I have this thought about AI usage (and new technologies in general), which is that when they're brand-new, there's this "shiny stage" where people play with them, and it's PLAY, and it's great. But then over time, what was once novel becomes ordinary, and we fall into ruts where we're just kind of mindlessly using that technology (whether we need it or not) because we can, and we become its prisoner, and Using The Thing is grim drudgery.
On the other hand, I keep telling myself I'm going to use AI _MORE_, but I gave myself this one requirement... that I'm going to be "in a place" where I'm really giving thought to each ChatGPT dialogue I'm having and not just spewing random questions, but trying to somehow (?) be intentional about my inputs. It turns out that giving myself that requirement + mostly only using browsers on a computer + whatever other habits & tendencies are relevant results in... "not much AI usage."
(Technically, my firrrrst thought of something I might say in response to my post... was an idea i had when I mis-read "identifying objects with wheels" as "identifying which objects are wheels..." and I was like "hmmm, wheels or Biblical visions of angels? Sometimes those are surprisingly hard to distinguish!")
Have you tried the newer AIs? The instant it got actually useful -o3 - I had already decided to stop using it. The newer ones are frighteningly intelligent.
For me, it was too easy to slip into random inputs... unfortunately I like to generate speech myself, so I have a lot in common with LLMs.
Also, now if I'm less than perfect, I can just blame it on lack of AI!
I used an AI to write the prototype for the "hide DSL threads" plug-in. (Then I never finished it!!)
And some other code-based thing I was playing around with.
And once the Zvi post about [4o being an] "absurd sycophant" hit, I sat down and had ChatGPT (possibly even 4o, ironically!) generate some satire, which thread is labelled "Flatteropocalypse Satire."
Oh, also, this was really interesting to me!! 'Traditionally we repeat the second to last verse so as not to end the reading with the word “bad”.'
I've turned that (and its implications) over in my head lightly (i.e. not thinking too deeply or strenuously) a few times.
ironically, the thing that idea made me think of was just... incorporating such a move in something i was writing! (just ending one paragraph with a somewhat-more-positive thing than what i would default end with.)
I have this thought about AI usage (and new technologies in general), which is that when they're brand-new, there's this "shiny stage" where people play with them, and it's PLAY, and it's great. But then over time, what was once novel becomes ordinary, and we fall into ruts where we're just kind of mindlessly using that technology (whether we need it or not) because we can, and we become its prisoner, and Using The Thing is grim drudgery.
On the other hand, I keep telling myself I'm going to use AI _MORE_, but I gave myself this one requirement... that I'm going to be "in a place" where I'm really giving thought to each ChatGPT dialogue I'm having and not just spewing random questions, but trying to somehow (?) be intentional about my inputs. It turns out that giving myself that requirement + mostly only using browsers on a computer + whatever other habits & tendencies are relevant results in... "not much AI usage."
(Technically, my firrrrst thought of something I might say in response to my post... was an idea i had when I mis-read "identifying objects with wheels" as "identifying which objects are wheels..." and I was like "hmmm, wheels or Biblical visions of angels? Sometimes those are surprisingly hard to distinguish!")
Have you tried the newer AIs? The instant it got actually useful -o3 - I had already decided to stop using it. The newer ones are frighteningly intelligent.
For me, it was too easy to slip into random inputs... unfortunately I like to generate speech myself, so I have a lot in common with LLMs.
Also, now if I'm less than perfect, I can just blame it on lack of AI!
I used an AI to write the prototype for the "hide DSL threads" plug-in. (Then I never finished it!!)
And some other code-based thing I was playing around with.
And once the Zvi post about [4o being an] "absurd sycophant" hit, I sat down and had ChatGPT (possibly even 4o, ironically!) generate some satire, which thread is labelled "Flatteropocalypse Satire."
And for... almost nothing else besides that.
I'd like to see that satire!
Oh, also, this was really interesting to me!! 'Traditionally we repeat the second to last verse so as not to end the reading with the word “bad”.'
I've turned that (and its implications) over in my head lightly (i.e. not thinking too deeply or strenuously) a few times.
ironically, the thing that idea made me think of was just... incorporating such a move in something i was writing! (just ending one paragraph with a somewhat-more-positive thing than what i would default end with.)
Yes, actually the full tradition is that you should end any endeavor with a hope for redemption for the world, if you really want to do it right!
I'm glad you enjoyed that!