I quoted Rich Hickey (the author of Clojure) in my research paper. 1 Perhaps his most quoted quote:
Programmers know the benefits of everything and the tradeoffs of nothing.
Actually it was from his 2012 keynote for the Rails Conference. 2 And it was in turn a modification of Alan Perlis’s Epigram #55: “A LISP programmer knows the value of everything and the cost of nothing.” 3 The quote itself is not that important. But the 13 years old keynote deserves to be revisited now, late 2025. Of course, the context is the topic I really don’t want to talk about it again, but I have to, “AI”.
Back in the non-AI age of 2012, Hickey said in the keynote that we optimized for the ease of development, but we don’t look at what we are getting along with that. We focus too much on ourselves, not the product. He called this “metaculture.” He used the example of Foo Fighters. If the Foo Figthers were a band that only think about how hard it is to themselves, they would think that learning to play drum and guitar is too hard. They would all of them playing Kazoo 4 and become Kazoo Fighters. However, you probably would not want to listen to Kazoo Fighters; also you also don’t want to be join a band like Kazoo Fighters that only plays Kazoo.
Hickey did not explain why we do not want to listen to Kazoo Fighters; or why we do not want to join a band that only plays Kazoo. It is now up to us to interpret what he meant. From my understanding, it is because the band does not care about their product and only cares about their comfort; and joining such a band is not inspiring because working with these ease-seeking people does not need to use any effort. One does not improve. In fact, one does not need to improve. In this “anyone knows how to make noise with their mouth can play Kazoo” and “there is no need to improve” environment, members of the Kazoo Fighters are replaceable. Alice can play a Kazoo, so she can join Kazoo Fighters. If Alice asks for too high a pay, the band manager of Kazoo Fighters can fire Alice without any worry. He could just hire Bob, who also can play Kazoo. No one cares about the product, therefore no one would notice the differences between what were produced by Alice and what will be produced by Bob.
I don’t want to do it like the previous post to say those who use AI are Eichmanns. 5 But I think the core argument of using AI is for seeking ease and comfort, not for producing better products. You can learn the techniques and then draw professionally for 20+ years, so that you might draw like an animator at Ghibli. The same can be said about a lot of things, from playing guitar, programming, speaking a foreign language, to simply having 60 minutes of attention span to attend a Zoom meeting. Of course I know these things are hard.
With all the hard things on this planet, here is where value starts to diverge: I am not asking for accomplishing those hard things. But did you even try to do these things on your own first?
With the new so-called AI, no, we don’t need to try. The so-called AI can do the thing for us. The same way we don’t need to try to learn how to play the guitar, because our mouth can actually make sound similar to a guitar with a Kazoo. But does a Kazoo sound really the same as a guitar? If we don’t care about the product, yeah, we accept the Kazoo sound as guitar. We did something and save ourselves from trouble. We are so comfortable because our brain and our body do not need to engage with any hard problem. It is so easy that any replaceable human being with the AI can do the same thing. We saved time too. We can use the saved time to keep doom scrolling our phone.
However, there are some people who try: They learn how to play the guitar but not the Kazoo. And in fact, people try, people also develop new skills, and create something new and (potentially) better. In the past, we encourage people to try. Do we still? If you don’t, that’s okay. It’s your choice. But please don’t laugh at people who try. Also, I don’t want play in your band.
Chan, C. H., Schatto-Eckrodt, T., & Gruber, J. (2024). What makes computational communication science (ir) reproducible?. Computational Communication Research, 6(1), 1. https://doi.org/10.5117/CCR2024.1.5.CHAN ↩
I do not particularly hate Kazoo. But the point here is that kazoo is easier to learn than guitar or drum, but not all people would consider kazoo the same thing as guitar or drum. ↩