Hacker News Reader

Comment by ainiriand | Hacker News Reader

In my opinion, massive automation of tasks is going to cause a massive demand of programmers, developers and QA engineers, so no robot apocalypse yet. Basic income is not the solution to this problem because this is not a problem. The solution to this 'problem' is having a better education and making tech more accessible for everyone.

View on HN

There will be more demand for a short while but a lot of that will also be automated away. Think about a period of 100 years aka your children and grandchildren. Not the coming 3 years as we programmers tend to do. Even if we will have another AI winter after this spring and summer; 100 years gives enough time for another spring and summer. Not only can not everyone do everything; many people cannot learn programming and QA. I would say by far most will not be able to. That is a difference with the industrial revolution; something like a horse carriage to bus driver transition is vastly different than a truck driver to a dev.

Sure for jobs like lawyer or doctor to programmer you can imagine but there are not many compared to physical labour anyway.

For both reasons, I do not see your opinion happening in the 100 year and after span. Maybe over the coming 10 but not 100. Wether I like it or not, my brain does not get over the choice of letting people perish, basic income or boycotting the future. Someone like Trump can bring jobs by just banning robots ; for instance by changing regulations so they cannot drive vehicles on public roads. But that would pretty quickly (again look at 100 year spans) lead to the US being no longer competitive on the world market for those particular (related) services.

Well, it is really hard to think about what is going to happen regading jobs in 100 years. I dont think that human mind is going to be removed from the equation. And if human intervention is, in fact, removed from most of the jobs, then what alternatives to get the basic needs are they going to have? Also, we have to be extremely careful because, even if we are not going to be there, we are paving their destinies now.

I agree we will need the human mind but for less and less jobs. And indeed my point is that we are paving that destiny now and therefor we need to work towards some kind of solution now.

> In my opinion, massive automation of tasks is going to cause a massive demand of programmers, developers and QA engineers, so no robot apocalypse yet.

The question is not how massive it is in absolute numbers, but how does it compare to the amount of jobs that are replaced by automation?

And you know the answer to that question. If all jobs so replaced would create an equivalent number of "programmers, developers and QA engineers", nobody would automate, because it would cost more than human laborers.

That's actually not true. The effect would be an increase in productivity. More things get built/done with the same resources.

The only resource conserved in this case is the amount of time spent to produce some good - it's not like robots can magically make more out of the same chunk of steel, say.

But if you need the same amount of people to produce the same quantity of product (jut in different roles), then, by definition, your automation does not conserve time.

On the other hand, if your automation does decrease amount of time spent per unit of output, and you run it at full capacity, then where do all the extra units go? If you say "other people", then that would imply that there was some heretofore unsatisfied demand that the market couldn't satisfy due to a labor shortage.

Automation is a great way to solve the labor shortage problem, but it's not always present. And when it's not present, you're just adding more goods without adding new workers (or increasing wages for existing workers) to afford those goods - so you'll end up with an overproduction crisis.

Note, this is not an automation issue. This is a capitalism issue, because in capitalism, the only demand that matters is the one that lets you make a profit. People who may need what you make, but cannot pay above cost of production for it, don't matter. Automation tends to increase the size of that category.

What's massive though? Programmers currently make up approximately 0.8% of the workforce. You could increase the size of the industry by an entire order of magnitude and you still would only be employing 8% of the workforce. About the same rate as manufacturing today; a sector often scrutinized for no longer providing enough jobs for the populace.

Massive automation of tasks performed by humans is pretty depressing because it indicates that some combination of technology focus/funding can't aim higher than what people are doing already.

You are assuming you cannot be automated away, why?

I wouldn't be surprised if QA is completely automated right about the time that facial recognition becomes viable. I guess that means many years from now, but if the hitch to automating QA is having person-eyes looking at a page and using the site in order to find problems, then I think they might ultimately converge their tooling because there's similarities in there. Something you can throw your PSDs or Balsamiq URLs at it and wait for the notifications. Web page recognition rather than facial. Heck, it might be simpler...or already exist.

Update: Here's a guy doing text recognition for localization: http://www.jebriggs.com/blog/2013/04/imug-computer-vision-an...

Also: http://browserbite.com/

Comparing screenshots from Selenium is a pretty nice way of doing it.

Reply on news.ycombinator.com