I kid you not, earlier today I was talking to a coworker and later to my wife (despite neither of them asking me anything) about AI and DSLs. And now you've worded it so much better than I did!
My angle was that programming is about making 1000s of decisions: from irrelevant to super important. Ideally we'd let AI make the irrelevant ones so we can focus on the ones that are super important for the problem domain. And how do you abstract away the irrelevant decisions to leave just the pure problem domain ones? DSLs. And hey, look at that, Ruby is exceptionally good at spinning up DSLs. Now, if only we can figure out how to teach LLMs to make consistently good DSLs with Ruby.
I’ve delightfully enjoyed this. Especially the otters. 🙏🏽
was wildly entertained by this
these are too good
These are also just good problem solving strategies in general. Reminds me of Kahneman and Tversky's "System 1" vs "System 2"-thinking model.
this is so great
so cool !!! THANK YOU FOR WRITING THIS !!
Anything that mentions abductive thinking is going to be interesting, but this was especially well-written and useful.
Spot on, especially Manuel. :)
I kid you not, earlier today I was talking to a coworker and later to my wife (despite neither of them asking me anything) about AI and DSLs. And now you've worded it so much better than I did!
My angle was that programming is about making 1000s of decisions: from irrelevant to super important. Ideally we'd let AI make the irrelevant ones so we can focus on the ones that are super important for the problem domain. And how do you abstract away the irrelevant decisions to leave just the pure problem domain ones? DSLs. And hey, look at that, Ruby is exceptionally good at spinning up DSLs. Now, if only we can figure out how to teach LLMs to make consistently good DSLs with Ruby.
lovely
Well said, it made me chuckle and snort a few times as well... marmalade windows! 🤣
> But knowing and doing kept arriving at different addresses.
🔥⚡