[TheOtherHobbes]

STEM work requires above-average intelligence, and even the most basic jobs are probably only available to 25% of the population. Hard-core mathematical engineering is only available to the top 1-2%. So yes - retraining is pointless if the raw ability isn't there.

But people don't need to retrain. What they need is an economic system that allocates their time usefully instead of declaring it worthless and wasting it.

The amount of work that needs to be done and can't yet be automated is almost infinite: infrastructure improvements, simple renovations, community projects, recycling - the unused pool of potential is huge.

[acjohnson55]

> STEM work requires above-average intelligence

I'm not entirely sure that's the case.

[mmierz]

It's not entirely the case right now, at least in sciences, there is still medium-skilled technician work.

But these jobs are subject to the same pressures as driving, cleaning and apple-picking. They're being increasingly automated and the technicians replaced by one person to operate the fluid-handling robots.

Edit: maybe I misinterpreted and you're saying that people don't need above-average intelligence to be an actual research scientist, perhaps on the non-tenure track level. Maybe that's true. But, it requires a strong interest and years of training, hence the long apprenticeship of the Ph.D / postdoc.

[acjohnson55]

I might have missed something -- I was thinking more of industry STEM jobs. I truly believe a lot of people can get into software development, for example, with normal intelligence, about a year of time investment, and perhaps above-normal self-motivation.

[jazzyk]

If you seriously think that all it takes is one year to become a competent software engineer/developer, sorry , but you have no clue what it means to develop good software.

Software development is devilishly complex (as in multi-dimensional, multi-disciplinary) to do right. It takes years to master it.

[acjohnson55]

I develop software for a living, and I dare say I'm better than most at it. I've been doing it about 20 years, since I was writing Z80 assembly as a middle schooler [1]. It would take a lot more than a year for someone to learn to do my job. But I'm a lead engineer, not an entry level developer.

However, it would take less than the total amount of time I've been building software to do what I do, because my path to my job took me through all sorts of software development I don't do anymore. That includes 8 years of higher ed, of which I apply only a tiny fraction on a daily basis. Don't get me wrong, I wouldn't trade all that education for the world. I draw upon it in plenty of indirect ways. I just know there are more direct routes for someone who simply wants to become an application developer.

I know this, because I know plenty of people who have taken less than a year to get into the industry, and are now productive developers. One of my friends even wrote about how he did it in detail [2]. There's a fantastic episode of Software Engineering Daily interviewing a guy who landed a lucrative job at AirBnB little more than a year after doing a dev bootcamp [3]. My own brother has been in the industry less than a year after coming through General Assembly and is doing quite well. I interview people all the time who have similar stories. They're not unicorns.

Developing software isn't easy. But it's also not that hard. At least the sort of software most of us build. You only need a couple people at the typical enterprise who are capable of doing architecture and the most technically intense stuff.

[1] http://www.ticalc.org/archives/files/fileinfo/61/6175.html

[2] http://alexkrupp.typepad.com/sensemaking/2013/11/2012-my-yea...

[3] http://softwareengineeringdaily.com/2016/07/11/salary-negoti...

[nradov]

A person of average intelligence but with above average determination can earn a PhD in some scientific fields and make a useful contribution. I've seen it done.