1. My blood pressure is a lot lower after having listened to Blossom (twice in a row). Egads, is there anything you can't do? Stop raising the bar like this, for the sake of the rest of us.
2. SO MUCH FUN. Thank you for the opportunity, and hope I can repay it soon.
3. The only actual Substack rule I know of is: there are no rules, so feel free to be weird and break stuff wildly in all directions. But since there are no rules, that's not a rule, so at this point I would like to vanish in a puff of self-thwarted logic. Thank you.
*****
Re. Cognitive Offloading and the role of AI:
Thank you for raising this one in a sketch. Oh boy. I have so many thoughts swirling around this right now, and yet for me the whole thing feels like trying to nail jelly to the ceiling, because there are so many smart people writing about this and I need to put time aside to understand them...
My biases at work: my creative process is messy and semi-analogue (I have to write notes on a topic on my Remarkable 2 tablet, physically writing them with a pen, in order to really feel I'm getting a grip on them - but I also use my phone and laptop and so on) and while I've tinkered with various AI services, I currently use none of them to do my work because I enjoy the process of stumbling over things myself - and I'm wary of automating and farming out a process that is teaching me to be better at my job *and* is a lot of fun. (Plus, I have enough ego that I want the sense of satisfaction from having done the job myself in a way that gives me enough self-confidence to do my marketing around it.)
Beyond the ethical arguments around AI tools right now - which I think usually boil down to ethical question-marks hanging over the humans behind them and the ways they use them, eg. AI replacing experienced employees, the copyright battle between the UK Commons & Lords right now, etc. - well, I can maybe see myself using AI at some point. But one analogy sticks in my mind...
I used to play a lot of squash. Great fun, great for my fitness levels, and the closest I've ever got to becoming obsessed with a sport. (I'd love to get my fitness back to somewhere near that level, and that's a goal for myself in coming years, but - later.)
The very first time I played squash, I was 100% knackered at the end. And I remember thinking, "Oooh, I look forward to getting better at squash, so I won't be anywhere near as exhausted at the end of a match."
But this is not what happens! And it's probably not what *should* happen, as you get better at a skill. What really happens is you push yourself to your limit every time, and therefore spend all of your energy, and it's just that as you get better at doing it, your capacity increases and you achieve more. But it's no easier at the end. Or maybe only the tiniest bit. You're pretty much still as exhausted as you were as a noob - it's just that you played 5x or 10x better.
With AI, I see way too many folk saying "AI will make creativity easy". After which they'll post something that's technically impressive but creatively impoverished & narratively incoherent. It could be just the usual bout of cosplaying creative work that comes with every technical leap forward. But - if it's easy, and you're not creatively knackered at the end, have you really done anything *better*? Was your cognitive load as heavy as it could be, for your own good, in the same way heavier weights will make you stronger, but where "heavy" translates to "deeper thinking"?
I love that squash analogy. It reminds me of the experience of someone working in customer support at a tech company (as for my last two experiences). Customer support is used to dealing with the problems, so you always get to hear about the things that don't work. As the product slowly gets better and fixes those problems it serves more people automatically. So a lot more people are getting their problems solved by the product, but for someone in customer support you're still faced with endless issues, just from a smaller percentage of the many more people being served by the product everyday. It's easy to feel like nothing's getting better (like, say, still being exhausted at the end of your squash match), but actually, the product and service have improved a lot.
I agree, that there's potential to allow us to do more, and better, and different things we couldn't do before. But rarely does important work get easier =)
I always took notes in talks and have the same experience you mention in that article. I like writing because it engages all my senses (ok, not taste unless you chew your pencil) and making myself think of what to write helps me process the information. Compared to, say, taking a photo of a slide that someone shows, which perhaps captures everything so much better, but doesn't actually help you learn it.
The main risk I see is that the automated thinking partner aborts the thinking early on and reduces the range of imaginable solutions.
If the thinking partner provides a list of ideas, the urge to think of the whole problem space -- to build a more compresensive onthology -- can easily go missing.
Great point. In my experience, if you think to ask for ideas from an AI thinking partner, you will get a lot of ideas. And you can always just ask for more. But it's easy to forget to ask, and it's easy to consider just the ideas you were given and not consider further as you say. It reminds me of this visual about the importance of getting past the same ideas. So easy to stop:
I've added a custom instruction to ChatGPT so it asks me thought-provoking follow-up questions, to a pretty good result. I'd say you can definitely customize it in a way that it generates more ideas without necessarily having to ask it yourself.
Wow, now that I've read this delusional wank about how chatbots somehow help you get ideas, I can safely say I won't ever follow your work again even if you do manage to leave the newsletter nazi bar. Best of luck with all that. Sheesh.
1. My blood pressure is a lot lower after having listened to Blossom (twice in a row). Egads, is there anything you can't do? Stop raising the bar like this, for the sake of the rest of us.
2. SO MUCH FUN. Thank you for the opportunity, and hope I can repay it soon.
3. The only actual Substack rule I know of is: there are no rules, so feel free to be weird and break stuff wildly in all directions. But since there are no rules, that's not a rule, so at this point I would like to vanish in a puff of self-thwarted logic. Thank you.
*****
Re. Cognitive Offloading and the role of AI:
Thank you for raising this one in a sketch. Oh boy. I have so many thoughts swirling around this right now, and yet for me the whole thing feels like trying to nail jelly to the ceiling, because there are so many smart people writing about this and I need to put time aside to understand them...
My biases at work: my creative process is messy and semi-analogue (I have to write notes on a topic on my Remarkable 2 tablet, physically writing them with a pen, in order to really feel I'm getting a grip on them - but I also use my phone and laptop and so on) and while I've tinkered with various AI services, I currently use none of them to do my work because I enjoy the process of stumbling over things myself - and I'm wary of automating and farming out a process that is teaching me to be better at my job *and* is a lot of fun. (Plus, I have enough ego that I want the sense of satisfaction from having done the job myself in a way that gives me enough self-confidence to do my marketing around it.)
Beyond the ethical arguments around AI tools right now - which I think usually boil down to ethical question-marks hanging over the humans behind them and the ways they use them, eg. AI replacing experienced employees, the copyright battle between the UK Commons & Lords right now, etc. - well, I can maybe see myself using AI at some point. But one analogy sticks in my mind...
I used to play a lot of squash. Great fun, great for my fitness levels, and the closest I've ever got to becoming obsessed with a sport. (I'd love to get my fitness back to somewhere near that level, and that's a goal for myself in coming years, but - later.)
The very first time I played squash, I was 100% knackered at the end. And I remember thinking, "Oooh, I look forward to getting better at squash, so I won't be anywhere near as exhausted at the end of a match."
But this is not what happens! And it's probably not what *should* happen, as you get better at a skill. What really happens is you push yourself to your limit every time, and therefore spend all of your energy, and it's just that as you get better at doing it, your capacity increases and you achieve more. But it's no easier at the end. Or maybe only the tiniest bit. You're pretty much still as exhausted as you were as a noob - it's just that you played 5x or 10x better.
With AI, I see way too many folk saying "AI will make creativity easy". After which they'll post something that's technically impressive but creatively impoverished & narratively incoherent. It could be just the usual bout of cosplaying creative work that comes with every technical leap forward. But - if it's easy, and you're not creatively knackered at the end, have you really done anything *better*? Was your cognitive load as heavy as it could be, for your own good, in the same way heavier weights will make you stronger, but where "heavy" translates to "deeper thinking"?
I want to believe these tools can be used in truly great ways. But I think about how using a paper map can be such a different and enjoyable experience to using a digital map, and how modern generations increasingly don't know how to use them - https://www.independent.co.uk/news/uk/home-news/map-reading-millennials-smartphones-ordnance-survey-a8932976.html - and I wonder what's being invisibly lost, the same way that the thousands of years we've spent learning to handwrite have also taught us how to think - https://everythingisamazing.substack.com/p/why-it-might-not-sink-in-until-youve - and is the speed we're moving towards keyboards getting in the way of our thinking in that regard?
Such a huge topic.
(And of course I had to go leave you the longest comment in Substack's history - which I wrote with a keyboard, so it may be total gibberish. Hooray!)
Wow. Thanks, Mike!
I love that squash analogy. It reminds me of the experience of someone working in customer support at a tech company (as for my last two experiences). Customer support is used to dealing with the problems, so you always get to hear about the things that don't work. As the product slowly gets better and fixes those problems it serves more people automatically. So a lot more people are getting their problems solved by the product, but for someone in customer support you're still faced with endless issues, just from a smaller percentage of the many more people being served by the product everyday. It's easy to feel like nothing's getting better (like, say, still being exhausted at the end of your squash match), but actually, the product and service have improved a lot.
I agree, that there's potential to allow us to do more, and better, and different things we couldn't do before. But rarely does important work get easier =)
I always took notes in talks and have the same experience you mention in that article. I like writing because it engages all my senses (ok, not taste unless you chew your pencil) and making myself think of what to write helps me process the information. Compared to, say, taking a photo of a slide that someone shows, which perhaps captures everything so much better, but doesn't actually help you learn it.
The main risk I see is that the automated thinking partner aborts the thinking early on and reduces the range of imaginable solutions.
If the thinking partner provides a list of ideas, the urge to think of the whole problem space -- to build a more compresensive onthology -- can easily go missing.
Great point. In my experience, if you think to ask for ideas from an AI thinking partner, you will get a lot of ideas. And you can always just ask for more. But it's easy to forget to ask, and it's easy to consider just the ideas you were given and not consider further as you say. It reminds me of this visual about the importance of getting past the same ideas. So easy to stop:
https://sketchplanations.com/get-good-ideas
I've added a custom instruction to ChatGPT so it asks me thought-provoking follow-up questions, to a pretty good result. I'd say you can definitely customize it in a way that it generates more ideas without necessarily having to ask it yourself.
Smart!
Not very happy to have a newsletter I'm subscribed to abruptly start sending me content from Substack. I'll be unsubscribing. You should seriously rethink this change and look into Ghost or Beehiiv, and not support a platform that has explicitly supported Nazis: https://www.techpolicy.press/substack-founder-defends-commercial-relationships-with-nazis/
Wow, now that I've read this delusional wank about how chatbots somehow help you get ideas, I can safely say I won't ever follow your work again even if you do manage to leave the newsletter nazi bar. Best of luck with all that. Sheesh.
Sorry about that. There were a lot of factors that led to this change. Appreciate you following this far.