7 ms·

> Instead, my job will be outsourced to a datacenter. Everything I do professionally, from reading and giving feedback on specs to estimating development schedu

by norswap 12y ago

> Instead, my job will be outsourced to a datacenter. Everything I do professionally, from reading and giving feedback on specs to estimating development schedules to architecting and typing in the code and writing unit tests to reviewing others' code, every one of these things will be subject to automation. We can't do it now, of course, not at any price. The whole thing derives from these flawed premises. Machines are incredibly dumb. Even these machine-learning wizz-things. They can do statistics, inferences, correlations. That's very very far from enough, especially for something like programming, which we can see as turning a spec into code. First, one still has to write a spec (so the "programming" isn't totally gone). Second, it means you essentially need to have "solved" natural language processing; as well as have assigned semantic values to linguistic constructs. Good luck with that. At this point, I feel compelled to point out that humans don't agree on the meanings of what they say. Most of the time, humans can't write software that meets a spec. That's because of ambiguity, and machines are way way worse at dealing with that than humans are. I think the error is simply to think the current rate of progress will keep going (and actually even the idea that there's been an acceleration of progress in AI recently is an illusion -- what we're seeing is increased application). It's the same error the fathers of AI made when they assumed everything would be solved in a matter of years. That's where the expression "AI winter" comes from.

oconnore 12y ago
Most people are paid to drive trucks, not solve novel problems or understand subtle linguistic constructs. http://www.npr.org/blogs/money/2015/02/05/382664837/map-the-most-common-job-in-every-state http://www.npr.org/blogs/money/2015/02/05/382664837/map-the-...
- DonnyV 12y ago
  Exactly and those jobs will be automated in about 5 to 10 year's. We as a society are not prepared for that.
TeMPOraL 12y ago
First of all, > Machines are incredibly dumb. Even these machine-learning wizz-things. They can do statistics, inferences, correlations. Do you think human brain does something more? All we can do is statistics, inferences and correlations, and we do it poorly. Moreover, there is an even bigger problem than 100% automation of everything - namely, the automation of 50% of jobs. What are we going to do with half of the society being unable to find any job? And that most likely includes your or mine parents, siblings and friends, who will no doubt come to us techies for assistance. And if current trends continue, it's 10 years away, not 100.
- dm2 12y ago
  Other technologies will make the need for jobs obsolete. It's called a post-scarcity society and it must be achieved at the same time or slightly before most jobs become automated. A molecular re-arranger would be the holy grail of a post-scarcity society. Also, you're not giving enough credit to the human brain. Whatever it does, we can't even get close to replicating it with current computer technology. Yes, they can emulate a bee or an ant with computers, but that creature doesn't have the ability to reason and come up with ideas, and dream (as far as we know and are emulating).
- nkozyra 12y ago
  At present the only difference is the human brain is far more malleable and adaptable than is a machine. Our most flexible AI is generally based on supervised learning. That still requires human interpretation and verification to work properly. I think you can also state that machines generally lack the ability to discern expansive context unless explicitly trained to do so. Granted, it will happen. True AI requires these things. At present all ML needs a push from a human to become more "intelligent."
- spiralpolitik 12y ago
  I don't think the current economic model can survive 50% unemployment and to be honest I think social unrest would take its toll before we even got that far. We may even see a partial collapse of society in some developed countries leading wiping out the advances of last century. The bigger question is can we as a species evolve into a more altruistic species and move beyond capital. Some days I don't think we can...
elihu 12y ago
I think the idea that programmers will be replaced by software-writing AI programs is silly and unrealistic, but that isn't to say that programmer jobs are secure. What's really going on is that developer's tools are getting better. Languages, debugging tools, version control systems and bug trackers are all rapidly reducing the number of developer-hours required to implement a given program, which we should expect to reduce the size of the programmer workforce eventually if we can't find new things for them to do. It is possible that we will find enough new problems for programmers to solve to counteract increased productivity, but for many applications, most of the general problems that one would hire a programmer to do already have reasonable-quality solutions. I think we'll see smaller and smaller teams of hyper-productive developers working on a shrinking set of increasingly esoteric projects, or things with a single specialized use (like calculating taxes for that particular year).

by norswap 12y ago

oconnore 12y ago

Most people are paid to drive trucks, not solve novel problems or understand subtle linguistic constructs. http://www.npr.org/blogs/money/2015/02/05/382664837/map-the-most-common-job-in-every-state http://www.npr.org/blogs/money/2015/02/05/382664837/map-the-...

DonnyV 12y ago
Exactly and those jobs will be automated in about 5 to 10 year's. We as a society are not prepared for that.

TeMPOraL 12y ago

First of all, > Machines are incredibly dumb. Even these machine-learning wizz-things. They can do statistics, inferences, correlations. Do you think human brain does something more? All we can do is statistics, inferences and correlations, and we do it poorly. Moreover, there is an even bigger problem than 100% automation of everything - namely, the automation of 50% of jobs. What are we going to do with half of the society being unable to find any job? And that most likely includes your or mine parents, siblings and friends, who will no doubt come to us techies for assistance. And if current trends continue, it's 10 years away, not 100.

dm2 12y ago
Other technologies will make the need for jobs obsolete. It's called a post-scarcity society and it must be achieved at the same time or slightly before most jobs become automated. A molecular re-arranger would be the holy grail of a post-scarcity society. Also, you're not giving enough credit to the human brain. Whatever it does, we can't even get close to replicating it with current computer technology. Yes, they can emulate a bee or an ant with computers, but that creature doesn't have the ability to reason and come up with ideas, and dream (as far as we know and are emulating).
nkozyra 12y ago
At present the only difference is the human brain is far more malleable and adaptable than is a machine. Our most flexible AI is generally based on supervised learning. That still requires human interpretation and verification to work properly. I think you can also state that machines generally lack the ability to discern expansive context unless explicitly trained to do so. Granted, it will happen. True AI requires these things. At present all ML needs a push from a human to become more "intelligent."
spiralpolitik 12y ago
I don't think the current economic model can survive 50% unemployment and to be honest I think social unrest would take its toll before we even got that far. We may even see a partial collapse of society in some developed countries leading wiping out the advances of last century. The bigger question is can we as a species evolve into a more altruistic species and move beyond capital. Some days I don't think we can...

elihu 12y ago

I think the idea that programmers will be replaced by software-writing AI programs is silly and unrealistic, but that isn't to say that programmer jobs are secure. What's really going on is that developer's tools are getting better. Languages, debugging tools, version control systems and bug trackers are all rapidly reducing the number of developer-hours required to implement a given program, which we should expect to reduce the size of the programmer workforce eventually if we can't find new things for them to do. It is possible that we will find enough new problems for programmers to solve to counteract increased productivity, but for many applications, most of the general problems that one would hire a programmer to do already have reasonable-quality solutions. I think we'll see smaller and smaller teams of hyper-productive developers working on a shrinking set of increasingly esoteric projects, or things with a single specialized use (like calculating taxes for that particular year).