Wednesday, February 29, 2012

TED Talk: The Case for Optimism and Abundance


Onstage at TED2012, Peter Diamandis makes a case for optimism -- that we'll invent, innovate and create ways to solve the challenges that loom over us. "I'm not saying we don't have our set of problems; we surely do. But ultimately, we knock them down."

22 Comments:

At 2/29/2012 10:46 PM, Blogger bix1951 said...

YES!

 
At 2/29/2012 10:56 PM, Blogger juandos said...

Ahhh, did anyone bother to talk to Uncle Ben Bernanke?

 
At 3/01/2012 5:03 AM, Blogger Ian Random said...

No one should be optimistic, we will all have to go through the 2038 date problem with computers that will cause the end of the world.

http://computer.howstuffworks.com/question75.htm

 
At 3/01/2012 10:26 AM, Blogger morganovich said...

i met him once at TED. he's an interesting guy, though a bit gormless in his own entrepreneurial ventures.

however, a couple of his key points are highly germane:

"We are living during an age of maximum personal impact. Today, individuals and small teams can accomplish what only large corporations and governments could once do. Exponential technologies and the tools of collaboration are allowing each of us to transform industries and address humanities grand challenges."

this is very true in a lot of spaces. it also explains a lot of wealth concentration. if a few can do so much, they reap outsized rewards.

"If you’re interested in driving and incentivizing breakthrough thinking within your company or industry; or if you’re interested getting experts around the world motivated to solve your insolvable problems, you need to create the proper incentives and establish clear, measurable and objective goals."

he seems highly fixated on outer space and i'm not entirely sure how useful space tourism is, but his X prize certainly seems to have driven some interesting innovation, though it has struck me that the size of the prize was quite small relative to the costs of chasing it. it seemed to be more about chasing glory (like an olympic medal) than economics.

i'm not sure how many times such a strategy would work, but it's good to see the private sector trying such things.

 
At 3/01/2012 1:54 PM, OpenID Sprewell said...

Huh, I was sure "gormless" must be some kind of typo but apparently it's a real word. I have a pretty expansive vocabulary and have never heard that one before, though I guess it's a briticism. Good recap of all the great stats of the last century in the video, though I wish there was more scoping out of upcoming innovation. I too wonder about these X-prize style tournaments, because if the net amount spent by all competitors is more than the prize, or worst case, if individual competitors are spending about what the prize gets them, they make little economic sense. So they may only make sense in limited niches, the tournaments of glory that morganovich references, but it's unlikely they'll break out to be more widespread. Which is fine, that's what the market is for, the ultimate tournament. :)

 
At 3/01/2012 2:27 PM, Blogger VangelV said...

LOL...

Let us begin with a few acknowledged problems by the speaker. He mentions climate crisis, species extinction, water and energy shortages.

Does any rational individual see a 'climate crisis?' Given the fact that we have seen a record high in agricultural yields, there is no crisis due to global warming, which is what he is talking about. How about species extinction? Well, if you look at the very small list of species having gone extinct you find that most became extinct more than 100 years ago and few species that have gone extinct recently.

I love the way that he deals with the energy issue. He claims that solar is cheap. Yet, the market tells us the opposite. Solar cannot survive without subsidies. Solar companies are going bankrupt every day and solar installations are idled all around the world as subsidies end.

It is because of naive optimism that we are in such trouble today. Yes, we are much better off than we were 50 years ago. But 50 years ago our access to more and more energy was going up. That is not the case any longer.

And let us look at one of the claimed positives and see how it affects reality going forward. Yes, we will have new more educated voices connected via the internet. What will they bring? Well, they will make some things that can help us. But one of the things that they will do is rise against corrupt governments that bind them and rule without their consent.

Like Mark and this guy I expect things to get better eventually. But along the way empowered people will rise violently against their oppressors and we will see plenty of bloodshed, poverty, and misery before freedom triumphs and people are finally freed to do as they wish.

 
At 3/01/2012 2:50 PM, Blogger morganovich said...

vangel-

that exactly why i stopped going to TED.

the preposterous, unsupported ideas there and wild techno optimism is cute once or twice, but then they start to sounds like brainwashed 8 year olds.

the amount of bad data about climate, alt energy, imminent technologies etc at TED has gotten absolutely stunning.

as an event, it has really lost its way has become, gormless! like many of it's participants.

most of these guys are serially failed entrepreneurs barracking about how great their philosophies and ideas are, but mostly seem unable to put them into practice.

a dazzling number of the talks seem to have no substance at all and just lurch from buzzword to buzzword.

TED was once filled with movers and shakers, not it has been overrun by the professional pontificators and become an echo chamber of groupthink and platitude.

 
At 3/01/2012 8:18 PM, Blogger VangelV said...

a dazzling number of the talks seem to have no substance at all and just lurch from buzzword to buzzword.

TED was once filled with movers and shakers, not it has been overrun by the professional pontificators and become an echo chamber of groupthink and platitude.


I could not agree more. Some of the older TED talks had some very brilliant speakers with a few scatterbrains sprinkled in here and there. Now we see the opposite. Most seem to be charlatans who make unsupported statements while very few are sound thinkers.

 
At 3/01/2012 9:47 PM, Blogger Hydra said...

Jesus, vange. Yes, rational people do see problems that you do not. Their horizon may be different from yours.

Yes, animals are going extinct. Their habitat is reduced, and climate change is moving what habitats there are.

If you choose to sneer at Gaians that worship mother earth, then you ought to at least offer some examples of how growth and innovation will allow us to solve problems we cause for species that cannot solve them alone.

There will be technical solutions for all kinds of problems, but this says nothing about who, or what, can afford, or will survive the answers we come up with.

 
At 3/01/2012 10:58 PM, Blogger juandos said...

"...climate change is moving what habitats there are"...

Really hydra but what planet and is that planet in in this arm of the galaxy?

 
At 3/02/2012 1:42 AM, Blogger Ron H. said...

"If you choose to sneer at Gaians that worship mother earth, then you ought to at least offer some examples of how growth and innovation will allow us to solve problems we cause for species that cannot solve them alone."

Well, some species have learned to survive by offering themselves to hunters in Texas.

 
At 3/02/2012 11:16 PM, Blogger markbahner said...

"Good recap of all the great stats of the last century in the video, though I wish there was more scoping out of upcoming innovation."

A key aspect missed by this talk is that the effective human population of the earth is about to explode.

I took Ray Kurzweil's estimates the amount of computer power (in instructions per second) per $1000, and historical and projected estimates of the number of personal computers sold in the world.

I compared that to the computing power of a human brain, which I took as 20 quadrillion operations per second (20 petaflops).

"Why economic growth will be spectacular"

From the graph, only one human brain equivalent (HBE) was added in 1996. And even in 2012, only about 100,000 HBEs will be added. But in 2024, 1 billion HBEs will be added. And in 2032, 1 TRILLION human brain equivalents will added. (In that single year!)

Arnold Kling and I predicted that spectacular economic growth would result from computer intelligence:

"Third thoughts on economic growth the the 21st century"

But I think even our estimates were low. I now expect world per-capita GDP growth rates above 10 percent per year even by the 2030s.

 
At 3/03/2012 7:58 AM, Blogger VangelV said...

Jesus, vange. Yes, rational people do see problems that you do not. Their horizon may be different from yours.

But they are not exactly rational when they make assumptions that are not supported by logic and claims that are not supported by data.

Yes, animals are going extinct. Their habitat is reduced, and climate change is moving what habitats there are.

They are? Fine. Provide us with a list of species that have gone extinct. If you bothered to look you find that the extinction rate was highest several hundred years ago when rats, cats, dogs, and other animals were brought to small islands on which species had no natural defence against them. Assuming that the planet is a small island and that island models are applicable is not rational.

If you choose to sneer at Gaians that worship mother earth, then you ought to at least offer some examples of how growth and innovation will allow us to solve problems we cause for species that cannot solve them alone.

I asked you first. What species? It is fine to claim an accelerating extinction rate but if you do you better back it up with evidence. Since you have none yours is a belief based position that has nothing to do with science.

 
At 3/03/2012 2:02 PM, Blogger Ron H. said...

"If you choose to sneer at Gaians that worship mother earth, then you ought to at least offer some examples of how growth and innovation will allow us to solve problems we cause for species that cannot solve them alone."

Do you mean these people?

 
At 3/03/2012 4:56 PM, Blogger Ron H. said...

I've always found this "accelerating extinction rate" argument to be rather specious. While it's true that humans have caused large numbers of extinctions, most notably large, very visible species, creative destruction is the norm, and has occurred for as long as there have been living things on Earth.

Previously unknown species are discovered on a fairly regular basis, so it's hard to say, with any accuracy, how many exist now, or how many have existed in the past.

With so little knowledge, it's silly to suggest that there is a "normal" rate of extinction, and that the current rate, although we can't even measure it, is higher than normal.

 
At 3/03/2012 7:27 PM, OpenID Sprewell said...

Mark Bahner, while it is true that computing will revolutionize the coming decades, your numbers are overly optimistic. There are fundamental physical limits to how small we can make transistors, which we're now reaching. Atoms are measured in tenths of nanometers and if you look at a chart of semiconductor scaling for NAND flash memory, you see that we're now making them at 15 nm feature sizes, so we're getting close to that limit. Extrapolating further scaling at the current rate, which is not a given considering all the problems that crop up as you get that small, means we'll hit the atomic barrier by the end of this decade. So you cannot simply extrapolate Moore's law out for decades, because it's very likely to end in a decade. Of course, we could always just go quad-core or octo-core and so on, building bigger rather than smaller, but that will reach a limit also: I don't think most will want to have refrigerator-sized computers in their homes. ;) So just as the predictions in your charts explode, those physical limits kick in and dampen your unrealistic estimates.

But the bigger issue that you're missing is that our software is nowhere close to reaching artificial intelligence, even with all that hardware at our disposal. The best we can do currently is domain-specific "intelligence," such as a computer that beats Kasparov at chess or dominates the Jeopardy champs. Try to get that same computer to do customer service with random customers at a bank desk and it fails regularly. Even with our vast success with mechanization you see the same problem: we can build machines that do very specific actions, like welding new cars, but those same machines cannot ever do a fraction of the general physical routines that you can put a person on, like preparing food or sewing clothes, leaving aside the mental training aspect. Robotics, where we create equivalents for the human hand and walking, is the next big step, before we could ever approach AI, assuming AI is even possible. We will be able to do great things with the augmentation of robotics and domain-specific computing, but that growth rate will not hit the heights that you are predicting.

Ron, haha, that's some funny shit in that video. :)

 
At 3/04/2012 12:03 AM, Blogger markbahner said...

"There are fundamental physical limits to how small we can make transistors, which we're now reaching."

I don't see that as a completely insurmountable problem. The human brain is 3-dimensional, not 2-dimensional.

"Of course, we could always just go quad-core or octo-core and so on, building bigger rather than smaller, but that will reach a limit also: I don't think most will want to have refrigerator-sized computers in their homes."

Maybe not refrigerator-sized, but I sure wouldn't mind a robot that weighed a couple hundred pounds, of which 50 pounds was brain.

"But the bigger issue that you're missing is that our software is nowhere close to reaching artificial intelligence, even with all that hardware at our disposal."

I think Ken Jennings and Brad Rutter would beg to differ. As would many of the of the owners of an Apple iPhone 4s.

"The best we can do currently is domain-specific "intelligence," such as a computer that beats Kasparov at chess or dominates the Jeopardy champs. Try to get that same computer to do customer service with random customers at a bank desk and it fails regularly."

Yes, intelligence is currently domain-specific. But the number of domains are increasing all the time (e.g., the Apple iPhone 4s).

And computers are currently too expensive to compete with tellers. But I can easily see the day when computers are far, far better than human tellers.

I think the last time I actually dealt with a human teller was more than a year ago. (Which tends to show that human tellers aren't all that big a deal already.) But as I recall, if I presented them with a signed check, my bank's tellers would give me cash...without checking my driver's license. I can easily see how a computer teller could not only check all the driver's licenses of their customers to check if I was the person on the check, but the computer teller could actually check the handwriting of the check against my handwritten signature for when I opened the account.

Creationists have long complained that evolution could not be true, because this or that missing link had never been discovered. Then when the missing link is discovered, they switch to another missing link.

The same basic thing is happening with people who say computers will never have intelligence equivalent to humans, or that such intelligence is impossibly far in the future.

Experts like Ray Kurzweil and Hans Moravec disagree (although Kurzweil's predictions see human-equivalent intelligence happening slightly sooner tha Moravec's). But assuming that a computers (costing several thousand dollars) become roughly comparable to human intelligence in the next few decades, economic growth will skyrocket.

As Julian Simon correctly noted, human brains are the ultimate source of economic growth. I'll stand by my predictions.

P.S. If you know any economists who have made long-term predictions of world per-capita GDP who are willing to challenge my predictions, please let me know.

Long bets prediction #194: world per-capita GDP will exceed $10 million by 2100

 
At 3/04/2012 7:58 AM, OpenID Sprewell said...

Mark, "I don't see that as a completely insurmountable problem. The human brain is 3-dimensional, not 2-dimensional."

wtf, you're joking, right? This response is so out of left field that it is hilarious. :)

"Maybe not refrigerator-sized, but I sure wouldn't mind a robot that weighed a couple hundred pounds, of which 50 pounds was brain."

I wasn't talking about robots, just how unwieldy it might get to deliver the level of computing capacity you're talking about, even in a box.

Sprewell -"But the bigger issue that you're missing is that our software is nowhere close to reaching artificial intelligence, even with all that hardware at our disposal."

Mark - "I think Ken Jennings and Brad Rutter would beg to differ. As would many of the of the owners of an Apple iPhone 4s."

No, I don't think any of them are confused enough to think those computers are artificially intelligent, as you apparently do.

"Yes, intelligence is currently domain-specific. But the number of domains are increasing all the time (e.g., the Apple iPhone 4s)."

While I hear Siri is nice and has millions of users, that is still practically nothing compared to US or world population, ie penetration is still practically nil.

"I think the last time I actually dealt with a human teller was more than a year ago. (Which tends to show that human tellers aren't all that big a deal already.)"

I think I've maybe dealt with a teller a couple times in the last decade: you can always constrain a problem to make it so simple that the computer can handle those simple parts. And that's great, I use an online-only checking account and don't have to deal with tellers almost at all. But when something goes wrong with my account, I still have to talk with someone over the phone, because the computer can't handle anything outside it's simplified domain.

"But as I recall, if I presented them with a signed check, my bank's tellers would give me cash...without checking my driver's license. I can easily see how a computer teller could not only check all the driver's licenses of their customers to check if I was the person on the check, but the computer teller could actually check the handwriting of the check against my handwritten signature for when I opened the account."

The notion that we'll still be using drivers' licenses or handwriting or cash at that point is what's crazy about your example. ;)

 
At 3/04/2012 7:59 AM, OpenID Sprewell said...

"Creationists have long complained that evolution could not be true, because this or that missing link had never been discovered. Then when the missing link is discovered, they switch to another missing link."

Completely false and unrelated example, because the evolution debate is about historical artifacts and the past, while AI is a debate about the future.

"The same basic thing is happening with people who say computers will never have intelligence equivalent to humans, or that such intelligence is impossibly far in the future."

I see no connection between a debate about what happened millenia ago to one about what might happen way in the future.

"Experts like Ray Kurzweil and Hans Moravec disagree (although Kurzweil's predictions see human-equivalent intelligence happening slightly sooner tha Moravec's). But assuming that a computers (costing several thousand dollars) become roughly comparable to human intelligence in the next few decades, economic growth will skyrocket."

"Experts" about the future are even more shitty and error-prone than other experts. Assuming that a computer could even "become roughly comparable to human intelligence," which I don't grant, at that point all bets are off, as it could go Skynet on us. It's sort of like predicting the nuclear reaction in the '30s and assuming that they will only be used to produce electricity, not for gigantic bombs. An AI might go in either direction.

"As Julian Simon correctly noted, human brains are the ultimate source of economic growth. I'll stand by my predictions."

Nobody denied the value of the human brain, only your overly optimistic predictions of if and how soon computers can mimic them.

"P.S. If you know any economists who have made long-term predictions of world per-capita GDP who are willing to challenge my predictions, please let me know."

I'm pretty sure almost no economist would agree with your predictions, but then, they probably wouldn't care enough to challenge you either. ;)

 
At 3/05/2012 12:21 PM, Blogger markbahner said...

Mark, "I don't see that as a completely insurmountable problem. The human brain is 3-dimensional, not 2-dimensional."

wtf, you're joking, right?

No, I'm not joking. And neither is Intel, IBM, Ziptronix, nor a host of other companies working in this field.

"This response is so out of left field that it is hilarious. :)"

Perhaps you should have simply searched for "three dimensional integrated circuit". Wikipedia has a decent article on the subject. The references in the Wikipedia article also point in good directions, such as:

"3-D Chips: IBM Moves Moore's Law Into The Third Dimension

This is hardly "so far out in left field it is hilarious"...as the ScienceDaily headline shows.

In fact, research as UCSB is looking at combining the memristors you mentioned into 3-D circuits:

"3-dimensional Circuits with Novel Nano-memristor Technology"

Were you not aware of the existence of 3-D integrated circuits?

 
At 3/05/2012 12:39 PM, Blogger markbahner said...

I wrote, "P.S. If you know any economists who have made long-term predictions of world per-capita GDP who are willing to challenge my predictions, please let me know."

Sprewell responded, "I'm pretty sure almost no economist would agree with your predictions, but then, they probably wouldn't care enough to challenge you either. ;)"

If economists "wouldn't care enough" then they're pretty ignorant. If I'm right, it's more important than anything that has ever happened in economics.

But tell you what, Sprewell. You obviously think of yourself as pretty hot stuff. So why don't you give us *your* predictions for world per-capita GDP growth rate (adjusted for inflation) from:

2010-2020?
2020-2030?
2030-2040?
2040-2050?
2050-2060?
2060-2070?
2070-2080?
2080-2090?
2090-2100?

 
At 3/05/2012 11:05 PM, OpenID Sprewell said...

Mark, I'm well aware that there are many firms thinking about or "working" on 3-d stacking of transistor planes. I'm also aware that nobody has gotten it to work and that there are significant heat issues with such stacking, which is why it is not going to happen anytime soon, barring some unforeseen breakthrough. Just because something could theoretically exist doesn't mean it will happen, or we'd all be driving nuclear-powered cars by now. ;) It is pretty self-important of you to claim economists are ignorant just because they don't care about you and your claims. Since it is almost impossible to predict these numbers, what you think they're going to be is almost irrelevant, particularly since you're unable to back your predictions up with solid arguments against the roadblocks I've listed.

As for my own prediction, I have none, because these phenomena are incredibly complex. Also, I'm not a big fan of aggregate measures like GDP, which don't take subjective and incredibly dynamic personal desires into account, particularly since I think the great trend of this century is going to be away from commoditization towards personalization. You'll be able to buy a cereal made just for you, every restaurant will make your burger exactly the way you like it, and your news feed will be customized just for you. Trying to aggregate such customized services will make less and less sense over time. But if you really want a prediction anyway, I'll say growth rates will hit anywhere between 5 and 15% by the end of the century, with a 5% chance of catastrophe, ie GDP falling by 50% or more in any single decade along the way. It is silly to presume to predict something so fundamentally unknowable, but one can set some reasonable bounds and your overly optimistic predictions surpass where I set those bounds.

 

Post a Comment

Links to this post:

Create a Link

<< Home