I’m not sure where you got the idea that we are modeling our AI integration after Apple when we have talked very little publicly about it.
My vision for integration of AI into Xojo is that of an optional Assistant that you as the user actively chooses to engage. If you don’t, it would be just a button in the tool bar you never use.
I have no interest in adding AI for the sake of adding AI. In fact until recently I would have said it’s simply not ready but I have now seen with my own eyes that it can be quite useful. But again, I imagine it as an assistant that you choose to work. The last thing I would want is for it to be a nuisance you can’t avoid.
If you’re still unsure how AI can be useful in software development without taking over, I highly recommend watching this short video about GitHub Copilot:
It clearly demonstrates how Copilot works with you, not instead of you. It doesn’t build your app for you — it watches what you’re doing, understands your codebase, and offers intelligent, context-aware suggestions as you write. For example, just by typing a well-named function like calculateInvoiceTotal, Copilot already knows what you’re probably trying to implement and offers the code — which you can accept, reject, or modify.
This is only possible because Copilot sees your entire codebase and understands the structure and intention behind your code. That’s why embedding AI into the IDE is so important. Outside tools can’t match this level of assistance.
It’s not magic. It’s just autocomplete — but with context, intelligence, and an actual understanding of your intent. That’s what makes it so powerful, and that’s exactly the kind of assistant I hope to see in Xojo someday — optional, but truly helpful when used.
Thanks for the clarification, Geoff. It’s good to hear that the AI integration will be entirely optional and something the user can choose to engage with. That’s definitely the right direction — AI should assist, not interfere.
What I was reacting to was your earlier post, where you mentioned that you think Apple is taking the right approach with “Apple Intelligence”, emphasizing the assistant-style branding over technical terms like “AI”. That sounded like Xojo might follow that path — which, in the context of actual AI capabilities, felt surprising.
Because frankly, of all the major companies, Apple has — so far — shown the weakest real-world AI integration. They’ve wrapped the tech in polished messaging, yes, but the depth of functionality and integration lags far behind what Microsoft, Google, and even GitHub (with Copilot) have delivered.
For developers, AI becomes really useful when it’s integrated tightly into the IDE — like Copilot in VS Code. It’s not about “build me an app” prompts. It’s about understanding my current code, offering smart completions, and helping me stay in flow. That only works if the AI has access to my source code and can actively assist me as I write — just like an advanced autocomplete system that understands context.
And this is not some niche use case. GitHub Copilot is now used millions of times every day in real-world production environments, especially inside VS Code — not just by beginners, but by highly experienced, professional developers. They use it to get rid of the repetitive, boring, boilerplate parts of development and to move faster without sacrificing code quality.
So I really hope that the Xojo assistant — when it comes — takes this path. Optional, yes. Respectful of user control, absolutely. But also deep, smart, and embedded enough to really help developers rather than being just a floating window that talks in generalities.
Since AI realistically won’t run locally, using it always means your code leaves your machine and gets processed externally – under the terms of the respective provider, such as GitHub Copilot. I hope and expect that Xojo will rely on established solutions here and allow full opt-out per project.
I think people are taking that satirical statement too literally. The point was LLMs work just fine without direct integration into desktop applications.
I don’t think Xojo are going to be expecting others to buy into the AI from a third party. They seemed to be looking to do their own, trained on the documentation. At least that was the jist at the German meeting, unless I misunderstood.
apple is not the same with tim cook, and lots of people think tim cook should be replaced, he is just managing the company, steve jobs paved the road he is just following the way.
In fact in france we heard Jula who made Siri at Apple, was loud about saying apple didn’t give much effort at the time, years later, nobody uses siri and if steve jobs would have been here, he would have understand, but he is not.
now google/amazon got ahead of apple.
wwdc was a disapointment concerning AI, and now apple is trying to buy perplexcity, which could be a good thing.
and in France we talk a lot about IA, since many IA directors are french, Meta with Lecun, and i think netflix, amazon ? and we have french AI Mistral, still trying to figure out why is that, math schools ?
well AI is changing a lot of things and one could expect that at some point, some LLM can write good xojo code within the next years , it does for other languages
I continue to be astonished at the number of people who think the role of CEO should be treated as an elected official, and that they know best who should fill it. Your average person on the street has ZERO actual idea of what it takes to run a huge corporation such as Apple. They seem to think that Jobs was a sort of emperor-king, ordained by a deity to reign over a grateful consuming public.
Shareholders who are not on boards of directors do not typically oust CEOs - that is the job of the directors, not the shareholders. But since you think it happens frequently, do feel free to prove your case with examples.
No, he wasn’t. He was sidelined by Apple’s board of directors, leaving him with little to do in the company. After his plan to execute a coup to remove the CEO was leaked, he resigned from both the company and the board of directors.
