Hacker News Reader

Comment by tengbretson | Hacker News Reader

It's odd to me that there is this bizarre certainty on hacker news that within 20 years the most common occupation, truck driving, will be nearly entirely obsolete, yet in an era where almost all human knowledge is available online for free, the idea that the education apparatus will either stay the same or continue to grow is somehow a given.

View on HN

It's almost as if pedagogy is a deep and rich field and you can't as easily absorb information without the guidance of a knowledgeable mentor. Who would have thought?

I don't know, after a few months of home schooling, I'm pretty sure that mentor could be an AI.

I really can’t tell if this is sarcasm.

AI is probably far fetched, but gamified learning where you have to imput your result and you immediately see the correctness is already possible.

I think you could go much further. We should work out how to monitor and test a students progress, push them harder if they are ready, or roll back a bit if we detect they missed something. On Khan Academy where everything is multiple choice, my kids started guessing rather than knowing the answers. It was up to me to notice the problem and take things back a step, and work out where they got lost.

> We should work out how to monitor and test a students progress

What you're describing is called "teaching".

Sounds like you need to seriously consider how well you're doing.

Yes, this could be a good opportunity to revisit how education is done.

I have a 6 and 8 year old and we had about 8 weeks of home schooling here in Australia. We took the time to check out a wide variety of online learning resources.

Khan academy is great, but a little dry. Some other things have a little to much flair and not enough substance. I think there is a lot of great work to be done here.

Professionally made resources with tests and assessment built in could make a big impact. Student's can learn at their own pace, and I think most of it could be made just as "fun" as any video game.

The hardest problems will be "social". Determining when a student is lacking motivation, or when they are confused. How does and AI encourage them or clarify things in a way that's not more demotivating.

Actually, I'm sure the hardest problem is political, but putting that aside, building an online platform for years 1-6 would be really interesting. Years 7-12 would actually be a lot easier I think.

That's because if you know anything about education, you know it's about human relationships as much as it is about pouring knowledge into empty vessels.

I actually think the are two separate concepts, and I'm pretty sure a AI could teach kids the times tables and test spelling.

My kids know a hell of a lot about the video games they play, and they didn't need a human relationship to tech them how everything works.

Tech can definitely teach some things to some people, but among other things it cannot create the context for teaching. This issue has been pretty well explored in the edtech community for a long time, and of course pedagogical research for much longer. If you're interested in learning more check out sources like EdSurge, which I started as a place to explore these issues.

Yeah, I don't think we need to say, all education will now be done by AIs, but I am a little surprised how under developed a lot of the software was we encountered in our home schooling adventure.

Totally. Part of the issue is there's at least two parallel markets in K12 edtech - school & home. I know there's lots of good stuff available for schools that just isn't designed for or available for home use. Meanwhile lots of the home stuff you find in app stores is more entertainment oriented.

> where almost all human knowledge is available online for free

And it has been so for nearly 20 years and changed very little. MOOCs are a failure as people will not finish them. Online certificates carry no weight with employers even if people do.

We have promising prototypes of self driving trucks. We don't have such models for education.

We have plenty of promising prototypes. The original MITx / 6.002x had in the nineties rate it as being as-good-as or better than in-person.

OLI before that had good prototypes too.

The low-quality junk came in scaling. Part of that had to do with culture and people (business people took over from academics). Part of that with economics (people didn't shop on quality). Part on difficulty training. Etc.

The problem is not in information quality. I agree that is a solved problem. But that was never the big challenge in education (besides the politics of it).

The problem is in getting students to complete the work, be honestly assessed about the work, get feedback on their work as well as for employers to believe that the assessment means anything.

> The problem is not in information quality.

Information quality can be interpreted many ways. I wouldn't say most teachers are good at their job or at presenting information. Most I had were not good at the job - a lot of them were babysitters and acted as such. At the college level, it's possibly worse where the people are so consumed with their research that they have absolutely no interest in teaching or getting good at it.

Textbooks are also very poor quality. History textbooks are probably the worst offenders.

I was mostly referring to why MITx isn't a solution to education.

I agree with respect to the current education system.

History is hard to correct as it is politically thorny and frequently disputed. If I built a e-learning system, I wouldn't include that.

Most of those are solved problems, technically. What's not solved are, as you put it "the politics of it." Business models don't align with good models. Markets are highly inefficient, timescales are long, and transparency is poor.

MITx became edX because of that misalignment.

There are promising models around learning and plenty of resources but I think the incentive is underdeveloped. This is not mentioning economic and political factors, but from a technical standpoint, the means are there.

Not necessarily. MOOCs and the opportunity to access free and open education are widely consumed by people outside the US in emerging countries.

I never seriously considered home schooling as an option. COVID has taught us otherwise. I’m pretty sure we’re all in on it for next year.

Five years ago there was the widespread belief that in 5 years most cars would be self driven. This was in fact the main reason for investors to give Uber extreme high valuation. As we see in this case, and many others, technology doesn't advance as fast as some of us wish.

It is absurd to equate something that requires human empathy and connection to transport.

I doubt truck driving will obsolete, given the track record of self driving.

Reply on news.ycombinator.com