Comments on: The Returns to College Going Forward

By: Andrew'

Andrew' — Sun, 18 Jan 2015 10:54:09 +0000

They can’t even play chess.

By: Slocum

Slocum — Sat, 17 Jan 2015 22:26:03 +0000

Yes, there are more things that could be automated, but the barriers are social rather than technical. The Khan academy could be supersized, but much of what is hypothetically achievable by Khan academy online videos would have been almost equally possible with the same content distributed on VHS tapes 30 years ago. Yes, everybody at McDonald’s could be ordering and paying via smartphone a la Uber. Same goes for waiters in restaurants (why on earth do we need somebody to take our orders via dictation, shuttle it to the kitchen, and then do the same with the bill at the end?) And those changes may indeed happen. But they won’t require any new technology that’s not already available.

By: Bryan Willman

Bryan Willman — Sat, 17 Jan 2015 02:17:21 +0000

One real problem in these “returns to education” discussions is that they always seem to regard all forms of education as equal. That is, the return to education in classical literature is lumped together with the return to education in physics which is lumped together with the return to education in computer science.

Today’s hint – the return to education of different areas/fields is different.

By: Buzzcut

Buzzcut — Fri, 16 Jan 2015 23:38:28 +0000

Meter reading is recently automated. On the road to truly smart meters and demand based pricing for utilities.

One wonders why McDonalds has not embraced self checkout.

A wireless network between vehicles has the potential to almost eliminate multi vehicle accidents.

There’s a long way to go with automation. Even education. Super size Khan Academy.

By: sort_of_knowledgable

sort_of_knowledgable — Fri, 16 Jan 2015 23:30:48 +0000

The claim was Increased supply of skilled labor -> increase wages for those who were _newly_ skilled and a reduction in inequality.

The wages for skilled labor would decline but overall wages would be up because more people would earn the wages for skilled labor.

By: Andrew_FL

Andrew_FL — Fri, 16 Jan 2015 22:37:53 +0000

Increased supply -> Higher price

Huh? Increasing the supply of skilled labor drives the wages of skilled labor *down* unless the demand curve is not downward sloping.

By: paul

paul — Fri, 16 Jan 2015 20:25:39 +0000

Moore’s Law is basically over. Yes, computers and smartphones will get faster and have longer battery life. But they won’t get cheaper any more. We have reached a sort of limit in semiconductor manufacturing (if you want the details, we are building all chips with 193nm wavelength light despite the feature sizes being more like 58nm or below, so we have to do multiple passes called double-patterning). It looks like 28nm will be a lower cost process than every process that came before (which is basically Moore’s Law) but will also be cheaper than every process that comes after too.

In the last 40 years we have seen a 1,000,000 reduction in the price of electronics. That will not happen in the next 40 years.

By: Slocum

Slocum — Fri, 16 Jan 2015 20:19:14 +0000

“Change in technology is being led by Moore’s Law. The core components of computers get twice as good every couple of years.”

Processing power is really not the obvious bottleneck for progress in AI. Nobody thinks that machine visual object recognition or, say, machine translation of text between French and English is going to get twice as good every 18 months.

When people talk about automation replacing labor, what examples do they point to? ATMs, automated checkouts lanes, voice mail, text search, and factory robots, and the like. What these have in common is that they don’t rely on AI, won’t really get much better due to Moore’s law, and they’re not *new* — they’re all at least 10 (and really more like 20) years old. I really can’t think of a new job category that has been automated within the last few years, can you?

Yes, there’s computer chess. But it is absolutely not representative or transferable. There have been huge investments in computer chess playing over many decades, and the results have been awesome chess-playing computers — that literally can’t do ANYTHING besides play chess.

By: JKB

JKB — Fri, 16 Jan 2015 19:06:55 +0000

Actually, electric motors themselves reached a kind of perfection very early on. The limiting, both cost and technology, element was in the motor controllers. Since the march of the microcomputer began great strides in electric motor efficiency and functionality have been made as the motor controllers have advanced. Recent advances in power electronics appear very promising.

Contrast that to the lack of advancement and innovation in the provision of instruction colloquially called education. Regardless of the tinkering with technology, almost all instruction at all levels is still based on the slow and inefficient “sage on the stage” which has a high probability to induce “school helplessness” in the student, that is lower initiative in school subjects than in other aspects of learning.

Learning ability could possibly be there but its usage is throttled by old controllers and, unlike electric motors, the longer a student is limited by such antiquated controllers the more limited their ultimate capacity becomes.