Well, for Xojo that’s true. But for most other programming languages and IDEs it’s not. For those AI is already a fact and useful in 2023.
If you leave out the justified copyright and security discussions, the tools are useful, very useful - even today. Less in the sense that complete apps are written, but autocomplete gets better, error messages can be interpreted, repetitive tasks are simplified via suggestions, etc. And very often you even get useful suggestions just by writing in a comment what should happen when next. Of course, AI sometimes fantasizes, but so do users on stackoverflow and Reddit.
I’m not talking about using chatGPT as yet another search engine, but included AI helpers (Github CoPilot, Jetbrain’s integrated AI in the lastest IDEs etc.)
BUT, I haven’t the faintest idea how this is supposed to happen in Xojo:
- The IDE is already slow today
- The autocomplete is already faulty and lame today
- In addition, a solid, public code base to feed AI with Xojo Code is missing
- I’ve always criticized Xojo for not separating Web1 and Web2 threads. How is an AI supposed to be able to differentiate correctly here? Not at all! The same will apply to API1 and API2…
- Parts of the AI magic (training the systems) happens through feedback. Are the suggestions useful or not, will the code get changed, etc. I’m afraid that the Xojo user base is by far too small to become competitive with any modern language in the AI universe.