A week after WWDC

The announcements at WWDC last week are still kind of blowing my mind. I wrote about Apple a few months back and how there was such a visible AI moat between their release processes and presence amidst the AI explosion, it was interesting to see how it appeared they might have leapfrogged ahead, but after some reflection not so much.

In a way only Apple can, they released an AI strategy that;

1. Positions them with the market leader without having to be the market leader (we have ChatGPT!)

2. Safeguards them against LLM shortcoming (it’s OpenAI’s fault!)

3. Starts to build out their AI tech stack in a very risk adverse way (Apple Intelligence, App Intents)

Such impressive moves! WWDC was filled with all the usual fanfare, I was aglow the first day’s keynotes.

But it faded; why? As someone who identifies as in the Apple developer ecosystem, WWDC used to bring meaningful changes you could begin to experiment with immediately. The beta iOS meant something. You would download the new Xcode and iOS and start tinkering with the new APIs.

I did the downloads this year, and there were literally no noticable changes. The sessions being all so damn polished about “to come later this year” advances are so crisp and I appreciate the time savings of sessions not all being 50+ minutes anymore, it does have a little “am I a mindless sheep” sort of feeling to it.

The biggest signal, the thing I have been complaining / excited about for months coming to WWDC? Copilot-esque code completions in Xcode. They announcenced them of course. But they’re nowhere in sight; likely to require a full MacOS release, and not available until later this year.

So ya, I’m disillusioned. WWDC seemed like Apple had gotten all caught up, but in retrospect it was performative, which now feels like the furthest thing from actually delivering on value to developers.

I am however bullish on a few things.

Their ability to deliver AI in the way they describe. LLMs are inherently hard to bring to production for strict must-be-correct use cases, and Apple really strives for must-be-correct experiences. Biting off bits with App Intents powered by AI that just do a specific thing is absolutely a great approach for them to inch them into AI. They ain’t ever gonna do AGI at this pace, but obviously thats not their goal.

I’m also bullish on Swift on the server. I really like Swift as a language, they’ve put a ton of thought into, and it’s lovely to build in. If their private cloud stuff does pan out to process personal data, if people actually jump on to that, using Swift for backend stuff will be pretty cool. None of this really has a business justification at this point, but its an interesting space to watch.

And Siri? Ok, not that bullish. But glad the announcements are out, and eager for later this year.

At least that’s how I feel this morning, maybe that’ll change when I no longer need to keep a VScode side by side my Xcode when doing Swift development.