šŸ‘½ Lee Tusman

ā† Nosebook šŸ‘ƒšŸ““

What I want to tell my students about AI

30 Nov 2024

Six-minute read

update: added some further thoughts at the end

Not long ago, in an online listserve Iā€™m part of, a digital artist announced they were working on a retro video game, programmed in an emulator, creating a game in that systemā€™s original language and operating system, a 40-year old game console, the Sega Genesis. ā€œIā€™m not much of a musician,ā€ they said. ā€œHow is..ā€ and then rattled off some machine learning models to find out whether folks recomended any particular ones for generating game soundtracks.

Art is about making interesting choices. If youā€™re trying out writing a game for an ancient platform, in an old language with particular constraints, thatā€™s a choice the artist made to try out something, outside their normal environment, to see what else is out there in the world, and to challenge oneself. Itā€™s also a part of learning a tradition and seeing what working in an older pattern or technique still has to teach us in the contemporary moment.

About a decade ago I felt like there was a lot of public discussion on ā€œ10,000 hours,ā€ a stand-in for investing years of time and experimentation that one puts in to master oneā€™s area of focus. It was widely understood that one needed to ā€œput in the timeā€ in order to reap the reward of being skilled at your craft.

And it reminds me of podcast host Ira Glassā€™s famous pronouncement that bears repeating:

All of us who do creative work, we get into it because we have good taste. But there is this gap. For the first couple years you make stuff, itā€™s just not that good. Itā€™s trying to be good, it has potential, but itā€™s not. But your taste, the thing that got you into the game, is still killer. And your taste is why your work disappoints you. A lot of people never get past this phase, they quit. Most people I know who do interesting, creative work went through years of this. We know our work doesnā€™t have this special thing that we want it to have. We all go through this. And if you are just starting out or you are still in this phase, you gotta know its normal and the most important thing you can do is do a lot of work. Put yourself on a deadline so that every week you will finish one story. It is only by going through a volume of work that you will close that gap, and your work will be as good as your ambitions. And I took longer to figure out how to do this than anyone Iā€™ve ever met. Itā€™s gonna take awhile. Itā€™s normal to take awhile. Youā€™ve just gotta fight your way through.

I donā€™t want students to do busywork. There is a power differential, so Iā€™m giving assignments and have authority. Itā€™s hard to upset that hierarchy, though I try. So believe me, I get that using tools to eliminate busywork is a strategy.

But trust me, or trust Ira, you cannot get good at making games, music, storytelling, website design, or whatever other creative or technical or studio practice weā€™re doing unless we put in the work.

The rise of the word ā€œcurateā€ to mean select is of a piece with using paid-for machine learning models to spit out lookalike art assets. If you think you ā€œcurateā€ people at a party or a soundtrack for a car ride then sure, it makes sense to externalize most of the work. But selecting is not the same as creating new original ideas.

If you want to get good, to make your own mark on the world, to make meaningful, compelling work, then you need to work through the challenges, feeling lost at times, and the open-ended exploration that can happen. In fact, these are necessary to be able to break through with your own vision, your own fresh voice.

When I was first getting into biking I was fearful of getting flat tires. What happens if Iā€™m far from home or a bike shop? Will I be stranded? I read a book on bike repair that included fixing flats. I watched some videos online. I watched someone do it in person. But then I myself had to try it. I didnā€™t want to have to go to a shop and pay someone else to do it for me every time for the rest of my life. And at first I wasnā€™t amazing. Patching isnā€™t equivalent to making art, but letā€™s go with my metaphor for a moment. Most of the time my patching worked fine and I was back on the road. Once or twice my patch didnā€™t take as I didnā€™t have the proper technique, and re-flatted minutes later. But I was okay, and putting in the practice, going past the fear of failure or confusion makes all the difference. Now I know how to do it correctly each time. Iā€™m not reliant on shops or cars to save the day.

When I was learning to program, to make software and games, I had previously mostly trained experience as a film photographer, and studying and making socially-engaged public art projects, and lots of practice making zines. I had a vision in my head for some ways I could fuse this work with programming, ways that differed from software art and digital art I saw in the world. But I had to build up the skills and experience to be able to make it. And it took me years, even into grad school. And Iā€™m still working to improve and refine my work. Thatā€™s a lifetime pursuit. Yet, Iā€™m proud of my work. I enjoy the process, and the refinement of my practice.

If youā€™re just starting out, this is even more reason to try your own hand and not the machine-robot-language modelā€™s hand. If you donā€™t care about the outcome too much, throwing some ā€œpromptsā€ into a black box and taking what pops out is what youā€™ll present. But if you want to build as a creative person, and learn and get better and better at your work, your writing, your art-making, and enjoy the process, and the challenge, then try doing it yourself. Youā€™ll be happier in the longterm.

I responded to the game designer who emailed the listserve about suggested they try using an original music tracker from that era. Itā€™s in keeping with programming using the old tools and techniques theyā€™re already using to make the rest of the game. It might take some time to make something theyā€™re happy with, but it will be a good experience, maybe even a good little struggle at first, and with practice theyā€™ll get better as they go, and maybe discover something new.

I truly donā€™t know if they followed that advice. But I hope they found their joy in the practice, and that theyā€™ll make work theyā€™re proud of.


Update: I wanted to add another thought as well: I donā€™t think all AI is the same, or that it canā€™t be part of a creative practice. Aside from environmental concerns of burning fuel to power GPUs, my argument above is aimed at a response to relying on generative Large Language Models to write wholecloth, generate new artwork, or write code, as just a few examples, particularly at the point when one is trying to learn or cultivate oneā€™s own body of creative work.