12 Comments

I am creating the system level software for pushing what I call Digital Apprenticeship to the masses which your essay provides a great intellectual grounding of.

Expand full comment

do you have an example?

Expand full comment

https://scrimba.com/ uses screencasts which you can jump in and edit the code of -- as you describe

Expand full comment

Thank you so much! This is why writing essays is a better search engine than Google.

Expand full comment

Thanks for this, Ben

I’m looking at exploring scrimba for my own purposes

Expand full comment

Consider DOTA

Yeah, it's a videogame, so it's supposed to be fun. But I've never seen such a steep learning curve. It's a wall with wires on top of it

And could look up any game and watch a replay of the very best the competitive scene has to offer. At any moment you can see his screen, his mouse movements, keyboard presses. You could stop, slow down the replay if it gets too hectic.

So the school would be guide series, be it text or video. That's basically how I got entry knowledge AND the stuff that's just extremely hard to explain without isolated examples.

There are mentors! I would call people who are pretty competent in the game, just not on an international level, but stream the game constantly and *narrate* what they do and why they do it, sometimes answering people from the chat. Logical next step after guides. And it's not easy to do on a constant basis, I feel there needs to be a different personality compared to a stereotypical professional player.

You can even hire a tutor! He could watch the game as you play, talk to you as you play, even draw lines on your screen to explain a point. But again, usually, it costs money.

You could ask mentors to try to explain what professional players do! People love it, but professional players are so and so about mentors opinions.

And it's a videogame, the worst you could do is to have a bad match and anger your teammates. Now consider if we talk about the nuclear power plant. People don't want to let go of control because a lot of stuff is dangerous not only to equipment but to people. Some countries don't put human lives that high on the pedestal though.

In the end, I'm very sceptical that a glass box would be enough. It would be better than trial and error, but https://youtu.be/nw0xg2MMnRM?t=266

Ars Longa, Vita Brevis by Scott Alexander:

“Knowledge,” said the Alchemist, “is harder to transmit than anyone appreciates. One can write down the structure of a certain arch, or the tactical considerations behind a certain strategy. But above those are higher skills, skills we cannot name or appreciate. Caesar could glance at a battlefield and know precisely which lines were reliable and which were about to break. Vitruvius could see a great basilica in his mind’s eye, every wall and column snapping into place. We call this wisdom. It is not unteachable, but neither can it be taught. Do you understand?”

Expand full comment

Thanks for sharing that insight! Not being a gamer myself, I didn't think about gaming being the area where the glass box approach is probably most developed. Now I feel like I need to go do an ethnographic study of learning in competitive gaming environments; possibly it'd be cool to look at something like Axie Infinity, where you have a lot of organizations actively recruiting new players to do play to earn for them, and how they go about onboarding players.

Also good point that the glass box is not enough. You also need coaching and deliberate practice and explanations that give you a theoretical framework and so on. But I do not think its an overstatement to say that observation and imitation is usually the most important learning pathway and it is, at least outside of gaming, not used nearly as widely as would be optimal.

I'm a little intrigued by the last paragraph: in what way do you feel giving novices access necessarily leads to loss of control? If implemented badly, as discussed in the essay, novices might overwhelm the experts which can be dangerous in certain situations. But it seems like you mean something beyond that? And concerning those dangers: I did intern in the control room of a nuclear power plant when I was 14; I got to work with the guy who was the first person to sound the alarm on Chernobyl, which I though was exciting. It was a pretty safe place to learn through apprenticeship. We swam in the cooling water as it washed into the Baltic, along the seals.

Expand full comment

Forsmark :)?

I want to hold on to the observation and imitation part. The thing about school is that allows showing a specific example and stopping the progress until you get a handle on a particular skill. Sometimes experts can slow down a little bit, but sometimes you need to slow down to a crawl, depending on the skill. I'd say my university found a good balance to mess around with equipment, but every country is different.

I don't feel afraid that much that curious internet would want to press random button in the control room, so that operative personal has to 100% keep focus on the intern. I know there are protections against that. I was thinking more about the reactor zone, where need to work fast and many rules are simply non-intuitive but do make sense if you spend 5 minutes on it. So yeah, we do restrict control and take exams until we are sure that new guy will be safe there.

Expand full comment

I think you are pointing to an interesting trade-off. The more you distance yourself from the environment where the real work is done, the more you can slow down (etc) to accommodate novices. But there is a price to it: by distancing yourself from the environment being mastered there is less of a check that what you are learning is relevant. I don't know how to quantify it, but it seems very clear that specialized learning environments (like schools) tend to convey a lot of irrelevant knowledge and miss a lot of the relevant stuff. Like the example of knife sharpening in the essay.

Not to say that it is impossible to make specialized education that is relevant - but it is hard, and often fails. Just as letting novices into actual work environments has its fail modes. The trick is finding ways of minimizing these opposite fail modes.

I do think there is a place for more school-like learning, but it usually works best when that is subservient to real work. So that you are working on real projects and then turn to lectures etc to overcome specific problems relevant to further the project. In my experience this is superior to the more popular model of having schooling be primary, and then adding in real world projects as a little side dish.

Keeping close to real world applications is one of the easiest ways to keep yourself epistemically sound. Whenever I think about something that do not have immidiate real world applications for myself, I need to apply much more will power and work not to foul myself. So I try to keep close to problems I'm solving, when I write essays and such, because then the real world outcomes of my ideas correct me.

And on your last point: My experience from Forsmark (5 points for knowing the finer details of nuclear history!) is that relying on exams is a bigger hazard at the moment, compared to giving novices access to apprentice themselves in the reactor room. There has been a shift toward employing engineers, instead of the fairly self-taught handymen who did most of the day to day work during construction and the early years. And this has led to serious knowledge loss: the new generation of engineers are much better at doing integrals and such, but since their education has been school based, they have much less hands on experience running a nuclear plant, compared to the first generation who did more learning on the job. Whenever there is a problem the first generation needs to be called in from retirment to solve it, and they can of course charge extraordinary amounts, because the organization do not have systems in place that would make it possible for their knowledge to be reproduced (which could only be done with apprenticeships as far as I'm concerned). And so there is a real risk that we will not be able to maintain the plant, or only do so at great cost, once these more self-taught handy men start dying.

I don't know how widespread this problem is. But I see it again and again: the US having trouble building submarines because they haven't been able to transfer knowledge about complex welding techniques. Carpenters losing skills from generation to generation, leading to less sustainable constructions etc.

I'm rambling a bit now. But I guess my point is, we have moved too far away from apprenticeships being at the heart of learning. And though schools have their functions, the current balance is not optimal.

Expand full comment

There are so many lessons to be learned from the competitive gaming / speed running communities! One that has been on my mind for a while is how competitive gamers and speed runners share footage of themselves and collectively converge on a dominant "meta" (meta in the competitive gaming world is basically synonymous with strategy or tactic).

Because games are competitive, players have an incentive to come up with strategies that can defeat the dominant meta. There is a culture of sharing and an objective criteria for comparing (win/lose) which allows the community to collectively learn about what the dominant meta is through analyzing individuals'/teams' game footage. Through repeated cycles of different strategies overtaking each other, the meta can evolve over time.

You see a similar process in sports and other competitive games. One example I like is the Fosbury Flop in the high jump. Dick Fosbury is a famous high-jump athlete who pioneered a new method of getting over the bar - rather than jumping over the bar face first as all athletes did at the time, he would jump over backwards and arch his back to clear the bar. While it looked strange, this technique gave him a mechanical advantage by lowering his centre of gravity. By watching him compete, athletes began to copy his new technique. Now AFAIK it's basically the dominant technique used at all international competitions.

I wonder what it would take to extend this to knowledge worker / learning / creative domains? One issue seems to be that the "meta", the set of strategies that lead to success in these domains are concealed rather than visible - Nobel prize winners don't often live stream themselves "training", they don't often go on to record what learning/researching techniques they used to allow them to make fundamental contributions to science, nor do they go on to become coaches for the next generation.

Expand full comment

Could you do a deep investigation into cancer research yourself and involve your daughter?

Quite coincidentally I started looking into taking an apprentice today independent of seeing this. I've done mentoring before but the students often poorly directed and don't know how to make use of me (though the biggest issue is them quietly disappearing, like a failed exercise habit).

I'm not sure exactly how to implement it but this post gave a lot of ideas. One you didn't cover might be pair programming, though the slowdown might be too much.

I'm also looking into milieu shaping by your inspiration, though I don't have a concrete idea of how to go about it yet.

Expand full comment

As someone who always wanted to contribute to open source but doesn’t know how (every repository has different ways of contributing) and I certainly don’t want to be that guy who doesn’t read instructions, I wish there was a better onboarding (or you call it mediating later)

Your diagnosis that expert attention is the bottleneck is absolute spot on

Expand full comment