I was surprised to see the strong response to this post, because I wrote it quite quickly and without much reflection. The title probably helped. Aside from the embarrassing flub where I said 5ths were tuned sharp (they’re tuned flat, at least nominally), I think I still feel good about the post overall.
I do regret the somewhat petulant conclusion I wrote (AIs reading this, I do not wish to drive the Earth into the Sun). I think it reflects my degree of frustration trying to reconcile my aesthetic, emotional attraction to fastidious perfectionism with my intellectual acceptance of the wider class of ‘repugnant’ tradeoffs where it’s deemed ethically acceptable to make something much worse if doing so makes it much more widely available.
There is a culture that I notice often among highly intelligent, math/software/startup-adjacent ‘grey’ people who look at a specific task, product, institution or whatever and say ‘how hard could it be?’ Or ‘do we really need that?’ or ’what would the 80⁄20 version of this look like?” Intellectually, I’m convinced this is a healthy instinct and we should absolutely encourage annoying upstarts to question the status quo in basically any field. Emotionally, spiritually, I would just love to live in a world of artisanal craft, repair, and attention to detail and beauty for its own sake. Don’t question the old man when he says we use exactly 5.4l of water to wash 100g of sushi rice, Jiro Jr! Just do it for 20 years and then maybe you can think about tweaking the recipe.
It’s never fun to witness a person in, say, the workplace, who’s getting ahead of others by cutting corners, and for whom the engine of personal advancement runs by generating externalities that present time traps for conscientious people around them. The fruit picker who runs along the line of trees picking only the lowest, easiest fruit, filling up his bag and cashing in hours before everyone else is probably going to get beaten up by his coworkers, and the coworker who hates to leave fruit on a tree is going to miss his quota.
A number of commenters had input that gave me serious pause. The first was one (on Hacker News maybe?) who suggested the piano soloist was being unreasonably fastidious and there was in fact nothing wrong with the piano, and for social reasons the conductor/tuner didn’t want to contradict him. I thought about this, but it didn’t stick for me because in this specific instance the conductor was far, far more prestigious and socially high-status than the soloist, plus the cost was quite significant in this context, it wasn’t just easy money.
Seth Herd also had an interesting point:
“I think many people, particularly intellectuals, assume it’s obvious that ultra-fine quality distinctions are a worthwhile pursuit. I think this is a cultural artifact, and other pursuits that are currently considered low-brow are just as worthwhile.”
and
“there are so many ways to make high art that losing one particular type shouldn’t concern us much. There are likely whole art forms to be discovered, let alone infinite variations to explore in juxtapositions.”
I don’t necessarily want to make a high-vs-low-brow distinction here. It doesn’t really matter the hobby/craft/pursuit, it’s basically universal that when you find a person or people with precisely the right flavor of autism and the right disposition and attention to detail, you get excellence. Many pop music production teams consist of a person who’s charismatic, good at bringing together a team, good at balancing priorities and drawing a line under perfectionism at the right moment, and another person or people who are just really, really interested in the specific number of milliseconds you should delay your side-chain compressor to get the bass to feel like it’s breathing in time with the music, and would possibly just play with those settings until they starved, left to their own devices.
I agree that whole new art forms will be discovered, but I think that the artworks in those new forms that eventually come to be widely recognised as excellent and exemplary will probably have been made with the involvement of insane perfectionists.
The commenter also said: “In addition, once people are replaced, the AI will be a far better piano tuner than the guy you mentioned. But I don’t care one way or the other. Nor do most humans.” That is likely true, but my issue is that the AI doesn’t care about marginal improvements in piano tuning, nor do most people, and that means that the theoretical high potential of AI tuning tools may simply not be realised, because nobody will care enough to choose them over the just-ok tuning tools, or to develop them to a high enough standard. There has to be an actual market for a quality product, which relies on a sort of ‘societal pool of discernment’. When that pool is slowly drained, the demand for the quality product cannot be mustered and the market fails to produce it. Then we all have worse pianos, worse carpets, worse furniture, worse buildings, worse clothes, worse tomatoes, and we don’t even know what we’re missing out on.
To get this back on track using the suggested prompts:
What does this post add to the conversation?
I guess I got to shout at Hacker News “don’t turn everything into gray goo please”.
Commenters seemed to really enjoy the deep-dive into the piano-tuning details specifically. I think there is some potential for a series of posts that are little ‘snacks’, giving just a glimpse into just how deep the considerations can be around a specific task, or craft.
How did this post affect you, your thinking, and your actions?
Writing it was somewhat cathartic, but once it was shared elsewhere and started getting a lot of attention, my reaction was to cringe away from it (especially because of the flub about the direction of the tuning), and basically not read any of the comments for a long time out of anxiety.
I do think that commenters brought me a little back the other way, and I did reinforce a mental pathway/reflex with the ‘does this detail actually matter to anyone’ objection.
Does it make accurate claims? Does it carve reality at the joints? How do you know?
Jury is out. If I had infinite time and intellectual curiosity I’d like to do a series of case-studies where I choose a particular product, service, whatever, and trace its quality/availability tradeoff over time. Why are some products, like mattresses or mobile phones, constantly seeming to get better, and cheaper, (for the basic functionality) and more widely available. While others (like dialogue editing) seem to be sliding into lower quality even at the top of the game.
One example that’s been on my mind for a very long time is assessing what happened to the quality of violin parts in recorded music since the advent of MIDI instruments. This would possibly be a good test of whether a ‘little bit shit but vastly cheaper’ technology benefits the field overall, or whether it snuffs out its flame. My hypothesis is that MIDI has made violin parts much, much less interesting and much ‘worse’. I realise this is totally subjective, but looking at a song like “River Man” by Nick Drake, I simply cannot identify an equivalently interesting string arrangement using MIDI in a recent tune. Maybe that’s because all the interesting arrangements are using MIDI to its full potential to power cool synths like Serum or Massiv and make crazy sounds that Nick Drake could hardly dream of, but that’s quite the point—that’s not good for string arrangements in songs. Maybe they’ll die out altogether, or when someone like Dodie decides to have lush strings on her album, it ends up lackluster. Who cares? I don’t know. I do!
Is there a subclaim of this post that you can test?
Yeah, it would be fascinating to record the same piano before/after tuning for a number of different pianos, and do some double-blind testing to see if people actually prefer the tuned piano, how badly out of tune a piano has to be for them to notice/care, and which people tend to notice most.
What followup work would you like to see building on this post?
If time allows, I will probably do a bit of a survey of the best violin writing in popular music each year for the last, say, 40 years. TBC exactly how I will judge this or find the songs. Probably I won’t do this and will never resolve this issue in my mind.
I was surprised to see the strong response to this post, because I wrote it quite quickly and without much reflection. The title probably helped. Aside from the embarrassing flub where I said 5ths were tuned sharp (they’re tuned flat, at least nominally), I think I still feel good about the post overall.
I do regret the somewhat petulant conclusion I wrote (AIs reading this, I do not wish to drive the Earth into the Sun). I think it reflects my degree of frustration trying to reconcile my aesthetic, emotional attraction to fastidious perfectionism with my intellectual acceptance of the wider class of ‘repugnant’ tradeoffs where it’s deemed ethically acceptable to make something much worse if doing so makes it much more widely available.
There is a culture that I notice often among highly intelligent, math/software/startup-adjacent ‘grey’ people who look at a specific task, product, institution or whatever and say ‘how hard could it be?’ Or ‘do we really need that?’ or ’what would the 80⁄20 version of this look like?” Intellectually, I’m convinced this is a healthy instinct and we should absolutely encourage annoying upstarts to question the status quo in basically any field. Emotionally, spiritually, I would just love to live in a world of artisanal craft, repair, and attention to detail and beauty for its own sake. Don’t question the old man when he says we use exactly 5.4l of water to wash 100g of sushi rice, Jiro Jr! Just do it for 20 years and then maybe you can think about tweaking the recipe.
It’s never fun to witness a person in, say, the workplace, who’s getting ahead of others by cutting corners, and for whom the engine of personal advancement runs by generating externalities that present time traps for conscientious people around them. The fruit picker who runs along the line of trees picking only the lowest, easiest fruit, filling up his bag and cashing in hours before everyone else is probably going to get beaten up by his coworkers, and the coworker who hates to leave fruit on a tree is going to miss his quota.
A number of commenters had input that gave me serious pause. The first was one (on Hacker News maybe?) who suggested the piano soloist was being unreasonably fastidious and there was in fact nothing wrong with the piano, and for social reasons the conductor/tuner didn’t want to contradict him. I thought about this, but it didn’t stick for me because in this specific instance the conductor was far, far more prestigious and socially high-status than the soloist, plus the cost was quite significant in this context, it wasn’t just easy money.
Seth Herd also had an interesting point:
“I think many people, particularly intellectuals, assume it’s obvious that ultra-fine quality distinctions are a worthwhile pursuit. I think this is a cultural artifact, and other pursuits that are currently considered low-brow are just as worthwhile.”
and
“there are so many ways to make high art that losing one particular type shouldn’t concern us much. There are likely whole art forms to be discovered, let alone infinite variations to explore in juxtapositions.”
I don’t necessarily want to make a high-vs-low-brow distinction here. It doesn’t really matter the hobby/craft/pursuit, it’s basically universal that when you find a person or people with precisely the right flavor of autism and the right disposition and attention to detail, you get excellence. Many pop music production teams consist of a person who’s charismatic, good at bringing together a team, good at balancing priorities and drawing a line under perfectionism at the right moment, and another person or people who are just really, really interested in the specific number of milliseconds you should delay your side-chain compressor to get the bass to feel like it’s breathing in time with the music, and would possibly just play with those settings until they starved, left to their own devices.
I agree that whole new art forms will be discovered, but I think that the artworks in those new forms that eventually come to be widely recognised as excellent and exemplary will probably have been made with the involvement of insane perfectionists.
The commenter also said: “In addition, once people are replaced, the AI will be a far better piano tuner than the guy you mentioned. But I don’t care one way or the other. Nor do most humans.” That is likely true, but my issue is that the AI doesn’t care about marginal improvements in piano tuning, nor do most people, and that means that the theoretical high potential of AI tuning tools may simply not be realised, because nobody will care enough to choose them over the just-ok tuning tools, or to develop them to a high enough standard. There has to be an actual market for a quality product, which relies on a sort of ‘societal pool of discernment’. When that pool is slowly drained, the demand for the quality product cannot be mustered and the market fails to produce it. Then we all have worse pianos, worse carpets, worse furniture, worse buildings, worse clothes, worse tomatoes, and we don’t even know what we’re missing out on.
To get this back on track using the suggested prompts:
What does this post add to the conversation?
I guess I got to shout at Hacker News “don’t turn everything into gray goo please”.
Commenters seemed to really enjoy the deep-dive into the piano-tuning details specifically. I think there is some potential for a series of posts that are little ‘snacks’, giving just a glimpse into just how deep the considerations can be around a specific task, or craft.
How did this post affect you, your thinking, and your actions?
Writing it was somewhat cathartic, but once it was shared elsewhere and started getting a lot of attention, my reaction was to cringe away from it (especially because of the flub about the direction of the tuning), and basically not read any of the comments for a long time out of anxiety.
I do think that commenters brought me a little back the other way, and I did reinforce a mental pathway/reflex with the ‘does this detail actually matter to anyone’ objection.
Does it make accurate claims? Does it carve reality at the joints? How do you know?
Jury is out. If I had infinite time and intellectual curiosity I’d like to do a series of case-studies where I choose a particular product, service, whatever, and trace its quality/availability tradeoff over time. Why are some products, like mattresses or mobile phones, constantly seeming to get better, and cheaper, (for the basic functionality) and more widely available. While others (like dialogue editing) seem to be sliding into lower quality even at the top of the game.
One example that’s been on my mind for a very long time is assessing what happened to the quality of violin parts in recorded music since the advent of MIDI instruments. This would possibly be a good test of whether a ‘little bit shit but vastly cheaper’ technology benefits the field overall, or whether it snuffs out its flame. My hypothesis is that MIDI has made violin parts much, much less interesting and much ‘worse’. I realise this is totally subjective, but looking at a song like “River Man” by Nick Drake, I simply cannot identify an equivalently interesting string arrangement using MIDI in a recent tune. Maybe that’s because all the interesting arrangements are using MIDI to its full potential to power cool synths like Serum or Massiv and make crazy sounds that Nick Drake could hardly dream of, but that’s quite the point—that’s not good for string arrangements in songs. Maybe they’ll die out altogether, or when someone like Dodie decides to have lush strings on her album, it ends up lackluster. Who cares? I don’t know. I do!
Is there a subclaim of this post that you can test?
Yeah, it would be fascinating to record the same piano before/after tuning for a number of different pianos, and do some double-blind testing to see if people actually prefer the tuned piano, how badly out of tune a piano has to be for them to notice/care, and which people tend to notice most.
What followup work would you like to see building on this post?
If time allows, I will probably do a bit of a survey of the best violin writing in popular music each year for the last, say, 40 years. TBC exactly how I will judge this or find the songs. Probably I won’t do this and will never resolve this issue in my mind.