Tim O'Reilly: Don't Fear AI, Fear Ourselves (wired.com) 72
Tim O'Reilly, publisher of geeky books, "seizes on this singular moment in history" for a futuristic new book of his own, according to this interview with Steven Levy. An anonymous reader writes:
When it comes to artificial intelligence, O'Reilly sees a reason for optimism in the fact that we're already discussing biased algorithms. ("We had plenty of bias before but we couldn't see it.") O'Reilly ultimately believes AI won't take away our jobs, and even argues that we're defining it all wrong. "What we now call AI is just the next stage of us weaving our intelligence together into a greater whole. If you think about the internet as weaving all of us together, transmitting ideas, in some sense an AI might be the equivalent of a multi-cellular being and we're its microbiome, as opposed to the idea that an AI will be like the golem or the Frankenstein. If that's the case, the systems we are building today, like Google and Facebook and financial markets, are really more important than the fake ethics of worrying about some far future AI.
"We tend to be afraid of new technology and we tend to demonize it, but to me, you have to use it as an opportunity for introspection. Our fears ultimately should be of ourselves and other people."
O'Reilly calls financial markets "the first rogue AI," while also priasing innovators like Elon Musk and Jeff Bezos for moving humankind in new and positive directions. And he also calls Uber "a good metaphor for what's right and wrong in tech" because of its clashes with both its drivers and city governments.
"It's interesting that Lyft, which has been both more cooperative in general and better to drivers, is gaining share. That indicates there's a competitive advantage in doing it right, and you can only go so far being an ass."
"We tend to be afraid of new technology and we tend to demonize it, but to me, you have to use it as an opportunity for introspection. Our fears ultimately should be of ourselves and other people."
O'Reilly calls financial markets "the first rogue AI," while also priasing innovators like Elon Musk and Jeff Bezos for moving humankind in new and positive directions. And he also calls Uber "a good metaphor for what's right and wrong in tech" because of its clashes with both its drivers and city governments.
"It's interesting that Lyft, which has been both more cooperative in general and better to drivers, is gaining share. That indicates there's a competitive advantage in doing it right, and you can only go so far being an ass."
Limits (Score:5, Insightful)
you can only go so far being an ass
Well, you can get at least as far as President of the United States.
Re: (Score:2)
you can only go so far being an ass
Well, you can get at least as far as President of the United States.
I call bullshit! Being an ass isn't enough! What you really need antisocial personality disorder, a (YUGE!) dose of narcissism and hordes of unhappy fools that will believe your endless transparent lies. The last part is critical because votes matter... but only in the electoral college.
Different (Score:5, Interesting)
Why should it be different in this era of cognitive enhancement?
Because when robots can do everything (or nearly everything) a person can do, what is the prospect for employment, let alone increased employment? Before, people always moved on to jobs that machines couldn't do. First from farming to manufacturing, then to services. The final refuges will probably be software development, some retail, and a little bit of hands on medicine. Also top corporate leadership, not because computer programs won't be better than they are at their jobs but because they're the ones in charge so will continue paying themselves enormous salaries for something a computer could do for free. If sexbots are legal, prostitution will diminish greatly. Pretty much everything else will be automated. This isn't happening in 10 years, but it's happening.
Re: (Score:1)
That's the thing, that day will never arrive. It is absolutely astonishing to me how so many people seem to have become little more than sci-fi futurists, and the absolute horse sh*t that flys out of their mouths. Technology is useful, hyperbole and fantasy, not so much.
FTFY.
Re: (Score:3)
Either computers and robots will continue to get better and better until they're better at more or less everything than us, or they will at some point stop getting any better. Just at a permanent plateau with no improvement ever. You're telling me you think it's going to be option B? And then you're telling me I'm the one having fantasies?
Re: (Score:2)
I think it's a stretch to assume we can reach human-style fuzziness without losing computer-style exactness. For example if you take the concept "birds", one might say that birds fly. But ostriches and penguins don't fly. So the understanding of what a bird is will be fuzzy, thus a command that depends on the nature of birds will also be fuzzy. Or simply incomprehensible, like if you ask someone to explain why AlphaGo thinks this is a good move there's no "human" explanation just a lot of weights that says
Re: (Score:2)
Or simply incomprehensible, like if you ask someone to explain why AlphaGo thinks this is a good move there's no "human" explanation just a lot of weights that says the computer thinks it's a good idea
You mean like how the human brain works? People are often confused as to their motivations. At least AlphaGo is clear about what they are. The brain appears to work like a scoring system. Sometimes the results are highly unpredictable. Just like a computer, we are prone to apparently inexplicable results.
The difference is that the computer can be programmed to never do certain things, or to only be allowed to operate within a certain context. You can't. This programming can be made at a level that supersede
Re: (Score:2)
I think it's a stretch to assume we can reach human-style fuzziness without losing computer-style exactness
I expect we'll always have some of each. Computer exactness where it's applicable, and fuzzy thinking where that doesn't work.
Re: (Score:1)
Because when robots can do everything (or nearly everything) a person can do, what is the prospect for employment, let alone increased employment?
No employment, no moolah, no resources to survive. You can only survive by eating other unemployed people. IOW, and in the words of O'Reily, "Don't Fear AI, Fear Cannibalistic Humans."
If something like this does happen, programmers and engineers (along with scientists) will be held accountable for making automated tools that eliminated human labor and made it possible for the man to give your jobs to cheap robots. The Man will inherit the Earth and her resources, even though he did not discover the princi
Re: (Score:2)
so we need to stock up on guns and ammo. may some c4 as well to blow open the locks at the food storage place so we can eat as well.
Re: (Score:1)
Re: (Score:2)
No, can't you read?. Those will become obsolete.
We all need to stock up on automated gun turrets like the Samsung SGR-A1 [wikipedia.org] and other advanced automated solutions.
Re: (Score:2)
Because when robots can do everything (or nearly everything) a person can do, what is the prospect for employment, let alone increased employment?
The process of replacement of human labor with automation is actually slowing down. Manufacturing jobs are mostly already automated, and service jobs are proving very hard to automate. Many, many jobs require general AI, and despite progress in machine learning, we are nowhere close to human level machine intelligence. For now, it is just science fiction, and when it arrives it will change the world in such fundamental ways that "jobs" will likely be the least of our concerns.
Before, people always moved on to jobs that machines couldn't do. First from farming to manufacturing, then to services.
This is obvious in hindsight
Re: (Score:3)
Some thoughts on why this time will be different:
https://www.youtube.com/watch?... [youtube.com]
Re: (Score:2)
Some thoughts on why this time will be different:
https://www.youtube.com/watch?... [youtube.com]
This is just the normal collection of nonsense, myths and fallacies.
Myth 1: "This time" change is happening much faster
Fact: Change is happening slower this time. Productivity growth is stagnating not accelerating.
Myth 2: Demand will be flat despite falling prices.
Fact: Rising prices means prices fall in real terms, and people consume more goods and services. In the past, this rise in demand has outpaced productivity improvements.
Myth 3: Everything will be automated.
Fact: This ignores comparative adv [wikipedia.org]
Your "myths" are actual facts. (Score:1)
Re: (Score:2)
Fact 1: "This time" change is happening much faster
Nonsense. Productivity growth is stagnant, and has been since 2007. That is very well documented, and is the main reason that wage growth has also been stagnant.
Between 1880 and 1920, 40% of the the population went from farms to factories. Nothing like that is happening today. 5 million people drive for a living. That is THREE PERCENT of the work force. Those jobs may be wiped out by SDCs over ten years or so. That is a far less dramatic change.
The "deep learning" revolution started in 2006. How man
Re: (Score:2)
"Productivity improvements not only bring prosperity and rising living standards, they are the only thing that will do so"
Far from the only things. Any efficiency improvement will tend to raise living standards. That could be on the demand side as well as the production side. So will opening new lands, perhaps by global warming.
Re: (Score:2)
I would be interested in any evidence for these:
- Productivity growth is stagnating not accelerating.
- There will always be jobs where humans are relatively better than machines.
By the second one, I assume you mean there will be enough of those jobs to be significant to employment on a national or global scale, not just that at least two people will be better at something than machines. And I am taking you literally when you say always - meaning for the remainder of the existence of humanity.
In have seen the enemy (Score:2)
I have seen the enemy; it is us.
--
Sir Donald Sinden
If slashdot was around in the 70's (Score:2)
Re: If slashdot was around in the 70's (Score:2)
Weren't you on slashdot in the 90s? There were just as many luddites as there are here now.
Re: If slashdot was around in the 70's (Score:2)
Whoops thought your uid was 6 digits not 7. Anyhow, yeah, even techo geeks are afraid of technology. Sometimes justified, often not.
Re: (Score:2)
We can generalize the skepticism (Score:2)
People who say "AI and smart automation will not (net) reduce jobs" are relying on a general denial that certain fundamentals will or can change.
They can't imagine fundamental change of almost anything. Either cognitively they can't imagine it, or more likely, emotionally, they can't accept it and so employ irrational psychologically defensive denial.
These are the same people who say things like "solar and wind power can never be more than X percent of our energy mix" where X is usually somewhere around 10
AI and automation has been good thing so far (Score:2)
Because when robots can do everything (or nearly everything) a person can do, what is the prospect for employment, let alone increased employment?
Automation and ever 'smarter' machines needing ever-less human intervention has been going on since the Industrial Revolution. If people work less because machines do all the work, that means more spare time for the people. In today's day, despite all the complaining, people have never had more spare time to do with as they wish.
Think of a Dickensian urchin in a garment factory, or that poor bastard from the 12th century who spent all day in the mud or behind some ox pulling a plow. Or leave the 1st world a
Re: (Score:2)
Re: (Score:2)
If people work less because machines do all the work, that means more spare time for the people.
That's absolutely right. The problem is that being able to eat and having somewhere to live generally depends on getting paid to do work. When most people have nothing but free time, how do they get food and shelter? We need to figure that out.
Re: (Score:1)
Fundamentally, work exists because work = food & shelter.
Machines work for free = cheaper food & shelter./p>
More and smarter machines = ever cheaper food & shelter.
Enough really smart machines = free food & shelter? Maybe. But so far that is the trend.
Re: (Score:2)
More and smarter machines = ever cheaper food & shelter.
Except no. Food is getting more expensive now due to climate change, and shelter is getting more expensive now due to bank fraud.
Re: (Score:2)
We would need to change our society though. Why would the robot owners give away the products of the robots for free? If there are very few workers, how will anyone have money to buy those products? Having robots make all our stuff for free sounds great, but I don't think we can get there from here if we just let it happen. Something fundamental has to change.
Re: (Score:2)
Or it means more spare (or if you prefer, superfluous) people for the ruling class.
With great power comes great responsibility (Score:2)
With the Logan's Run level of turnover in engineering, engineers are rarely in a position to control the technology they create. And
Re: (Score:2)
Thanks. As the post says: "Our fears ultimately should be of ourselves and other people." Tim O'Reily seems to have great optimism about the outcome of "weaving our intelligence together into a greater whole." But misses two major problems: 1) there will be a lot of dysfunction and evil woven into that whole and 2) what emerges from weaving of our intelligence together will not be controlled or even really accessible to individual humans. Biological minds have been individual for hundreds of milliion
As it is with Everything! (Score:2)
There is always a risk when new change comes. Since the real world never stays static for long.
As an example from the world of Pretend.
Assimov's 3 Laws
The 3 Laws of Robots [youtube.com]
Revolution! But Whose?
The 3 Laws are Perfect! [youtube.com]
When ever anyone says they know for a fact, that yada, yada yada IS how things will work out.
Or such and such can never happen. I tend to take a closer look.
First we have to agree on objective reality. (Score:2)
...and that seems to be a serious problem.
Before we talk about bias in results, we have to be able to speak openly about reality without condemnation and public lynching.
If an "AI" system (ok we all know they're not actual AIs) says "black defendant X is more likely re-criminalize than white defendant Y (with almost exactly the same background), parole Y but do not parole X"...there are two possibilities.
Possibility 1 is that the system has derived its data from biased human sources, and thus is reflecting
When AI Takes Over... (Score:1)
Re: (Score:2)
When AI takes over, it will be with our consent.
"Our"? Who is this "us" you are implying? There is no "us", no "we", without a "them".
Kurzgesagt couldn't have said it better (Score:1)
It's not his ox getting gored. (Score:1)
AI only wins when you give up on learning (Score:2)
I really don't know how else to say this. The people complaining about "DEY TOOK OUR JERBS" are literally retarded. They have no mental capacity to learn new skills or to engage their own ambitions. All they do is sit around like vegetable complaining that they have no jobs. If someone hands them a job, what do they do? They sit around complaining how terrible their jobs are, and they fuck up until they are fired because who the fuck wants these retards around anyway? Fucking cancerous shits.
AI isn't going
Re: (Score:2)
Re: (Score:2)
I really don't know how else to say this.
For someone who clearly thinks they're one of the smart people your posts demonstrate a remarkable lack of vision, a dearth of wisdom, and a stunning disregard for the welfare of other human beings.
You should probably let go of some of your bitterness and bile, before you do yourself (or others) an injury...
Re: (Score:2)
I'll happily support anyone who dares challenge preconceptions or who try to make their own way in life. I will not, however, babysit incompetents or carry people who have no business being carried.
Fat people who eat themselves to death, stupid people who can't conceive of something to do with their lives, and fear mongering idiots who are afraid of their own shadows. These are all people who have no place in the world of tomorrow.
It's about time we go back to letting evolution do what it's supposed to do.
Collective (Score:2)
So, he's suggesting that we'll all become some small part of the Borg? Yeah, that's progress. Thanks, I'll pass. And really, you want to hold up Bezos and all of his anticompetitive practices as an example of not being an ass? Sorry, I don't share the opinion.