Hacker News Reader

Comment by snrplfth | Hacker News Reader

That's been the "theory" for a long time - say, the threshing machine will mean that farmers will be unemployed, the powered sawmill will put all the sawyers out of work, the spreadsheet program will put all the accountants out of work.

The trouble with that theory is that people invent and move into new services and jobs. Sure, the old work was automated, but that just means that people can redeploy their labour into an work that simply wasn't being done before.

Why doesn't this mean that resource-owners will control everything? Because "resources" are transformed by technology in the same way. Things that were previously not thought of as resources, become so. Secondly, one still imagines competition. There is no "he who owns the resources", there's many, and as long as they have to compete for the custom of the other 95% of the economy, they can't control, or coordinate to control, much of anything.

View on HN

Newer jobs have always replaced older jobs, and the shift in job titles disguises this.

When we say "farmer" now, we mean someone who uses electronics-enhanced tractors to plant and harvest at high yields on 1000 acres. But it used to mean someone who produced food from 5 acres with only human-operated tools. Later, it was someone who used ox-powered tools to produce on 20 acres. As the tech advanced, the job responsibilities changed.

Accountants are no longer responsible for adding the numbers correctly in handwritten ledgers. They use computer spreadsheets. But they are still accountants. The market can either use less of them for the same work, or require that they do more work for the same number of laborers.

The problem that is arising now, which did not arise in the industrial revolution, is that more employers are choosing the former option, simply because they cannot imagine more work for the same number of people to do that could not also be automated away.

If the only way for cab drivers to compete with self-driving hire-on-demand cars is to become tour guides as well, it only takes a few more years to add a synthesized voice recanting trivia to passengers about nearby points of interest to the cab-driver-replacing software. As long as anyone has the intent to replace expensive human labor with cheap machine labor, no offer of additional added value is sufficient to save everyone's jobs.

If no job is safe, how do you know which field to move into when your old one goes obsolete?

I'm sticking with software development. If the robots ever get better at writing their own software than us then we'll have other problems to worry about besides work :)

Plus there's a chance that it will open up a market for artisanal hand-written software (when people long for the little bugs and imperfections that make if feel authentic).

If the robots ever get better at writing their own software than us then we'll have other problems to worry about besides work :)

I wouldn't sleep on that. It's coming.

To be fair, writing code better than the median software quality is not a tremendously high bar to pass. There is some truly awful code out there, and plenty of mediocre CRUD apps to add bulk around the middle.

It is far more worrisome if software can write better software than the best software-writing humans. After that point, there is no job any human could do that would be competitive with AI.

At that time, meat-humanity will just have to retire from the workforce, move in with its AI grandchildren, surrender its drivers' licenses, and eat discounted ramen.

It is far more worrisome if software can write better software than the best software-writing humans. After that point, there is no job any human could do that would be competitive with AI.

I don't agree with this.

From an industry perspective It should be worrisome even if it's not better than the best because as you know, it's not required to be the best to take over an industry.

From an AGI perspective, you are generally correct that there will be no job that a human does better than an AGI, however between the point at which it kills programming jobs and when it dominates everything there will still be bespoke jobs. Even though Nike dominates, there are still cobblers.

Exactly. I'm sure there will be room for a locally produced software stand at the farmers market.

Generally we don't, and really can't, know ahead of time which new work will be invented and done in the future. That is the point of entrepreneurialism, and of a market more broadly. What I'm pointing out is that it has consistently been the case that people have declared that, because they could not imagine what new kinds of work would exist in the future, that therefore automation would throw people into permanent unemployment. This has consistently been mistaken.

Much work could be automated - but automation is not free. Sure, you can imagine all sorts of work being automated away - but is it worth the cost? In what locations? For what applications? All the time or just some of the time?

The most common way of adjusting to automation is not just to add new value to an old "job title", but to move into new fields entirely. A farmer today has very different responsibilities to a farmer a hundred years ago - but there's also far, far fewer people farming at all. Most people who do things that have been automated will do things that simply weren't done before, or were done very little.

Consider also that in the past, many of the jobs shifts took place on a generational scale. A farmer--instead of having 10 children that were farmers or farmers' wives, and one that was a non-farmer--would have one child (or less) that became a farmer, and the rest would become non-farmers.

The economy changed slowly enough that you could have the same job your entire life and not be suddenly unemployed or unemployable when the technology changed.

After the industrial revolution, the pace of economic change accelerated. You can now make an entire job obsolete in a shorter period of time than the ~50 years it would take for a new worker to reach a socially acceptable retirement age. And what is that person going to do for a living? Can they afford to restart their lives from the second-lowest rung on the ladder?

It's not necessarily always easy or painless. I certainly didn't claim it was. But you can't really grow the economy without introducing labour-saving devices that change the composition of jobs. Economies didn't have 90% of people working on the farm because they preferred it to anything else; it was that way because it had to be - the productivity of labour was so low. Likewise, if we want the economy to continue to grow at reasonable rates, it must be understood that job change is part of that.

Reply on news.ycombinator.com