Good article. I recently wrote a piece that echoes some of themes.
AI is most useful when it's taking away boring drudge work, freeing up people's energy and creativity for the good stuff.
I don't use it much directly for creative output, but I do find it helpful as a thought partner or editor. Sometimes I will ask the AI to steelman an argument, critique my position and help me evaluate my it versus contrary ones. AI can also be a helpful copy editor, suggest ways to tighten up my language or give me alternative ways of phrasing something. I prefer to have it make suggestions that I implement though, rather than asking it to do the work. That way what I end up producing reflects my voice vs that of the AI.
Filing and categorizing. Although while doing those repetitive tasks also allows for day dreaming, often the fertile soil of creativity. Perhaps too, the many childhood hours spent practising scales birthed much of my future.
Yup. If I like doing it, I’m not going to outsource it to an intern… Not even a magical wish granting intern.
The current crop of solution-in-search-of-a-problem genies aren’t solving any problems yet that I don’t want to solve for myself. Some people, perhaps most, will always prefer to rub a magic lamp or a lottery scratch off than to do the hard work of making something. They can’t afford a ghost writer, but maybe a free chatbot will make them a sudden billionaire?
The silly thing is: If the corporations trying to figure out how to monetize the intern genies really had a wish granting factory, why would they be trying to sell it instead of granting themselves infinite wishes and taking over the world?
AI is very helpful for my data analysis confirming hunches using large data sets that would take forever running formulas, scrubbing, and sorting. I click a few buttons and its done.
I am also a big fan of bulletpoints and summarizing large blocks of texts, which AI is fantastic for.
Yes AI for data work makes sense, totally diff use case than art (although to be fair, sometimes that can be creative work too, asking the right questions of your data etc)
Good hot take per usual. I think AI eventually becomes the “paintbrush” meaning it will come down to the human who wields it and not the technology itself. The visuals I use on my newsletter are all technically AI generated though I start with a sketch and I sweat the details. AI rarely nails it on the first try. It takes a good deal of going back and forth.
Interesting you and another commenter talked about requiring multiple prompts to get what you want. I get that, and feel similarly, but also at some point I would just rather open up creative software to be precise. For imprecise work, sure, let Jesus-bot take the wheel.
It’s a strange phenomenon. Reminds me of being a creative director and working with designers. Sure I could sit down and design it myself, but something I discovered early in my career is that I’m a better director than designer. So AI is my personal design team when it comes to my visuals. I get it where I want it and it’s not unlike creative directing a designer. Actually pretty similar. Will only get better with time
To be clear I think for day job work AI might be great. I don't think corporations really want you to do the same type of raw creative work you might do on your own. It's mostly the opposite because you are adhering to a brand that isn't you - there's totally different stylistic guidelines etc. Which is why you should really only ever work at a company you like the style of if you are a creative, otherwise you'll be miserable.
I've used AI a lot - mainly writing and to produce film. The main difference for me is you need to be able to deal with probability. For example I made a film using AI, one of the shots called for a satellite in orbit passing across the shot and blocking the sun on the way through. It took dozens of runs to get it right, and occasionally it threw in something unexpected - like a rocket launching from earth far below. But it was (much) quicker, (much) cheaper and just a different way of tackling a problem.
Good article. I recently wrote a piece that echoes some of themes.
AI is most useful when it's taking away boring drudge work, freeing up people's energy and creativity for the good stuff.
I don't use it much directly for creative output, but I do find it helpful as a thought partner or editor. Sometimes I will ask the AI to steelman an argument, critique my position and help me evaluate my it versus contrary ones. AI can also be a helpful copy editor, suggest ways to tighten up my language or give me alternative ways of phrasing something. I prefer to have it make suggestions that I implement though, rather than asking it to do the work. That way what I end up producing reflects my voice vs that of the AI.
Filing and categorizing. Although while doing those repetitive tasks also allows for day dreaming, often the fertile soil of creativity. Perhaps too, the many childhood hours spent practising scales birthed much of my future.
It's definitely a good argument to allow some boredom - fertile area for creativity to be nourished
Yup. If I like doing it, I’m not going to outsource it to an intern… Not even a magical wish granting intern.
The current crop of solution-in-search-of-a-problem genies aren’t solving any problems yet that I don’t want to solve for myself. Some people, perhaps most, will always prefer to rub a magic lamp or a lottery scratch off than to do the hard work of making something. They can’t afford a ghost writer, but maybe a free chatbot will make them a sudden billionaire?
The silly thing is: If the corporations trying to figure out how to monetize the intern genies really had a wish granting factory, why would they be trying to sell it instead of granting themselves infinite wishes and taking over the world?
AI is very helpful for my data analysis confirming hunches using large data sets that would take forever running formulas, scrubbing, and sorting. I click a few buttons and its done.
I am also a big fan of bulletpoints and summarizing large blocks of texts, which AI is fantastic for.
Yes AI for data work makes sense, totally diff use case than art (although to be fair, sometimes that can be creative work too, asking the right questions of your data etc)
We are seeing the backlash of AI on the creative side. Substack booming from people fleeing trad media pushing out AI Gen nonsense.
Imax selling out and films only being done in 70mm. Even saw Kodak is making a camera film comeback.
It turns out the real world is awesome and most of the AI slop will get rejected.
Good hot take per usual. I think AI eventually becomes the “paintbrush” meaning it will come down to the human who wields it and not the technology itself. The visuals I use on my newsletter are all technically AI generated though I start with a sketch and I sweat the details. AI rarely nails it on the first try. It takes a good deal of going back and forth.
Good thoughts. Keep em coming
Interesting you and another commenter talked about requiring multiple prompts to get what you want. I get that, and feel similarly, but also at some point I would just rather open up creative software to be precise. For imprecise work, sure, let Jesus-bot take the wheel.
It’s a strange phenomenon. Reminds me of being a creative director and working with designers. Sure I could sit down and design it myself, but something I discovered early in my career is that I’m a better director than designer. So AI is my personal design team when it comes to my visuals. I get it where I want it and it’s not unlike creative directing a designer. Actually pretty similar. Will only get better with time
To be clear I think for day job work AI might be great. I don't think corporations really want you to do the same type of raw creative work you might do on your own. It's mostly the opposite because you are adhering to a brand that isn't you - there's totally different stylistic guidelines etc. Which is why you should really only ever work at a company you like the style of if you are a creative, otherwise you'll be miserable.
I've used AI a lot - mainly writing and to produce film. The main difference for me is you need to be able to deal with probability. For example I made a film using AI, one of the shots called for a satellite in orbit passing across the shot and blocking the sun on the way through. It took dozens of runs to get it right, and occasionally it threw in something unexpected - like a rocket launching from earth far below. But it was (much) quicker, (much) cheaper and just a different way of tackling a problem.