Learning from a Computer the Way Socrates Would

Martin Rezny
Words of Tomorrow
Published in
5 min readOct 24, 2017


How a single person can learn alone in a room with a computer


What a wonderful and entirely practical conception of learning using new technologies — learning to actually solve problems using the wealth of knowledge available on the internet, instead of using computers in the classroom only to play movies, the way a caveman would use electricity exclusively to light campfires. Also, it should be more surprising that schools resist this provably effective method, but alas, it of course isn’t.

Apart from a major kudos to Mitra, what I’d like to add is that in my experience, it doesn’t even require a hive mind to learn from a hive mind. I’d say it doesn’t even require a computer that’s actually online, though that should by all accounts accelerate the process tremendously. On the level of principle, the most genius aspect of Mitra’s approach, at least in my opinion, doesn’t need computers at all and could be applied universally.

The truly brilliant part is the teacher taking a hands-off approach, letting go of the steering wheel in the classroom. What Mitra calls an “attractor”, which is a great way to conceptualize it, is the core of what makes the Socratic educational method work. An ultimate goal that’s challenging and not entirely specific, especially regarding the trajectory of how one’s supposed to get to it, is probably the best way how to motivate learning.

Ultimately, I’d argue that computers specifically, provided without an instruction manual and operating in a foreign language, are the best technology to pair with this approach, as by their very nature, they generate indirect, un-obvious attractors by themselves. Without a computer, a Socratic teacher can provide attractors and answers to factual queries the way a search engine would, but computers are also puzzles.

Natural Learning by Interfacing

To venture a guess at how such emergent learning may work in individual cases, at least to some extent, when you look at children as learning machines, you’ll see they’re hardwired to interface with the surrounding reality by interacting with it, from genetics to memetics. They’re supposed to learn language essentially by osmosis even if nobody is specifically teaching them to do so, and, from not only my own personal example, I know that this does sometimes include self-learning to read and write.

This process seems to be spontaneous, driven by curiosity, which means that the best way to kill it is to impose boring, unreal structure all the time that punishes creative interaction, i.e., schooling. Computers are in this sense a reality-in-a-box, the structure of which is geared toward maximizing the intensity and effectiveness of creative interaction and spontaneous exploration. A computer placed in front of you doesn’t tell you how you’re supposed to explore it, not directly or overtly, as it is itself a puzzle.

When you put a puzzle in front of a kid, you don’t need to tell the kid what he or she should do with it. If anything, it makes more sense to use reverse psychology like Mitra and say that it’s too hard to solve, anyway, so it’s fine if they don’t even try. The difference between solving a regular puzzle and “solving” a computer is that when you solve a puzzle, you maybe train your memory or pattern recognition, but when you “solve” a computer, you can learn languages, mathematics, coding, any number of facts, and even arts.

I can now see exactly what by sheer coincidence happened to me that lead to me learning so many things without instruction. I was given a computer when I was nine, mostly in English and with a text-only DOS interface mind you, and no one in my family or the general surrounding area understood either. Even if my parents wanted to instruct me, they couldn’t. I was just left alone with it, without any internet connection or manuals, armed only with floppy disks and with what my friends could similarly discover.

Through just that, I have learned how to use any software from word processors to video editors and how to install hardware, English, coding, and musical composition, and those are just the apparent, easily identifiable skills that most people would agree are constructive. People around me have always considered most of these to be impossible to learn even with instruction, certainly to the level to which I’ve mastered them.

The Blessed Ignorance of Limits

With a computer, a programmable thinking machine, anything is possible in a very practical sense. Including, undoubtedly, many things that are at any given point thought to be impossible. New tools can be created, new forms of organization discovered, and any information known by humans or theoretically representable can randomly mix with any other information to produce a breakthrough. My current favorite is concept-mining on the Library of Babel, where one can encounter all possible English words.

Check out some of these gems of philosophy and literature produced by sheer random — boobook, introsuspection, caffeinism, octopush, vociferate (noun, as in a type of regime), exceptless, afterbrain, stasimorph, interpoint, redeath… There’s literally infinity more of such thoughts unthought to be discovered there, and only there — in a computer. You tell me what these mean, or better yet, explore for yourselves and find more.

I hereby coin this approach to writing “iterature”. Think about it.

I could try to dance around the issue to be polite, but truly only an idiot wouldn’t see how much more of a potential there is to learn from a machine that could write books itself, rather than just from books, and I dare you to convince me otherwise. Heck, the Library of Babel alone actually contains all books that will ever be written in English language, in a way, and the internet as such contains very nearly all that currently exist in all literary languages. Denying access to that while learning is nonsense.

In a true Socratic fashion, I believe it’s important to recognize how little we know, especially as teachers regarding what is or isn’t possible to learn or how. To decide the limits of what someone else is supposed to learn and exactly how is arrogance of the highest order, even before you consider that it effectively dissuades learners from trying to learn by discovering new solutions and problems, including new ways to learn. Or, in other words, actually learning. Late great Terry Pratchett also seems to agree:

“…it is well known that a vital ingredient of success is not knowing that what you’re attempting can’t be done.”

In Equal Rites, the book where this quote comes from, a young witch manages to teleport herself without a counterweight simply because no wizard was able to explain to her beforehand that it can’t be done. Educational system based wholly around teaching things and using methods that are known effectively prohibits any improvement that has to be stumbled upon. Which is to say, in all probability, vast majority of all meaningful improvement ever. I wonder what teachers would say to this.



The Casual Fascism of Strategy Games

7 min read

Feb 18, 2016

Game Design with the Help of Social Sciences

7 min read

Feb 27, 2016

How to Make Mobile Games That Aren’t Brain Poison

7 min read

Dec 6, 2017

How to Unruin the Diablo Franchise

19 min read

Jan 21, 2019

Why I Won’t Play No Man’s Sky

10 min read

Aug 28, 2016

For Tron’s Sake, Get Out of my Head

8 min read

Jan 1, 2017

The Advanced Math of Choice

9 min read

Feb 5, 2016

Fixing the Game of Education

8 min read

Mar 10, 2016

You Have No Choice

6 min read

Feb 4, 2016

Why New Master of Orion Is Already Old

10 min read

Sep 16, 2016