Specifically related to this aspect, currently the closest thing I'm aware of that comes to mind is the Godot "game" engine: https://godotengine.org
While it's not an exact match for Hypercard in many aspects, three of the related areas where it is a particularly strong contender are:
* A small initial download size (<100MB) and short load/launch load time.
* Supports an "iterative"/hacky development process well--GDScript is very well suited to the initial "exploratory development stage" of "making something that works". (As GDScript itself evolves it's also getting better at being robust over time via recent additions such as gradual typing.)
* Godot's distribution/installation story is very strong--I would posit that it is light years ahead of any other ecosystem[0] in this area: Download the export templates (which are generic platform-specific binary executables) for the platforms you want to support and then export your project (which packages your project-specific code/resources for use with the executable)[1].
Currently, I would say the place where Godot is currently most lacking (somewhat surprisingly) in comparison to something like Hypercard or ("classic") Visual Basic is in relation to a lack of "drag and drop UI" creation. Godot has a quite powerful UI system (albeit at times... opaque) but the current UX for building the UI forces one to be very conscious of the scene's node tree rather than being able to just drag and drop widgets from a toolbar, or similar, onto a window.
(This UX is also an aspect that seems could be improved "relatively easily" via an Editor Plugin and maybe one day I'll get around to prototyping something for that particular rabbit hole. :) )
Obviously, it's not a perfect solution for numerous reasons I won't go into now but it's certainly my default choice for any GUI-based utilities these days--some of which I actually finish enough to "release". :D
[0] This aspect is where Python in particular falls down catastrophically.
[1] This also extends to being able to produce WASM exports for the web which opens a two-pronged distribution approach where people can initially use a web-based version of your utility and, if it's useful enough to them, they can also download an executable for local use.