Arnold Kling  

Diamandis and The Diamond Age

PRINT
Property Rights ARE Human Righ... A Libertarian Defense of Black...

Peter Diamandis is trying to think of an X-prize to fix education.


he has considered multiple directions that an Education X Prize could take, such as coming up with better ways to crowd-source education, or rewarding the creation of "powerful, addictive game" that promotes education. But he isn't sure which way to go.

Neal Stephenson has already pointed the way: The Young Lady's Illustrated Primer. Create a friendly, personal coach that constantly talks to the student and adapts to the students' needs.

It seems to me that the technology to do that is pretty close.


Comments and Sharing





COMMENTS (8 to date)
wintercow20 writes:
Create a friendly, personal coach that constantly talks to the student and adapts to the students' needs.

You mean a parent? :)

Glen S. McGhee writes:

But what would be the point? Without mediating social institutions to utilize human capital thus accumulated, it would go to waste. Like all those PhD's driving taxis.

R Richard Schweitzer writes:

"Problems" or "Issues" arising in social organization (such as "Education") can not be approached (successfully) in the same way as those of the physical or medical spheres of information and "knowledge," as tackled by other X Prizes.

Wherever the enterprises of change are formed as systems, the results will be
ultimate institutionalization (the basic source of present defects that have
created education systems ) of what might develop as a series of civil
(non-governmental) instrumentalities where each segment of the social
organization deals with the motivations of its members producing and resolving
the conflicts and cooperations that produce the results.

Unless that process is regenerated (it used to exist) all else will be social
engineering; just more planning requiring new bureaucracies.

SWH writes:

From all parties involved in the K-12 education universe, the overwhelming consensus is that if you "fix" parental involvement, there is no education problem.....and if you don't, the problem is unfixable.

drobviousso writes:

SWH - Pretty convenient for those involved in the K-12 education universe.

Becky Hargrove writes:

The trick about 'fixing' education is not to think of it as a world unto itself, but instead as education relates to the actual needs and desires of a community, as it reaches out to the world for inspiration.

infopractical writes:

"From all parties involved in the K-12 education universe, the overwhelming consensus is that if you "fix" parental involvement, there is no education problem.....and if you don't, the problem is unfixable."

Parenting can certainly be a problem, but that bad parenting prevents an education is simply not true. My parenting was so horrible (no need to go into details), but I was motivated to learn, so I did. I'm far from the only case.

There are people within the educational institutions (I do work within education now) who like to use parenting as the variable that provides an excuse for their lack of success. It might be in some cases, but on the whole, it's overly used as an excuse.

From my experiences talking with many educators, there is hardly any kind of consensus about this fact, except that the teachers unions, and those tied together with them politically, seem to want the world to believe that there is consensus.

infopractical writes:

"You mean a parent? :)"

One with expertise on a wide range of topics...whose time and energy are not scarce resources.

Working on rudimentary forms of this technology now, I don't share Arnold's optimism that we're nearly there. We'll have basic tech like this soon, but it won't be all that great for a while, and when it comes to difficult or technical material (particularly math), getting the curriculum/pedagogy right is its own problem that gets less attention that it should -- and while some people think they understand this problem, we have good data that says otherwise. We'll have decent AI for some subjects soon, but the bots won't yet be particularly good teachers. What they'll be is (a) always ready, (b) inexhaustible, and (c) full of facts.

I still haven't had a good conversation with anybody's bot yet, and the idea that one will do what a good teacher/tutor/mentor does and hone in on a student's misunderstanding of the material they need or order to grip the next step -- any time soon -- is unfortunately unlikely.

Arnold is looking in the right direction, but underestimating the engineering tasks, unless by "pretty close" he means "a couple of decades" (maybe longer).

What we will have sooner than this are games that assess mechanical application of ideas. Khan does this on a simplistic level, but the problem is understood by a lot of people, and has been for a while. Right now Khan does this for the most simple mechanical tasks, but the problem gets more challenging very quickly when math reaches a level of problem solving applications. This developing tech will be a very good thing, but still limited. For our while our bots will "fake it" with artificial conversation based on results from these kinds of mechanical exercises.

Comments for this entry have been closed
Return to top