IBM is the Tech Investment of the future. Sure Watson is not human - It is so much better. IBM recently signed a contract with a major health provider to have Watson give second opinions on Doctors diagnosis and tests ordered with the (for now) caveat if the Doctor and Watson don't agree it will still be reviewed the old fashion way.

This is only the tip of the Iceberg of potential uses for Watson. Watson will help business owners make decisions based on past history and future probabilities. I would love to have access to Watson for my business right now. In these hard economic times I could ask Watson for advice on how much Inventory I should order for upcoming seasons for instance.

IBM is a solid Investment for the long term (IMO) no matter what but Watson puts IBM out in front of any potential competitor for the future of shared computing!

[ModE98]

"This is only the tip of the Iceberg of potential uses for Watson."
Believe that a excellent prognostication. And other things yet to come out of the age of electronics.

[yclept]

The chip linked in reply 69 above is the biggest piece of news on this thread. It's a learning/self-programming chip -- no human programmer required.

I do think IBM, the company, is pushing some of the most advanced hardware applications forward and that will bode well for them for a few years. This will be a chip that can most likely design a better version of itself, and once that happens, the cat is out of the bag. Assuming the chip will be an item for resale, IBM will no longer have any exclusivity on further development. And they'll either have to sell it for incorporation into other people's machines, or they will have to develop a new business model wherein all they sell is its computational power, which, I think would very much limit the overall sales potential (many potential users will probably not want their business-specific data to be out of their hands). IBM will be damned if they do, and damned if they don't (sell it). In any case, other chip designers will eventually have similar architectures -- if nothing else, universities will push computationally parallel geometry designs and have them fabbed by other chip makers.
I don't know about IBM as a stock. I haven't looked that deeply into the fundamentals. A quick look makes it seem pretty fairly valued to me.

[The Virginian]

I agree, IBM's "Cognitive Chip" is just another example.

Many Investors still look at IBM as the Old Giant of Selective Typewriters and Data Processing. IBM is totally recolonizing the Computing Industry and I don't believe many are paying attention. This is not science fiction but major breakthroughs.

As for being a threat to humans? That depends on the humans of the future - maybe our grand kids or great grand kids but I would be doubtful of anything to worry about in the next fifty years. But who knows, as soon as chips can fend for themselves changes will come more quickly than we can imagine. Humans have to rest, play, and live..... the AI will be able to bypass all those and "think" 24 hours a day - 7 days a week.

[tyfighter3]

Our TEC has been advancing but we Humans have been de-evolving since Adam and Eve. Just think about all of the Bad Genes we have since then.

[Aman A.K.A. Ahamburger]

Sept 17, 2011 7:51:17 GMT -5 The Virginian said:

I agree, IBM's "Cognitive Chip" is just another example.

Many Investors still look at IBM as the Old Giant of Selective Typewriters and Data Processing. IBM is totally recolonizing the Computing Industry and I don't believe many are paying attention. This is not science fiction but major breakthroughs.

As for being a threat to humans? That depends on the humans of the future - maybe our grand kids or great grand kids but I would be doubtful of anything to worry about in the next fifty years. But who knows, as soon as chips can fend for themselves changes will come more quickly than we can imagine. Humans have to rest, play, and live..... the AI will be able to bypass all those and "think" 24 hours a day - 7 days a week.

I agree, all these advancements are going to help. Humans like to humanize things, names for boats, cars. She is a nice car, ect,ect.

The biggest thing to remember, who will robots be doing all of this for? We are programming them for our uses, and it still doesn't change the fact that they NEED power that we give them. They will be "evolving" with the "mind set" to make human lives better, it's like you said Mr V. they don't need water, food, praise, ect, ect, all the will want to do is think, and LOGICALLY I might add.

[yclept]

www.computerworld.com/s/article/9220044/Intel_runs_PC_on_CPU_powered_by_solar_cell?taxonomyId=64

[Aman A.K.A. Ahamburger]

This CPU, also called near-threshold voltage processor, may not become a commercial product, but the research could be integrated into future processors and other circuitry

I was saying it a while back.. IMO, Chip into brain, singularity. AI= Artificial implants, humans have Artificial intelligence, the next step in human evolution.

[The Virginian]

humans have Artificial intelligence, the next step in human evolution.

Considering what we have done with our "real" Intelligence maybe it won't be such a bad step. It could give people that "Logical" side many are missing.

I just want to make sure my implant is fully capable of receiving Upgrades!

What happens when you get a "Virus" ? Some European Criminal could have us doing all sorts of evil things!

[yclept]

www.kurzweilai.net/breakthrough-proton-based-chips-that-communicate-directly-with-living-things?utm_source=KurzweilAI+Daily+Newsletter&utm_campaign=4694bd72c7-UA-946742-1&utm_medium=email

The beginnings of a biological link.

[Aman A.K.A. Ahamburger]

Sept 20, 2011 7:31:44 GMT -5 The Virginian said:

humans have Artificial intelligence, the next step in human evolution.

Considering what we have done with our "real" Intelligence maybe it won't be such a bad step. It could give people that "Logical" side many are missing.

I just want to make sure my implant is fully capable of receiving Upgrades!

What happens when you get a "Virus" ? Some European Criminal could have us doing all sorts of evil things!

I hear you about the upgrades Mr. V. The questions you're asking are ones that will be answered as we see how things develop. How open people will be to it, will be another thing.

[yclept]

www.hpcwire.com/hpcwire/2011-09-27/on_wall_street,_the_race_to_zero_continues.html

[Aman A.K.A. Ahamburger]

It's funny, I realized the biggest point to this conversation was overlooked, while it was guiding this and every human interaction. Emotions. CPU, AI, machines, are only "humans" because we like to humanize things. The desire to take over stems from the emotional side of human nature.

The reason that CPUs have a slight advantage over us is because they only use logic. Like a Vulcan. Which brings me to what made me realize this. All the Star Trek movies are on Netflix, Data, an superior AI gets an emotion chip and is rendered useless basically. When you think about it, if we where to make robots "human" we would probably have a "punk" robot movement.

This position was re affermed today when I was watching Weird or What on history and this exact debate was on it. A very logical older scientist brought it up very plainly, as I find logical old men do. The question is why? CPUs have NO desires, and the reality is that WE, being humans, think that we have built these advance CPUs when in fact bacteria is smarter than a CPUs are right now. Bacteria can live and thrive on it's own, a CPU still needs a human.

Thanks to Ycelp for this discussion, I hope you read this and put your input in sir.

[Aman A.K.A. Ahamburger]

I was reading this article and immediately thought of this thread. A big part of this discussion is....

How do we govern our technology?
www.theglobeandmail.com/news/opinions/opinion/how-do-we-govern-our-technology/article2371917/

[sunuva]

Shadows of the Mind by Roger Penrose sets out to prove that consciousness is not algorithmic (quantum mechanics and all that). And since it isn't algorithmic it can't be "programmed." It is very simplistic the way I have described it and this is by no means definitive, but he lays a foundation that basically says AI won't surpass the human mind. I quite liked the book (but I read it way back in the nineties and haven't read anything else on the subject since).

[Aman A.K.A. Ahamburger]

Right now it took 16,000 supercomputers to "learn" what a picture of a cat was. Just in case we do face a robot uprising though...

HOWTO survive a robot uprising (just add water)
boingboing.net/2012/08/01/howto-survive-a-robot-uprising.html?utm_source=feedburner&utm_medium=feed&utm_campaign=Feed%3A+boingboing%2FiBag+%28Boing+Boing%29

[safeharbor37]

In regard to the competition between humans and computers: Consider gambling; Humans compete against Slot Machines [computers], Roulette Tables, and card games against the dealer who has a statistical advantage ~ and claim to win. The theoretical trick is to "leave when you're ahead."
To some degree, the strategy to use against computer trading is not to compete directly, but to take advantage of the quirks of the computer programmers or the computers themselves. That is; to avoid competing directly against the computer, but to understand its strategy and work around it.
Theoretically, it would be possible to devise computers to work every angle, but it isn't likely that will be done. To some degree, it is like the strategy used by retailers who find themselves in competition with Walmart. The smart thing to do is not compete, but complement Walmart. Likewise, it is not smart to compete directly with a device which can calculate several millions of times faster than you, but to cooperate with it to mutual advantage.