What LLMs do when they code is actually two things:
1. Translate from natural language to executable code
2. Expand a very high-level overview into a lot of detail
Each of these is great. Having them only available inextricably combined is a problem.
Thoughts?
HTMXNative is taking shape: my little ToDOMVC demonstrator is running identically on macOS, iOS, Android and the web.
No Electron.
Binary size 4MB.
The only JS is a modified htmx.
And the application logic is the same single Objective-Smalltalk Software-IC (bundle).
Streams can use blocks and objects interchangeably.
] stdline ! { :s | s ! 'Hello World!' }
Hello World!
] stdline ! 'Hello World!'
Hello World!
What if the reason dynamic languages are so suitable (even necessary) for component composition / scripting is not because we actually need dynamism for that.
But because dynamism lets us work around the fact that our model(s) for composition are wrong.
On my blog: “And it seems retro in the worst way that we’re still using anything other than a scripting language for most of our code. We should be using something simple and light that can configure toolbars, handle networking callbacks, query databases, manage views, and so on.”
I am starting to see the need for a stdout that always appends a newline at the end, so println: but on the object side.
What should this be called?
Jordan Maris 🇪🇺 🇺🇦 #NAFO
Brent Simmons
Patricia Aas