Squarespace is actually a good example; someone still has to write all the nifty widgets and figure out how they'll work together for a pleasing experience. If putting an attractive visual layer on top of code was all it took, then Smalltalk would have won decades ago and everyone would program.
The main barrier to "no code" and "everyone can program" is still the mode of thought, not the mode of representation.
This has parallels in artificial intelligence, where as soon as a problem is solvable by a computer, solving the problem no longer counts as "AI" (chess, natural language processing, even sentiment analysis!). Similarly, as soon as a problem is made easy and accessible to non-algorithmic thinks, programmer move on to the next thing that requires that mode of thought. Maybe we'll exhaust the problem space of potential programs someday, but I don't see any evidence for it.
> It's thinking algorithmically in enough detail, with edge cases, fallbacks, error handling, etc, that most people struggle with.
And this is why we need better languages, not better editors.Python was the last real language to bring prepackaged solutions that are easy to use and cover (some) edge cases to reduce development overhead. Adding a use-specific framework, such as Django, often takes care of the remaining 80%.
I'd love to see a new language or framework that makes adding new features - not features that the framework devs created but rather real new features - easy to code would be a boon to development.
In fact, by providing an entry point into an online presence, squarespace and other click-to-get-online products may actually increase the pool of businesses interested in migrating to a specialized, purpose-written website as they expand past the capabilities that squarespace offer them.
So yes, easy stuff will be automated, as in many other industries, but this will create new possibilities that only specialized developers will be able to pursue.