Comment by ssivark | Hacker News Reader

Comment by ssivark | Hacker News Reader

Parent Full thread

Amara's Law: "We tend to overestimate the effect of a technology in the short run and underestimate the effect in the long run."

Not to seem like a luddite, but let me play the devil's advocate here. 15 years, IMHO, a huge exaggeration. The standard technology of today: IVRS systems suck so bad that they make me want to slam the phone down after 2 minutes, if I cant talk to a human. Siri or GoogleNow aren't much better. When you force a human to interact with a computer, you are handicapping them and taking away most of their expressiveness. Guess what, humans don't like that. You reduce them to the level of the computer which can only understand a few pre-programmed options. That seems like a far cry from human communication which it will try to replace. And we don't even know if that's possible even in principle. All we have is the hope from a bunch of people who are generally excited by technology, and the ensuing hype.

Keeping Amara's law in mind, I would imagine that such a drastic change would take a longer time (almost 1 generation's working life?). Incremental changes summed up might displace jobs over decades, but if someone can adapt to those incremental changes, I would imagine that they can manage their whole working lives. At least as applicable to people born till (say) 2000 and would presumably live till 2100 or so. Any "predition" beyond that would be a wild leap. As for humans being born circa now...

However, I believe that a few questions are still of supreme importance and we need to tackle them for the sustenance of mankind (pun unintended).

1. If a computer/robot could exactly replicate a job (and do it on a macroscopic scale), will such a solution be energy efficient? [Assume that you already have a population of X billion whose needs you will have to meet]

2. What is the real aim of automation? Is it to save costs from the perspective of the business owner, or to give people leisure?

3. Let's say robots/computers replace a large segment of jobs. Presumably, a small number of people will own and operate those systems of scale (you just have to make a 3D printer which can print other 3D printers). Then what is the role of the "unemployed" humans? And most importantly, any such structure we come up with, must lead to a stable society.

I find it interesting how many smart people seem to truly believe that things will just work itself out. Kind of like the idealist libertarians or communists, only with technology as the solution to every problem.

They generally also believe in the invisible hand of the capitalist market.

That is something that can only "believed" in or "disbelieved" as much as gravity can be "believed" in or "disbelieved."

One has to be careful when applying the "invisible hand" idea. You can't just take an "ideal" situation (fixed point) and say that the system will evolve towards that. For an example, I refer to my comment on this HN thread -- https://news.ycombinator.com/item?id=6313482

Gravity can be tested just by dropping your pencil. The invisible hand takes a lot more faith and fervent wishing, IMO.

To test the invisible hand, as described by Smith, you just need to buy something useful from someone who just sold it for money, and not out of any desire of helping people.

People just inflate the concept to ridiculous levels, which were clearly not intended by the author.

I believe in the invisible hand, but I also believe it's drunk.

That's probably because there's no good reason things won't just work themselves out.

Eventually, all blue collar jobs will be replace with machines. Only the people who make the machines will remain employable. Eventually, white collar jobs will be replaced too; we're already seeing this in the field of law where an army of clerks is no longer necessary. I'd imagine health care will be a prime target given how outrageously inefficient it is in the US and the percentage of the GCP that flows into it. Its very easy to imagine a society where at least 25% of able bodied, working age people are categorically unemployable not just working itself out.

Look up the luddite fallacy.

When we, as a society, reach that point, goods will be so cheap that people won't need to work to live.

"goods will be so cheap that people won't need to work to live."

That sounds like a nice, optimum fixed point. Let's think about how we might get there. It's highly unlikely that things become cheap and then humans lose jobs. It will probably be the other way round. Consider a time slightly before we get to that utopia of yours -- things are not quite cheap enough, but lots of human jobs have been "obsoleted". In such a situation how do the unemployed people survive? How does the system evolve into your utopia, rather than in a direction where the "have-nots" are neglected while the "haves" cater to their own needs? In other words, your utopia looks like a nice fixed point, but it's not a stable or attractive (in the technical sense) fixed point, of this dynamical system.

This is an example which shows that correlation is not enough. It is important to keep in mind what the causes and effects are. You can't have the effect happening first and inducing the cause. The causal order is important.

Most engineers/technologists probably never see the inside of an ethics or philosophy course.

Reply on news.ycombinator.com