Generative AI takes away the opportunity to develop taste

I was listening to a recent episode of The Vergecast and Nilay Patel said something:

We write sentences for a living, it's hard. It was hard to write sentences when we got started, and you had to be good at it. There's a part where [AI]'s robbing people of the time it takes to develop a craft… I can see the laziness in the craft and I think that's actually very bad… The idea that everyone was going to become a great artist or a great writer because they had the AI tools, it won't happen because nobody will have taken the time to develop their taste. (YouTube video with timestamp, starts at ~47:35)

The first comparison that comes to my mind when I hear this is the well-known piece from Ira Glass about "The Gap" between your taste when you first start making something, and what you're actually able to make. That when you start making things: "they're trying to be good, they have ambitions to be good, but it's not quite that good".

Or take journalist Ryan Broderick who, when reflecting on a vibe coding experiment said (paywall):

my mind tends to treat writing an article, making a video, writing a song, cooking a meal, drawing an image, and, apparently, designing software the same way. It’s not a matter of just “generating” something perfect from my head, but exploring the tension that exists between what I’m imagining and the limitations of my stupid meat body... Vibe coding, like every new trend coming out of Silicon Valley, turns this process — the entire act of creativity, itself — into a slot machine. One more pull on the AI and maybe it will figure it out for you.

It's good to state the obvious sometimes: not everyone is equally good at everything and not every activity is equally enjoyable. I'm an awful visual artist, and I simply enjoy nothing about it. Some people are exceptional fiction writers, but the vast majority of us are not.

It's strange that two people can go through broadly similar levels of education, and even achieve identical recognition or qualification, yet still have meaningfully different abilities from a subjective experience.

What do I mean by that? Think back to the people you went to school with, or people who last did any kind of thing with you (course, qualification, project), and ask yourself if a) all of the marks awarded to every participant corresponded strongly to "ability"; and b) if people who got roughly the same marks had the same level of interest or taste for the subject.

I for one have grown a lot more appreciative of maths, but sixteen year-old me absolutely did not care. He still did okay at secondary education levels. I left my undergraduate with a substantial interest in computers and statistical modelling, and others left the same degree with an interest in climate policy or financial regulation. Even if we arrived at those interests semi-randomly, our long-lasting enjoyment of them came from choosing to repeatedly engage with them.

The internet is too full of people making hyperbolic statements like this, but this one is true: I genuinely find this fascinating. I have a doctorate in education motivated in no small part by this question. How do you get (or can you get) someone to engage with a subject in a way that helps them build a "taste" for it. If areas of knowledge have an "essence" - how do you get people to see how cool that is?

I'll give you an example: I watch a lot of contemporary dance and ballet. And there's a feeling you get when a piece is well constructed - the geometry, the flow, the orientation. They all point in the same direction. I am willing to bet that functionally every single working professional dancer or choreographer has had that feeling and wants to achieve it. I am also willing to bet that more often than not, they won't. But why else would you do that job if you didn't at least want to achieve that feeling? It's just... hard to do.

You listen to a record, you read a particularly well-written article, you hear a good interview. And you think "that's good".

Or maybe you hear an awful mix-tape demo, or read a bland article from your local authority, or hear someone ask "how are you so fantastic?" in an interview. And you think "this is bad". Or worse: somebody's creation slides off you like water into the river and you never think about it again.

There's a million ways that people become better at something. That our tastes get refined, as we learn to recognise and then name the flavours in wine that we love vs. those that we don't.

No matter what you're making, or building – the taste and abilities that we're building improve because we engage with it. What's more, as we engage with things more we start to live a more fulfilled life. No matter if it's pottery or family ancestry — go deep on either of those and you'll probably find your life enriched in some way.

To put my tinfoil hat on for a second: I worry that we start to conflate interest with consumption or engagement with a product. That we flit wholly between chess.com, Duolingo, and goodreads.com. And the marketing teams there would love your attention (or money). Alright, tinfoil hat off.

Generating, viewing, and then discarding huge amounts of something from a generative AI technology doesn't help you get anywhere, really. At least no more than watching every Oscar-winning movie at 4x speed on 2 monitors at the same time will make you 8x "better" at cinema in the same amount of time.

If you're looking to do a thing, and you don't have the framework to engage with it, you'll be stuck with a sense of "no that's not quite right". And the only tool the generative AI will give you is a button that says "regenerate". And the company that sells you the generative AI service would love for you to hit that button. OH NO — MY TINFOIL HAT, I DIDN'T TAKE IT OFF RIGHT.

If you really, really cared about building taste you would use libraries. I mean public libraries - physical buildings full of organised information and people who are paid to help you access it.

And we have libraries. And nobody is out here writing headlines, raising billions of dollars about the place you can go (FOR FREE) to learn mostly anything. Probably because they're old, or because it's hard work.

There's something here about how the appearance of expertise is increasingly getting mistaken for actual expertise. Calling yourself a journalist does not make you a journalist. We simply don't have time to go into that today.

Look, I think it's worrying that the generative AI has let us rob ourselves of the chance to engage in the difficult part of problems. I think that having a discerning taste in something is extremely rewarding personally, and possibly professionally. I worry what happens when that gets taken away, and replaced by a private, for-profit service.