I remember the wonderful talk @Geoff_Perlman did in Germany on integrating AI into the Xojo IDE. Plus I understand this may take a long while to add into the IDE in an effective and useful manner.
But in the month that Apple is belatedly adding AI to their macOS, well after Google and Microsoft have thrust AI into every conversation, product and keyboard, this makes we wonder why AI IDE integration isn’t even an item on the Xojo Roadmap (Roadmap — Xojo documentation) — even if it were at the bottom of the list?!
For what? To make a regretful decision? When Microsoft integrates AI into their IDE such IDE will use Microsoft AI to spy the users contents and sell more contents and services to their users. When Google does it, they do the same, Google’s AI. Apple the same, Apple’s AI. $$$ into their own ecosystems, just it. There are lots, I mean LOTS of experiments going on right now. And companies, and choices, and integrations, and costs. What to integrate? How integrate? What’s the cost? What’s the benefits. What’s the penalty for a bad decision? What if the current choice you pick now is considered the worst choice in 6 months? You, by yourself, can pick your choice right now, play with it, and regret your choices yourself, and pick another one. Is right now the correct time to compromise the IDE with things in a so early stage?
A roadmap is just that: a roadmap. Not a promise, a general plan. Plans change when the context evolves. Roadmaps should be taken for what they are, not more. I, for one, would be disappointed that effort is put on integrating AI in the IDE instead of improving core features and renewing the Windows desktop controls, for example. Just saying.
Back in 1996, I was told to ‘add internet’ to a major application I was working on.
It didn’t need it, and literally had no use for it.
So after much WTF-ing, we compromised on adding a toolbar button marked ‘Internet’, which launched Internet Explorer open to the company website.
Next release had ‘Internet Enabled!!’ all over the box.
I felt dirty.
Personally, I don’t want AI anywhere at all, but definitely not in the Xojo IDE
There are better things to work on.
To err is human… to really screw things up, add AI.
OTOH… iPads have been using keyboards for over a decade, and Xojo Mobile still can’t tell if a modifier key on an external keyboard has been pressed, nor can you add keyboard shortcuts to aXojo iPad app, so adding AI is probably about a decade and half down the road.
We are still hard at work on it. I have spent a lot of time with LLMs as you can imagine. They are prone to hallucinate and that’s unacceptable in the context of an assistant. I also think Apple is taking the right approach. Apple is almost NEVER first to market. They are often second and better. They also don’t tend to use technical terms in their marketing. So it’s Apple Intelligence rather than AI. The feature we add will almost certainly be labeled an assistant rather than an AI assistant. Companies that use the term AI in the feature name itself I believe are making a mistake.
I can’t entirely be sure you feel about Apple Maps. For me, I find Apple Maps superior to Google Maps. I know others that feel the opposite. I find it depends on how you use them. I use them to get from on known destination to another. I find that Google Maps is not as smart about gauging traffic conditions, accidents, road work, etc. The overall user experience of Apple Maps is far better at this for me. I used Apple Maps extensively in my last car which had a CarPlay compatible radio. My current car is a Tesla Model Y which uses Google Maps. So I have experience with both under the same conditions.
I believe he’s referring to the disastrous initial rollout of Apple Maps, with its post-apocalyptic satellite imagery and inaccurate data. This resulted in a rare public apology from Apple and is thought to have cost Scott Forstall his job as Senior Vice President of iOS Software at Apple.
