I've spent some time with Apple Intelligence, using version 18.1 for several weeks and now 18.2 for most of yesterday.

I really appreciate the notification summaries. A few times, just by glancing at my phone, I could instantly see and understand what was waiting in my inbox. It’s not flawless—occasionally, the order of information is off, or it says the opposite of what the message actually contains. However, these issues are rare, and overall, I find it to be a genuinely useful feature.

The image search via the camera button is particularly impressive. I snapped a picture of my dog, a Corgi-Shepherd mix, and when I asked it to identify the breed, it nailed it. It worked for everything I tried, even pulling up information about the game I was playing based on a picture of my TV. This felt like the first time AI truly seemed magical to me. I have no idea how it works in real-life scenarios, but I’m looking forward to trying it more during a work trip I have in early November.

As for the playground features, they’re about as gimmicky as I (and many others) expected. I doubt I’ll use them beyond some initial testing, and fortunately, they can mostly be removed. While there are still image generation features scattered throughout the system, they’re easy enough to ignore.

Overall, as I said during my initial thoughts at WWDC, I’m impressed with Apple’s slower approach compared to the more aggressive, desperate, borderline dangerous paths taken by Google and Microsoft. Here’s hoping Apple stays the course.

I found WWDC24 to be a mixed bag.

I’m excited about the quality-of-life updates across Mac and iPhone, including the new Passwords app, updates to Mail, and continuity from the phone to the computer. I’m also interested but cautious about some of the intelligence Apple showed off.

I’m disorganized when it comes to the things on my phone and computers (ex, I have documents spread across 3-4 different apps), so having a more personalized (and private) “assistant” on my phone could be beneficial. I like the ability to teach Siri how to use apps to automate repetitive tasks, like what Rabbit was promising for the R1. Lastly, our household uses Siri a lot via the HomePod in our living area, so the improvements there will be nice.

On the flip side, the image generation stuff looked awful, and the whole idea felt very un-Apple to me. I honestly believe the presentation would have been better received without any of that included. It’s also a shame to see these sorts of features announced without any mention of how this impacts Apple’s environmental efforts.

Above all, it’s nice that Apple will be rolling this out in stages as a beta rather than forcing all sorts of half-baked AI “updates,” as we’ve seen from Microsoft and Google.