I’d not heard of websim, but it’s really cool. Just spent some time playing with it.
What this most reminds me of is what it was like to build things using HyperCard, but 100x easier because there’s no need to do any scripting or custom work to make your idea real, you just type text and iterate. Same sort of idea, but a lot faster and more accessible to people with less computer experience.
It’s also got some pretty clear limitations right now, but I expect that to improve with time and effort.
I spent about 45 minutes with it and got it to create an app I’ve been thinking about for the last couple years, but never find the time to spend the several hours to get it going. Really excited about the possibilities of this!
Yeah, it’s funny, it feels like the realization of the promise of computers. I guess we all told ourselves this narrative, “one day we’ll settle on a really good programming language and set of APIs we’re ever going to need, and we’ll get it memorized, and we’ll be able to express programs as fast as we can speak english, and control our computers, and create all that we need to create” and of course that never happened for most of us. The APIs are disgusting and complex and we end up straddling different programming languages and occasionally having to abandon one for another.
And the meaning-aware translation abilities of language models basically end up solving that.
I now feel emboldened to go ahead and create my weird tree maximalism post-textual programming language, because now, if people don’t want to learn another set of APIs in order to use it, they won’t have to learn it.
I’d not heard of websim, but it’s really cool. Just spent some time playing with it.
What this most reminds me of is what it was like to build things using HyperCard, but 100x easier because there’s no need to do any scripting or custom work to make your idea real, you just type text and iterate. Same sort of idea, but a lot faster and more accessible to people with less computer experience.
It’s also got some pretty clear limitations right now, but I expect that to improve with time and effort.
I spent about 45 minutes with it and got it to create an app I’ve been thinking about for the last couple years, but never find the time to spend the several hours to get it going. Really excited about the possibilities of this!
Yeah, it’s funny, it feels like the realization of the promise of computers. I guess we all told ourselves this narrative, “one day we’ll settle on a really good programming language and set of APIs we’re ever going to need, and we’ll get it memorized, and we’ll be able to express programs as fast as we can speak english, and control our computers, and create all that we need to create” and of course that never happened for most of us. The APIs are disgusting and complex and we end up straddling different programming languages and occasionally having to abandon one for another.
And the meaning-aware translation abilities of language models basically end up solving that.
I now feel emboldened to go ahead and create my weird tree maximalism post-textual programming language, because now, if people don’t want to learn another set of APIs in order to use it, they won’t have to learn it.