regardless of how taste is designed or created or … is that any formula requires practice in some mechanical way. i can’t historically identify any one who eventually was told they had “unique” taste without the repetitions, iterations, and experience that building anything great requires.
so much of taste developes in the dark, way before any public applause can be heard.
Taste is depth outside the comfort of the herd, so that we can show them "no, *this* is better". Taste is someone's experience & opinion, along with a number of people willing to follow them (otherwise, it would be called *bad* taste). What everyone is trying to achieve is critical mass, so that the new herd can sustain itself instead of dying off.
But how many tokens do we want to burn to send opinionated LLMs on that journey? Will we even like them when they come back? Who will be the early adopters of what they return with? Who will lead the charge, given that no one really invested into it nor is particularly attached to it? No pain was felt, and no tears were shed.
And then we're right back to square one with the tyranny of choice: an endless number of agents with an endless number of marginally "better" results to choose from, without any clear direction (e.g. "they moved the button -- is that good?"). How big do you have to be to stand out in the ocean?
People follow people more than taste, because there are countless example of great taste that went undiscovered. It's all about culture, tribes, and vibes. We are social creatures, not art critics. The logo and the status matter more than the design.
So it seems like taste has more to do with attracting *people* than identifying good design ideas. Nikola Tesla was one of the greatest geniuses the world has ever seen, but he had terrible taste/leadership, so he constantly failed to show us "no, *this* is better".
“Taste” is just being good at predicting what other people will like. It’s more pattern-matching than art-making, which is why LLMs will probably eventually be great at it.
But of course even here I’m not talking about *my* taste. I’m talking about other people’s taste. My taste is of course singular and unassailable.
It's also possible to lead people though. They may not yet know that they will like something. Or we can *convince* them to like something, even something that's not great at all (e.g. Hitler). And charisma, culture, tribes, and vibes have a lot more to do with it than we think. Moving the herd is always the hardest part (especially with so many people trying to do it / pulling in their own direction), and it can hardly be done with good design/taste alone.
Unless what we mean by "taste" is charisma+culture+tribes+vibes all rolled into one? In that case I guess it goes back to your pattern matching idea, and who has the best antenna to pick up the working patterns/vibes. But you still can't dress up a mouse as a lion or a king. Some people know how to read the crowd, but some people are visionaries. But then again, are the top products really on top because of their immaculate design? Did the Nazis get into power because of their nice boots or their cool armband? It's like we're selling dreams (and vibes), and people get excited about those potential futures (generally, whatever seems to work in their favor).
If we took all of the theatre away, people would like very different things. But on the stage, everything changes.
An admirer recognized the artist at a café and asked for a quick sketch on a napkin. Picasso quickly obliged. But when he asked for a large sum of money for the drawing, the woman was shocked, pointing out how fast he had drawn it. Picasso famously replied: "No, it took me a lifetime."
Looks like things that can be done quickly are perceived as worthless. Makes me wonder about the value of vibe-coded apps, or anything AI produces. Are we racing to the bottom?
I love the way you use the word "Tyranny" here. It captures the very nature of the work we do in tech/data/AI.
The more powerful tech it is, the more empowered the "center" is. No matter what this "center" is (corporation, governments, individuals, etc.). Right now, it is still humans sitting in the "center" wielding that power. And I don't really believe AI should be given that much of power, as I ain't sure if we as humanity could bear the worst possibility.
Centralizing power is generally a bad idea. That is exactly what evil people want, and the very first thing they do as soon as they step foot in office. So why should we do it for them, or make it any easier? Are we trying to roll out the red carpet for them? But we're going the wrong way because we're lured in by the shiny idea/dream of an AI utopia (which will fail to materialize, because the [mostly social] foundation is rotten -- and the solutions to that problem in the hands of centralized power are terrifying).
The problem is a very simple one, because guess what: when people have attained power (and/or wealth), *they want to hold on to it* (which is the central root of evil, and leads to much conflict & war: because people who do not have power do not want war either).
Fortunately, we live in a society that has embraced the peaceful transfer of power (although some people have been in lesser positions of power for ages, and just won't let go). Now we just need to do the same thing with wealth also (in order to alleviate inequality, gentrification, and inflation). A prosperous economy is built on the peaceful transfer/circulation of wealth -- so when the rich start to hoard away from others, they are actually harming their own nation.
Generating 80 concepts was never the hard part though. Picking the one that's worth machining glass for is, and that judgment only shows up after you've shipped a few wrong ones and felt it. Cheap generation doesn't teach you which of the 80 is the iPhone.
I would worry with a company like Taste Labs that this is more displacement of the commercial artists and industrial designers who are paid to have aesthetic sense that ordinary people who might try to design a product might not
"A place where we know we are loved" is so specific to what Don is going through at that moment that it is not transferable to many other products even as a joke
regardless of how taste is designed or created or … is that any formula requires practice in some mechanical way. i can’t historically identify any one who eventually was told they had “unique” taste without the repetitions, iterations, and experience that building anything great requires.
so much of taste developes in the dark, way before any public applause can be heard.
This.
Taste is depth outside the comfort of the herd, so that we can show them "no, *this* is better". Taste is someone's experience & opinion, along with a number of people willing to follow them (otherwise, it would be called *bad* taste). What everyone is trying to achieve is critical mass, so that the new herd can sustain itself instead of dying off.
But how many tokens do we want to burn to send opinionated LLMs on that journey? Will we even like them when they come back? Who will be the early adopters of what they return with? Who will lead the charge, given that no one really invested into it nor is particularly attached to it? No pain was felt, and no tears were shed.
And then we're right back to square one with the tyranny of choice: an endless number of agents with an endless number of marginally "better" results to choose from, without any clear direction (e.g. "they moved the button -- is that good?"). How big do you have to be to stand out in the ocean?
People follow people more than taste, because there are countless example of great taste that went undiscovered. It's all about culture, tribes, and vibes. We are social creatures, not art critics. The logo and the status matter more than the design.
So it seems like taste has more to do with attracting *people* than identifying good design ideas. Nikola Tesla was one of the greatest geniuses the world has ever seen, but he had terrible taste/leadership, so he constantly failed to show us "no, *this* is better".
“Taste” is just being good at predicting what other people will like. It’s more pattern-matching than art-making, which is why LLMs will probably eventually be great at it.
But of course even here I’m not talking about *my* taste. I’m talking about other people’s taste. My taste is of course singular and unassailable.
😆
It's also possible to lead people though. They may not yet know that they will like something. Or we can *convince* them to like something, even something that's not great at all (e.g. Hitler). And charisma, culture, tribes, and vibes have a lot more to do with it than we think. Moving the herd is always the hardest part (especially with so many people trying to do it / pulling in their own direction), and it can hardly be done with good design/taste alone.
Unless what we mean by "taste" is charisma+culture+tribes+vibes all rolled into one? In that case I guess it goes back to your pattern matching idea, and who has the best antenna to pick up the working patterns/vibes. But you still can't dress up a mouse as a lion or a king. Some people know how to read the crowd, but some people are visionaries. But then again, are the top products really on top because of their immaculate design? Did the Nazis get into power because of their nice boots or their cool armband? It's like we're selling dreams (and vibes), and people get excited about those potential futures (generally, whatever seems to work in their favor).
If we took all of the theatre away, people would like very different things. But on the stage, everything changes.
This also reminds me of the Picasso parable:
An admirer recognized the artist at a café and asked for a quick sketch on a napkin. Picasso quickly obliged. But when he asked for a large sum of money for the drawing, the woman was shocked, pointing out how fast he had drawn it. Picasso famously replied: "No, it took me a lifetime."
Looks like things that can be done quickly are perceived as worthless. Makes me wonder about the value of vibe-coded apps, or anything AI produces. Are we racing to the bottom?
I love the way you use the word "Tyranny" here. It captures the very nature of the work we do in tech/data/AI.
The more powerful tech it is, the more empowered the "center" is. No matter what this "center" is (corporation, governments, individuals, etc.). Right now, it is still humans sitting in the "center" wielding that power. And I don't really believe AI should be given that much of power, as I ain't sure if we as humanity could bear the worst possibility.
Centralizing power is generally a bad idea. That is exactly what evil people want, and the very first thing they do as soon as they step foot in office. So why should we do it for them, or make it any easier? Are we trying to roll out the red carpet for them? But we're going the wrong way because we're lured in by the shiny idea/dream of an AI utopia (which will fail to materialize, because the [mostly social] foundation is rotten -- and the solutions to that problem in the hands of centralized power are terrifying).
The problem is a very simple one, because guess what: when people have attained power (and/or wealth), *they want to hold on to it* (which is the central root of evil, and leads to much conflict & war: because people who do not have power do not want war either).
Fortunately, we live in a society that has embraced the peaceful transfer of power (although some people have been in lesser positions of power for ages, and just won't let go). Now we just need to do the same thing with wealth also (in order to alleviate inequality, gentrification, and inflation). A prosperous economy is built on the peaceful transfer/circulation of wealth -- so when the rich start to hoard away from others, they are actually harming their own nation.
Generating 80 concepts was never the hard part though. Picking the one that's worth machining glass for is, and that judgment only shows up after you've shipped a few wrong ones and felt it. Cheap generation doesn't teach you which of the 80 is the iPhone.
Maybe someone should tell Taste Labs about Jakob's Law.
tl;dr: people expect your design to work just like the ones they already know.
Says the man with an Android!
I would worry with a company like Taste Labs that this is more displacement of the commercial artists and industrial designers who are paid to have aesthetic sense that ordinary people who might try to design a product might not
What is "ordinary people"?
People whose full-time job is not designing products but might have something like a side hustle making things to sell on eBay
This isn't a phone, it's a time machine! (Generally booed reference)
"A place where we know we are loved" is so specific to what Don is going through at that moment that it is not transferable to many other products even as a joke