Hacker News Reader

Comment by Alex3917 | Hacker News Reader

He does frequently bring up retraining, but in the context that government-sponsored retraining programs have between a 0% and a 15% success rate. E.g. the graduation rate for former military employees who then go to college is 0%.

So retraining isn't a solution to jobs being lost by automation and even if we get significantly better at retraining it's still probably not a solution.

View on HN

You're going to need to provide a source for your claims that 0% graduate. A handful of my classmates from Carnegie Mellon were veterans (some with trauma) and all graduated within the typical timeline. The NVEST report from 2017 on student veterans [0][1] show an approximate 54% graduation rate within 6 years (18% were still enrolled, 28% dropped out) greater than the general averages. While the washout rate for 2 year programs is particularly high, your idea that retraining isn't achievable is flawed.

[0]https://nvest.studentveterans.org/wp-content/uploads/2017/03... [1]https://www.theatlantic.com/education/archive/2017/04/why-is...

If they got into CMU they're already the cream of the crop. The low rate covers all the other people who never made it to CMU, dropped out of the labor force and got onto disability.

The NVEST numbers (54%) include everyone though.

I see, I missed the part where it was veterans. I'm not informed enough to know why that program's success is so much higher than the ~37% success for manufacturing workers or Michigan's No Worker Left Behind (30% unemployment vs 40% for those not retrained).

If I had to guess, it could be because vets are younger and more able to learn new things, or it could also be that measuring graduation rates is different than measuring employment (plenty of college grads don't find employment in their field of study).

Hmm apparently I'm just wrong. I could have sworn there was a Wired article about this from a few years ago, but I can't find any evidence of it online or in my notes or in my previous HN comments.

Are there any retraining programs that have more than a 50% success rate? Has retraining ever been a realistic option for job loss? Seems more likely the next generation would make up for the lost jobs on paper

Nope. Retraining has never worked. Yang is quite explicit about that.

And he's wrong. Retraining has been effective in several European nations like Germany, Sweden, and Switzerland, especially for older adults (there's a reason even DeVos and Trump name drop them). The fundamental difference (from the US programs attempted in the past) primarily lies in the amount of money per worker spent and the level of state control/centralization. Furthermore, companies there seem actively enthusiastic about ensuring they have a skilled population pool in the future regardless of the short term ROI.

Personally I think the 2017 tax cuts should be rolled back for a lot of industries and used as an incentive for companies to develop proper apprenticeship programs and collaborate with universities/community colleges/trade schools to teach the exact skills they're looking for.

People don't have that kind of elasticity on skills, the ability to retrain a 20 year old and 50 year old are very different

I always hear about these “retraining programs” but I’ve never met anyone who has done one. Where do these programs even happen? What are they training people to do? Is it offered to people who file for unemployment or meant to be more proactive?

I teach data science at a masters programs (University of St Thomas graduate programs in software in the twin cities) and we have some students who are part of retraining programs. Apparently if a big company lays off a significant chunk of workforce and off-shores it, there are programs funded through corporate taxes that the laid off workers can use to go to school. The students in my anecdotal experience (2 semesters so far) are pretty good. I think there must be more than one such program because some students required a sign off from me but others didn't. It's just a few students that I know of so it's not a super large scale but it's nice to know that such programs exist at least.

Very interesting, thanks for sharing. I'm pleasantly surprised to hear the programs include such high-level courses.

Lol @ software engineers complaining about ageism and at the same time thinking retraining a 50yo truck driver is a reasonable solution

> Lol @ software engineers complaining about ageism and at the same time thinking retraining a 50yo truck driver is a reasonable solution

I'm not sure who these software engineers are. Anonymous online comments are unlikely to be a good source of determining how many people truly believe what you seem to think some of them believe.

Is retraining a 50yo truck unreasonable? Retraining them to become a software developer is probably unreasonable for most. But what about other jobs and types of labor? What about training them in more general skills that are valued in today's workplace and economy, as a part of a broader approach to tackling the issues facing the job market and economy?

I fear your military stat is inaccurate and taken out of context, making it meaningless for any inference on this matter.

Military population jas a selection bias for people with high thresholds of tolerance for difficult tasks amd by end of training usually have otjer confounding factors affecting their graduation, ie. Ptsd, lack of purpose, loss of fraternity, etc.

Can you back that zero percent number up with any sort of data? This article https://www.theatlantic.com/education/archive/2017/04/why-is... says that the number isn’t even officially tracked, though it’s probably low.

However, it also points out some places do manage to graduate a decent proportion of veterans, so it might be a matter of finding the right approach rather than an impossibility.

I'd be interested to know how many of those vets end up getting a job after they graduate though, other retraining programs measure job retention as opposed to program completion.

Reply on news.ycombinator.com