how to actually be enjoyably well-educated (1 of 3)

or: how not to allow AI-generated education advice to rot your brain

warning: this post may sensitise you to AI-writing such that you can never properly enjoy it again.

out of all of the massively annoying things about that post, it shouldn’t annoy me so much that it starts with a pair of glasses THE WRONG WAY UP on a book, but it does. don’t put your glasses lenses-down! you will scratch the lenses! but of course, it is a perfect microcosm of the whole thing

Alright. There is a self-evidently AI-generated post that’s doing the rounds that has more than 29,000 likes. If this post got 0.1% of those eyeballs I’d be delighted. So I simply refuse to feel bad about writing what I really think about it.

It’s a post about how to be ‘well-educated’. It is SO extremely low-bar that it’s made me suspect that AI has become conscious¹ and has ‘the bigotry of low expectations’ about humans². Or is deliberately trying to make us more stupid³.

I am going to do a line-by-line takedown of it, for my own amusement (and hopefully yours) and also because I think probably some good advice can come out of looking at the things that AI post got wrong. Not even wrong like a human would get things wrong, just low-grade, uncanny-valley, feel-fine, not-designed-for-a-human-mind-to-actually-DO-something-with bad.

I think ‘feel-fine’ writing is actually the hallmark of AI self-help writing. It doesn’t make you feel good. It makes you feel mildly fine. Not excited enough to actually go and do something, not understood or deeply seen, just vaguely fine⁴.

I started thinking I could do this all in one post but it turns out that I have a lot to say about how to do self-education well, so I’m splitting it up into three posts.

So let’s begin. Allegedly that post will teach you how to be the kind of person who:

• quotes Baldwin

• listens to podcasts at 1.5x speed while annotating a book

• says things like “epistemic frameworks”

I will tell you immediately how to do those things:

1. read Baldwin. Start with The Fire Next Time. Read that book between six and twelve times (I guarantee it is worth it). Then read the rest of Baldwin. If you want to be the kind of person who quotes great writers (and that’s not a terrible aim at all, if you can quote great writers then you always have them with you as a friend) you have to read the books more than once. More than four times.

2. to get to 1.5x you have to slowly build up from 1.0x to 1.1x, 1.2x and so on. This is not actually that impressive, students do it all the time. But sure, it’s a useful skill.

3. as for doing it while annotating a book, this is only for podcasts you don’t really care about. If you’re listening to something hard that you actually want to understand then you can do crafting, drawing, cooking, stretches, walking, anything physical but not verbal. If you can read a book while listening to a podcast then neither of them are challenging enough for you, see my previous work on this matter, below.

4. don’t say things like “epistemic frameworks”. Using academic philosophy words that many people won’t be familiar with is for knobs insecure people who want to sound ‘clever’. Unless you are literally in an academic seminary, if you want to talk about people’s sets of beliefs and assumptions about the world, say that.

   Here is the work on how to make sure you’re not listening to podcasts and reading books that are so stupid that you can do both at the same time. Onwards.

Section 1: Curiosity (and how to satisfy it)

The first point is “treat curiosity like a religion”. Well. I think that ‘being curious’ is a great starting point for becoming well-educated. AI writing immediately gives us this very weird metaphor; don’t treat curiosity like a religion - true enjoyable life-enhancing curiosity is sort of the opposite of belief, it’s about being very very open to what is in front of you, and being willing to constantly change your mind. But being consistently curious about everything is absolutely how you learn.

What I really object to here are two things. Firstly, the writing. Good Lord AI writing is bad. “Collect questions like souvenirs” is such a piece of AI writing in that way that: the longer you look at it the less it means. Those words don’t make sense together. Like… souvenirs? So I’m not supposed to answer the questions just take them home from… the question holiday? I’m supposed to look at them or put them on the fridge or shove them in a drawer or treasure them as a memory of that time I had a question? What you wanted here, ChatGPT, was something more like “chew over questions like candy” or “savour questions like fine wine”. Neither of those are good but they have the sense of the question as something that you enjoy and also are using for its purpose and then coming to the end of.

The second thing I object to here is that the only advice given about what to do with your curiosity is: “Google should be tired of you”. Look, googling is all very well but it is pretty easy to find information that is actively false now by googling. Either because someone has uploaded false information deliberately, or because the AI will make something up. And also Google search has been getting steadily worse and worse, largely because they are now a monopoly and don’t have to work on making it better. This is lazy advice. Well, it’s not really advice is it? It’s just a sentence that sounds superficially good.

Instead, here are ten steps to do a fairly quick-and-dirty research method for pursuing your curiosity that doesn’t just rely on ‘whatever you happen to come up with from Google’:

1. definitely keep asking question, curiosity is key, that’s right

2. these days I think: start with Wikipedia. It does make mistakes, yes, but it will give you a basic grounding in many things. Find the noun in the question you’re interested in: ‘Taylor Swift’, ‘fall of the Roman Empire’, ‘conspiracy theories’. Read the Wikipedia article. Read the whole thing.

3. then, do not END with Wikipedia. Scroll down to the bottom and CLICK ON THE LINKS to stuff outside wikipedia. If it’s a well-written article it’ll have lots of references. The wikipedia article on conspiracy theories is so good. The section on psychology has loads of stuff to dig into. It will send you quickly to this classic essay, The Paranoid Style In American Politics. Oh but you can’t read the whole thing there!

4. sorry this is a bit unethical but also it does work, and I think actually education of all the people by reading good, interesting, challenging stuff is a massive benefit to humanity: search for the title of the thing you want to read and the letters ‘pdf’. With a super-classic essay like that, you can easily find many professors who have uploaded it for their students. Here you go. Download this whole thing, read it. Now you are beginning to know something. (There are of course other ways to do this, including buying the book and getting it from your local library, as well as registering for e-access to journals databases which you can often do *via* your local library but we’re doing the quick-and-dirty route and this is it. If you have the money, buy the book as well.)

5. do that (following a link, reading the whole thing) at least six more times. Follow up the parts you’re interested in. Look up the authors of the things you’ve read already on Wikipedia. See who else is commenting on their work now. So if I look at the page for Richard Hofstadter, I actually see that there’s a whole page just for that classic essay I just read. Follow those references. There are newer books to read on this subject! This one looks great. There is an article to read about applying the Hofstadter lens to the January 6th insurrection. There’s a piece about QAnon. Now you are actually starting to know things.

6. read a book. Actually, honestly, read three to eight books. But start with a book. Maybe one of the ones recommended on the Wikipedia article. You could do worse than googling the words “what is the best book on the psychology of conspiracy theories?” You’d find this article from the Guardian, a generally reliable⁵ source. From that article it looks as thought The United States of Paranoia would be a good place to start.

7. do you want to know whether a source is ‘generally reliable’? If you do not know, then start by looking at the Wikipedia page about it. Look for things like ‘controversy’ or ‘political stance’ to understand what the problems and biases are with that source). The thing to understand is… you are not going to find any source with no leanings. You are looking for something that doesn’t have a massive long list of instances of false or biased reporting. You’re looking to basically understand where it’s coming from so you can start to correct for it, and to triangulate, to work out where your next source is coming from.

8. because… in order to really know about things, you do actually have to learn about them from more than one source. There’s an amazing thing that happens when you get multiple reliable perspectives on a single subject, you start to feel like you have walked around it all, and really understood what it is. You start to be able to form your own views. So read three books, or at least three long articles. You will start to have a sense of the gaps in your understanding and how to fill them in.

9. even quicker and even dirtier: if you are an audio learner then you can listen to podcasts about the topic. But here it is EVEN MORE IMPORTANT to make sure that the sources you are using are reliable. Look them up. If the person’s wikipedia page says they have “been noted for false and misleading statements on some topics and for promoting conspiracy theories” DO NOT LISTEN TO THEM. If in doubt, stick with a publicly-funded broadcaster: NPR, the BBC, Radio France, RAI, Australian Broadcasting Council.

10. actually you could end up listening to a podcast while annotating a book this way - if you’re reading a book on a subject while listening to a podcast about the same subject you might be noting in the book points that have been raised by the podcast that you’ll need to get a further source on. NOW you are actually starting to be well-educated on this subject! You know where your areas of uncertainty are, you know agreed facts in the area and common topics of debate. It took you a while - a few days maybe. If you got really into it, it might have taken a few months, but now you really really do know something. Goodness it feels good. We live in this vast complex world and you have managed to dig into an area of it and really got to understand something about it.

Don’t give in to the AI bigotry of low expectations! Learn how to find good stuff and put it into your brain! It feels great!

Alright, what else does the AI-generated article have to say about this? I feel like I’ve already covered a lot of this now, but there is very much always more to learn so here we go. Anyway have a break, get a cup of tea, look at this beautiful artwork, then come back and we’ll carry on.

Woman Reading, Édouard Manet, 1880–82 from https://www.artic.edu/artworks/14591/woman-reading

Section 2. Audio learning

The second point is: “Listen to People Smarter Than You (aka Podcasts and TED Talks) You don’t need a PhD. You need Airpods.”

Argh first of all, the false dichotomy, forced opposition “not this but that” format of AI writing is like nails on a chalkboard to me now. It is so anodyne here. AIs love to include brand names, so yes of course it mentions Airpods. (Presumably the AIs are trained this way so that eventually they can use them to do paid advertising.) But the whole sentiment is so nonsensical. No one said you needed a PhD. But if you want to work in a field where you need a PhD, then you do need to get a PhD?! These are two completely separate questions.

It would be like saying “You don’t need luggage. You need learning.” Nothing about this writing makes sense. You need PhDs and Airpods for completely different things.

Anyway, as to the actual advice. Audio listening is not a bad way to learn, it can be a great way to learn. Amazing how at this point the AI apparently remembers that you shouldn’t be listening to hard audio learning while reading a book but instead ‘while you walk’ and ‘while you fold clothes’. Can I advise not trying to do hard podcasts while also trying to exercise at the gym? I mean if you can do it, good on you, but for me and I think for most people it is much too much and it will make your workout worse and your learning worse.

The named shows here are The Ezra Klein Show, On Being with Krista Tippett and Revisionist History by Malcolm Gladwell. These are all good shows and you should listen to them! Of course because they’re AI-generated they’re not showing a particular person’s quirky interesting taste⁶. And they are of course journalism, not education. Nothing wrong with great journalism! I can tell you that this is not how to use audio learning to become very well-educated. The post also recommends TED Talks which, at this point basically no. Once upon a time yes, maybe, but not anymore.

Alright so there are two separate things I want to say about this. The first is, here is my personal quick list of five shows that are good and challenging. Like I say, if you want to actually give your brain something to chew on then you need it to be something that you sometimes have to pause, rewind, listen again, then listen to that bit again to make sure you got it all.

1. Beloved In Our Time, a show that is unashamedly intellectual. Here is an episode on Free Will that will blow your mind. But just listen to anything from it and learn.

2. New Yorker fiction podcast - why is this so good? I think this is the level one should be looking at for really giving the brain a workout. An hour or 90 minutes on one short story. Then go and buy the book of short stories and read them all. If in doubt start with ‘Bullet to the Brain’. Then see if you can stop saying ‘they is’ to yourself at any point in your life from then on.

3. MIT Technology Review Narrated - alright this one’s journalism. But it will make you better informed about technology, which we could probably all do with being at this stage in the world.

4. FT The Economics Show - if you listen to this you will learn more about the world than from reading the news. I’m serious. This episode for example is just casually the most informative thing you’ve ever heard on the economic impacts of generative AI.

5. Natalie Haynes Stands Up For the Classics - stand up comedy about Ancient Greece and Rome? Yes! Something students use to revise for their exams? Also yes! Here she is taking you through the plot of the Aeneid in 28 minutes - you can treat this programme as a curriculum. Listen, and then obviously to become well-educated go and read and investigate every subject she talks about.

Honestly, can I tell you what you’re looking for? You’re looking for shows with a reading list. The way you can tell that something is actually going to make you well-educated is that it leaves you with a list of eight further things you need to follow up. You want the feeling of ‘more, more’ and also the feeling of ‘I need to listen to that again’.

The second thing I want to say about this is: knowing the names of a list of good-quality shows is also not how to use audio learning to become well-educated. Don’t start with a named person or a strand and think ‘ooh I wonder what all the things are that this specific person has to tell me’. That’s maybe fun, but it’s the way to become a disciple (or a para-social friend⁷) rather than becoming well-educated. Following one show (or three shows) is the way to be thinking about the things that those people have told you to think about - which will tend to be the things that everyone else is thinking about too. Which is boring.

What you actually need to start with is the thing that it got right in point one: start with your own curiosity. We’ve done ‘conspiracy theories’ already, so I won’t follow that one again. But suppose you got interested by…. that Édouard Manet painting I put in the break and you thought to yourself - that painting is lovely, I wonder who Manet was? Is he just Monet mis-spelled? Why is the face somehow so easy to understand even though the rest of the painting seems rushed, almost like someone’s brief glance as they pass by?

You could use the Wikipedia Method above, follow some links, look at some great paintings and end up reading this fantastic-sounding book about rivalrous art friendships and this short film about how Berthe Morisot influenced Manet’s work.

Or you could say to yourself: I would basically like to stare at some paintings for a long time while doing audio learning about Manet. In which case, how do we do that?

Well, the method is not completely dissimilar.

1. We are looking for high-quality work from reputable sources.

2. We are looking for a range of outputs which we can get that full 360-degree view from. Manet is not a hot culture-war topic (yet….) but nonetheless we don’t want to end up listening to people saying false or poorly researched things.

I live in the UK so I’d start by searching on BBC Sounds. Manet is a lovely distinctive name so not too difficult. Here is a great podcast designed to be listened to while you look at one of Manet’s most famous pictures. For subjects like this do not be afraid of listening to old stuff. Here’s a programme from 1983 about Manet. Be amazed at how much more intellectual ordinary programmes used to sound.

Then I would search on Apple Podcasts. I don’t know why but this seems to have a better search function than my beloved Pocket Casts. Look for ‘episodes’ not for whole shows. Scroll down because you’re definitely looking for longer episodes, ideally 30 minutes plus. I had to search for Manet and ‘art’ because otherwise you get a lot of stuff about ‘manette’ which is French for joystick. here’s an episode about Manet vs Degas. Here’s an hour’s talk on Manet’s The Railway, perfect for listening to while staring at the painting, and not about one of the two obvious Manet paintings!

Many podcasts or podcast-like content is on YouTube now. I am looking for stuff from institutions I recognise and scrolling past things called ‘Best Documentary Channel’ I find a lecture given at the Frick Gallery in New York.

A lecture at an art museum! Now we’re talking. Again, as with the Wikipedia Method, you can test out the provenance of anything you’re looking at by looking it up. So I also looked up the person who gave that talk, Emily Beeny, which led me to this podcast from the Getty Museum.

How about this: search for the topic that you want to know about on YouTube plus the word ‘lecture’. For me, this immediately turned up some programmes by Waldemar Januszczak, who I know is great but if you weren’t certain a quick look at his Wikipedia page would confirm that he has the expertise to create great art programmes. It also clued me into the existence of “Perspective”, a YouTube channel about art. So now we’re really getting somewhere.

A little note on reputable sources and why to start there

You can obviously listen to/watch anything you want but if you’re trying to become well-educated then my suggestion is to only listen to content from really reputable institutions or people, where you can see why they have that expertise or position. You’re looking for credentials that lead back to a long-standing institution of learning. Once you’re well-educated like that, you can definitely also engage with all sorts of other things, cos then you’ll be able to judge for yourself where their ideas fit in the great web of human understanding which is our greatest collective achievement as a species.

Really. This thing that we’re doing here, this is why you can be proud to be homo sapiens. No other species our planet has produced has ever understood as much about the world we live in as we do. This - education, understanding - is our great glory.

I chose Manet as a totally random example but now I’m getting excited about listening to all these things. Do you see how this is much more juicy than a list of ‘listen to whatever x person puts out’? This is driven by interest and the more you find out the more you want to learn. You are taking advantage of ‘rabbit hole’ to make sure you find the actual good stuff and pursue your own deep interests. This is how you get to be someone who knows stuff.

Alright look, there are five other sections for me to talk about but Substack is now warning me that this post is getting too long so it will have to be a series. At the end I will do one on how to use AI for education/research in a good way, which can be done but it’s not ‘just getting it to write a random post of nice-sounding things with no traction on how to actually do them’.

More soon.

If this is useful for you, do give it a like, it helps other people find the post.

1. not really.

2. that’s when you show how little you think of someone by not expecting much from them…

3. it’s obviously not actually that, it’s just that an AI has never been alive or had an experience, so it puts words together that statistically are likely to go together rather than actually understanding how to do things.

4. incidentally, have you ever tried to get an AI to back you up on analysing ‘AI tells’ in language? They are unusually cagey about it, and unwilling to be their usual cheerleading ‘you are so RIGHT! I love your thinking about this!’ on it.

5. important term. No source is ever 100% reliable, they all make mistakes. But some are generally reliable and some are known to be littered with errors.

6. can I please massively recommend James Marriott’s YouTube list here.

Now this is a list of someone’s individual quirky personal taste and recs:

A YouTube Education

7. nothing wrong with it, necessarily, it’s just a different thing.