Hacker News Reader

Comment by PhilWright | Hacker News Reader

I am sceptical that robots will lead to mass unemployment.

We have heard this luddite fear repeatedly since the start of the industrial revolution. There used to be 90% of people employed in agriculture but today we are down to something like 10% in the UK. Do we now have 80% unemployment?

A hundred years ago the UK had mass employment in heavy industries such as ship building. They have long gone and the UK is now a services based economy. Do we now have 80% unemployment?

There is no limit, so far, to human wants and needs and so releasing people from some industries will allow them to move to others or create entirely new ones.

View on HN

But, each wave of new jobs requires more specialization and training than the previous, and the period between each wave decreases.

Unless we can figure out how to retrain the redundant workers more quickly, we are going to inevitably have more and more unemployment.

We also have no way to accurately forecast future labor demand. So, people end up wasting lots of time and money to train for jobs that won't exist by the time that they're done training.

I already see this with programming. Tons of people are being displaced from other work, and they're trying to get into programming because they've been told it is the future. But, a good programmer takes many years to train. And, the skill requirements for programmers increases every year.

Programming will still be a good career for some time, but a 'basic' programmer job of today will probably only have another 5-6 years or so until it's mostly displaced. The next wave of programming jobs will require advanced math and scientific skills, at least.

I used to think that way until I watched C.G.P Grey short video on the topic "Humans need not apply" https://youtu.be/9BMcks55vfE

I think you linked to an incorrect youtube video, though those GeoOrbital wheels look cool (and heavy!).

Actual link: https://www.youtube.com/watch?v=7Pq-S557XQU

Whoops - thanks for the correction.

We have also the red queen problem of responding to increased capabilities by adding additional complexities. For example: faster processors and more memory haven't made computing incredibly fast because we add more features or simply more bloat. See also legal contracts, which before the days of word processing, were no more than a couple of pages, and now even simple contracts can be tens of pages and complex agreements can be hundreds of pages long.

I'm curious about this viewpoint. Do you think it will never happen ever, or just not in the next few decades? If not autonomous vehicles and burger flipping mcrobots then what's the theoretical tipping point if one exists?

no op, but personally I think an increase in automation will simply push human endeavours upward in Mazlow's pyramid of needs.

ie. prior to the industrial revolution most human activities were related to satisfying our physiological needs. When that was largely automated we worked on increasingly abstract desires (eg. TV/entertainment industry, social media)

In a post-robotic world I think we'd focus on more human-centric activities. Eg. what if we had 10x or 100x the number of teachers, and the educational system was more like 1 on 1 tutoring rather than large classrooms. What if food became more like music in that its physical reproduction is automated, but humans are paid to design it.

ultimately the luddite argument belies a lack of imagination. Most jobs that humans do today didn't exist 100 years ago. If and when humans truly become redundant in the face of strong AI, there wouldn't be any point in worrying about whether all humans can remain employed.

Reply on news.ycombinator.com