* 80% fewer lawyer jobs because algorithmic contracts will kill the rest of the lawyer jobs that e-doc review didn't kill already. The 20% that remain will be lawyers/programmers.

* 80% fewer medical doctor jobs because all surgery will be done by robots because it's safer, radiology will be automated because it's more accurate, and most diagnoses will be automated through DNA analysis and readings from wearables. Doctors will need training in presenting (quite possibly bad news) results to patients and training to run the machines. The 20% that remain will be doctors/programmers.

* 80% fewer jobs in "driving vehicles" or "packing things" (truckers, factory workers, Uber drivers) because self driving cars will be mature 20 years from now.

* Restaurant jobs will still be around. But the cooking will be automated for 80% of the food sold in America.

* Business process automation paired with private equity firms and competition from startups will automate away a surprising amount of managerial roles.

* 80% fewer retail jobs because Amazon has helped make it so you can just walk in and pick something up and walk out and be charged for it, without any human involvement.

20 years is the difference between 1994 and 2014. I might be wrong about some of it, but I think people underestimate how much the job landscape is going to change and how important of a skill "programming" (which means a lot of different things) will be. Everyone should learn to program. And basic income, but that's a whole different can of worms.

Interesting; this contrasts with the claim of liberal arts graduates who maintain that simply being competent critical thinkers is all that's required for almost any job.

Programming requires “training and education”, too.

But for many jobs, especially if you have connections that let you bypass filters designed to winnow the applicant field down to a manageable number that throw out lots of otherwise-atrocious candidates, you can demonstrate that other than by the degree that is “expected” for the field.

Computing is by no means unique in this respect (and the big players in computing are as much known for near-total exclusive preference for top schools as is the case in the worst of other fields.)

Are you implying that you don't need skill and knowledge to program?

Or that it's impossible to get skill and knowledge needed in "other reasonable paying jobs" mostly outside of formal education industry?

That only works in industries like software etc. If one wants to be a lawyer, doctor etc they have no choice but to go to college, right? Some American students are going abroad (Germany etc), that could be a short term option.

For lawyers in some US jurisdictions, including California, no; California, for instance, requires general education of two years of college or “demonstrated equivalent intellectual achievement” prior to study of law, and several options for the study of law, including what amounts to apprenticeship in law office or judge's chambers. So traditional college or law school is not strictly required.

The top law firms employ small percentage of employed lawyers. And not even a huge percentage of graduates of top law schools. And aren't even the kind of work some people going into law want to do.

Confusing “what you need to get a job at a top law firm” with “what you need to be a lawyer” is like confusing “what you need to get a job at Google” with “what you need to be a programmer”.

> And the legal field is generally oversupplied with lawyers so if you don't work at a top law firm you won't be making a great living for a long time.

If you avoid undergraduate and law school debt (and especially if you are a paid employee instead of a paying student during your legal education), you can make a great living at a substantially lower salary than if you don't; and olif you do a legal apprenticeship, you come out with real-world experience most law students won't have (and quite often a job in the office you apprenticed in, if it was a law office and not a judges chambers.)

If you don't get an apprenticeship, then your options after passing the bar with self-study are essentially (it seems to me):

1. low-paid government work (like public defenders)

2. setting up your own practice (hard to build up a client base initially)

3. working on contract/temp basis for another law firm with variable prospects

Law school isn't just about credentialling and signaling; it also sets you up with a valuable alumni network that helps you get work. Anecdotally it's easier for an aspiring programmer without a network to get a referral into a tech company; go to a few meetups, hackathons or conferences, build a portfolio/github, meet a few people and ask them if they'll refer you (I've both gotten a job and helped someone get a job this way). And very few people ask or care about where you went to school when they can see your work instead. Again I may just be biased because I know the tech field better but it seems to me that it's harder to demonstrate your legal skills since you can't exactly post a sample contract or legal opinion you wrote, on Github (or maybe there's a portfolio site for lawyers too, who knows).

> Confusing “what you need to get a job at a top law firm” with “what you need to be a lawyer” is like confusing “what you need to get a job at Google” with “what you need to be a programmer”.

Again, not knowing the legal field well it seems to me that even mediocre programmers do better than mediocre lawyers because of the supply-demand differences between the two fields.

There are very small numbers of people who choose that route (and very low awareness that it even exists); as I understand, there are mere dozens actually doing it any time in California recently.

One of the more common routes, reportedly, is people who are already employed (e.g., as paralegals) in the office they become apprentices in.

and getting into medical school is basically luck of the draw; there are far more overqualified applicants than necessary. thinning that herd a bit isn't going to hurt.

https://www.youtube.com/watch?v=wdAfILoJWcI

And his lectures are available for free.

If you get a degree, lots of what you learn will be outdated fast. The most important things you learn will be on the job actually making things.

> Unless maybe you are going to a top 1% CS program, there is little need for a degree in software engineering.

I think you missed what I just said. You say there "is" little need. But I'm not looking at the present. We still don't know if we do for the long-term. (Well, lots of people think they do, but they're only guessing, not going on evidence obviously.)

Note that it's simply not true that people would learn the same material independently that they would learn in a classroom. Yes, it is theoretically possible, and some people can pull it off, but it's not at all universal or even common. A lot of people need the environment to learn better. Online lectures just don't cut it for everyone. So yes, "you need to go to college" can be true regardless of how good online lectures are, and we still don't know if what people get out of a CS degree will on average give them an advantage compared to those who don't get one in the longer term (though I don't really have statistics for the shorter term either).

That's not always true. Most people seem to think they "need" a car, yet the price of cars isn't astronomical.

But you're right that even though the cause is noble (education available for everyone) our current mechanism of funding it results in runaway prices and for-profit scams.

I don't think everyone should not go to college and learn to code though. I think in the near future coding will be a basic piece of literacy, but only one piece. I think it's reasonable to imagine a future where just like people study english and learn to do some writing in english 101, they will also learn the skills to munge a csv file or write an sql statement in programming 101, and both will be part of a well rounded education.

The last thing this industry needs is a flood of high school graduates creating shitty insecure websites. I say this as someone without a degree, who knows first hand that although it's possible to make it w/o a degree, it shouldn't be anyone's first choice.

Cars are not subsidized in the same way that healthcare or education are subsidized, though. There are fancy, expensive cars for people who want/can afford them, and there are more modest cars for the rest of us. There are lots of manufacturers of cars from around the world (and those cars from foreign manufacturers are available locally).

Of course, there are fancy expensive schools, and cheaper schools, but the subsidies (or maybe some other force I'm unaware of) seem to make the 'cheaper' schools still get more expensive every year.

As to how to fix it, it's complicated. If the fed wants to continue to pump money into universities, they could federalize them, but it would be a huge transfer of power to transfer control of universities from the state to the fed. There's accreditation, the board of regents, and a variety of other organizations that would then be federally controlled. On top of this, since a university is such a large piece of a local community and requires infrastructure to support it, it makes sense to manage it locally. Further, the state would probably just play the same game if a university was federally controlled and still milk money. Really, the states need to realign their priorities and increase funding for the university system if they want stuff like this to stop. Removing federal student loans would help alleviate some of this pressure, but there's still pressure on universities to move to adjunct faculty to reduce costs as well, which is what the article was about.

More money entering an economy will certainly allow buyers to bid up prices (as with houses and college tuition). The question is which other prices are being bid up even more rapidly and which are remaining flat or declining on a relative basis.

These sorts of problems are why both illicit and legalized counterfeiting are so pernicious.