Like me, he is a fan of Neal Stephenson’s illustrated primer as described in The Diamond Age. Diamandis writes,
Your AI companion will have unlimited access to information on the cloud and will deliver it at the optimal speed to each student in an engaging, fun way. This AI will demonetize and democratize education, be available to everyone for free (just like Google), and offering the best education to the wealthiest and poorest children on the planet equally.
Stephenson’s primer is a great idea, but his idea of multiple, voluntary, non territorial cultures is also illuminating.
How does he define “best” education? What mechanism will he employ to stimulate competition among providers to generate new and better approaches? Google isn’t really “free.” Is he suggesting the education service will be ad supported. Wasn’t there a controversy several years ago when some sort of ads accompanied educational materials?
The definition of “best” education will be decided by millions, or billions, of people in the marketplace. Thousands of trial and error attempts, along with the verdict rendered by the market will help entrepreneurs optimize the process. This is already underway. There are hundreds of web sites providing low-cost education already. Not all of them are good, but the best ones will grow and dominate in the future.
A think it is one thing to complain and object to advertisements in today’s classrooms, which are a government monopoly. (That is where those offending adds happened, so long ago). Offering people free or low cost education in the future will be an easy sell, if people know they can avoid spending huge amounts of money to get a college degree.
People already have free or low cost education in K-12, yet students cheer when school is cancelled. Offering education is not an easy sell. On the other hand, if you are offering certification of Bryan Caplan’s trinity of intelligence, conscientiousness, and conformity …
“yet students cheer when school is cancelled.”
I’ve always thought this was a misrepresentation. In my experience, if a teacher constantly cancels classes, the students will get upset with her. This is also true if the teacher is very poor at conveying the information. There will be a lot of complaints.
Cheering when a single class or day is cancelled is just a recognition that there is some slack in the schedule, and we can take a holiday with minimal ill-effect.
You are correct. Most kids hate school (I didn’t much care for it).
I made good grades in high school while others barely even tried because making good grades, graduating high school and going to college were just assumed to be normal in my family. It really did not occur to me that not doing those things was an option.
Telling people that a path is open does not send them down the path. People have to believe it is the only path because the best path isn’t the funnest. The rewards come later.
Question: What if this AI were also a personal trainer? It could give you workouts to do and lifestyle changes to make. It could tell you how many calories the food you were about to eat contained and what you would have to do to burn them off. Would that lead to a population that was physically fit and not overweight?
With well beyond 95% certainty, I predict it would not. It would be great for the people who are motivated and able to make themselves “defer gratification” (or if you prefer, don’t have a “present orientation”). However, for many people the effect would be minor. And at least in America, there would be a large overlap between the people who would use the AI personal trainer and the people who already are moving and not overeating.
I predict with the same certainty that an educational AI would have the same kind of effects. In America, lack of “education” is mostly a matter of demand, not supply.
Or to put it another way, an AI companion will not defeat the Null Hypothesis.
(On the other hand, the results in non-WEIRD countries might be different.)
“This AI will demonetize and democratize education”
I know what people mean when they say that something is democratized, but it still makes me twitch when I hear it. That’s not democracy at all. In fact, when someone describes greater accessibility to a thing as democratized, it’s usually the quite opposite of democracy.
If it equalizes access to education, remaining differences will be due to differences in ability to benefit from education. That’s basically a mix of IQ and conscientiousness (and probably minor contributions from a few other things like health). There’s no particular reason to think that social competition would get less intense or that life would get better for those on the bottom; the smart and conscientious among that demographic are few and far between.
That already exists mostly. Standardized tests scoop up the diamonds in the rough. Most people trend towards life outcomes that equal their genetics. It’s not perfect, there are some advantages and disadvantages from environment/upbringing (though many don’t show up in SES statistics, more things like family stability, comfort in social settings, freedom to take risks).
However, would you want to eliminate ALL advantages? Most people are motivated by wanting to make a better life for their kids. Given that 80% plus middle age SES isn’t effected by environment, shouldn’t we leave parents to scramble over the last 20%? I think we are at diminishing returns as far as equality goes, even after adjusting for genetics. Especially since further harping about inequality always gets channeled into attempts to brute force around genes, with bad results.
Yeah, that’s pretty much where my argument was going. But I don’t see it doing much damage if things work out like Diamandis predicts, either. The remaining 20%* isn’t a huge deal, usually.
* using your estimate
The AI agent in your hand knows your wants and desires. That information gets embedded in the market, as you buy and sell. All the AI agents connected in a network via a pure bearer asset automated trading system. The network itself becoming a singularity, as it re-arranges the world of goods such that all of us equally represented. The network always seems to know a bit more than individuals about who and what we are.
Anything that will compete with the ossified, cronyist degree and credential printing industry has got to be a good thing.
But as more and more of our lives gets sucked into online activities, this drift into AI may open up other opportunities for purveyors of offline experiential education. So many kids today have never worked with hand tools, never changed their own oil, or even a flat, never set foot in a barn, or killed, skinned, and cooked an animal.
One of my regrets in life is that as a parent I was remiss in getting my kids into basic automative maintenance summer classes. Tried, but never worked out and I never showed them how to change the oil on a car. And I really wish they had gotten at least some practical exposure to electrical and plumbing work. I’m guessing somebody who could organize a gap year or summer school program in which kids spent time building a house, working on a farm, working on cars, and learning small engine and appliance repair would catch a nice ride on the backlash against current trends in the higher education industry but also make some money and give kids some confidence and understanding about how the world works. And a little bit of time away from devices would just be icing on the cake.
My dad taught me some of that stuff, and he liked buying me tools. I miss that. He tried to teach me how to skin a rabbit but I was having none of it. You probably taught them more than you think you did.
I think what has really changed it for me has been Youtube, which is related to the post here. I’ve repaired a lot of stuff by watching it done online. Refrigerator. Washing machine. Auto side mirror. Lawn mower. Plumbing. You name it. People put the damndest things up. I found one YouTube for my particular factory defect in one model of Samsung refrigerator. I go through the steps about once a year and have saved countless hours of aggravation.
I’m motivated. By saving money and the time spent waiting around for some other guy to come do it. I think if you find motivation within yourself a lot of it is right there waiting for you. Nobody has to take you by the hand anymore.
I imagine if we had a free-ranging AI it’d be most used by people who were already motivated to learn. I doubt you could use it as the focal point of widespread education because most students won’t be motivated enough to use it properly.
However, I think partial AI could be very useful for expanding student-teacher ratios in traditional education, which might make it one of the most pragmatic ways to alleviate the cost-disease in education.
Anybody who wants to talk knowledgably about this should probably spend some time learning about Sudbury schools and their peers. It’s an imperfect proxy given that they tend to draw from high SES demos, but they still seem relevant.
https://en.m.wikipedia.org/wiki/Sudbury_school
the future is mental uploading.
A free K-12 education on the web for everyone to use, shouldn’t be that hard to achieve.
But I would not trust an “AI companion.” Like today’s news sites, all of them would be biased. Maybe or maybe not all in the same direction (depending on who finally gets control of the Internet’s DNS system, now that it’s in UN hands, or creates substitutes for it).
As for Stephenson’s world, it has some truly nice features but doesn’t appear defensible. At least the way Atlantis falls at the end of the book could not have been prevented in any obvious way.
Access to information is already here. There’s tons of free content online or very cheap paper textbooks.
Normal humans need social context; they need to see human peers and human coaches.
Consider a physical fitness analogy: quality workout videos are free on YouTube. Normal people need human peers and coaches to be motivated to show up.