Yesterday, Apple delivered its long-awaited WWDC 2024 keynote and I’ll admit, it was pretty awesome! Craig is the man, and I love how much screen time Apple gave him. His notable power stance was elevated this year with a rock performance and some serious Jason Bourne move swhen introducing the new goods! Well done Apple, well done. The keynote was worth watching just for him!
On a more serious note the first hour of the event showcased the new OS-specific features coming out in the next major software updates across all of their platforms. We’ll get to those cool features towards the end of the article. What I was most interested in was Apple’s take on AI (that’s “Apple Intelligence” by the way, thanks Apple for re-branding yet another universal concept). Going into the keynote, I wanted to keep a level head, being speculative but also still letting myself be blown away by how they would integrate AI into their product line. I was not disappointed, both from a security perspective and from a tech-nut perspective. So let’s dive in.
“Apple Intelligence”
It’s hard to have good conversations with people about AI these days because either they believe it’s a crazy concept and we’re all going to get taken over by robots or turn into cyborgs, or they fall into the unrestrained techie category. I on the other hand believe that what players in the AI space need is vision. It’s not about what the AI can do for them, its about how AI can truly enrich their lives. And I don’t mean save time on writing school papers or doing menial tasks on their behalf, no. I mean truly enrich the interactions they have with other people.
While yes, there are some AI features announced this year that definitely fall into the lazy-student category, there are also a lot of amazing features that I personally can’t wait to try out with my family and friends. But as much as I look forward to using them, I want to know that my privacy and data are not going to be mismanaged. I don’t want my devices to take screenshots and upload them to their AI model (I’m looking at you Microsoft Recall), and I don’t want my data being stored in the cloud or being used to train a model. For those reasons, I was so happy to hear that Apple valued those things too.
Architecture
There are three major considerations Apple made to the construction of their AI architecture:
- Data is never stored.
- The AI is used only for your requests.
- Verifiable privacy promise by third party professionals.
Starting from the lowest level, Apple is committed to maintain privacy on the local device by isolating the personal context from the AI model itself. They do this by essentially using a feeder function to feed the generative model with just the barebones data provided by the user. The generative model then takes that “dumbed-down” data, processes it to perform the task desired by the user, and then feeds it back to the user without ever collecting or storing the raw data itself. The way Apple describes it, the AI is only “data-aware”, it doesn’t store it for processing.
However, while most of the AI capabilities on Apple devices will be managed locally, Apple designed Private Cloud Compute which upscales the performance capabilities of local AI in a cloud-like environment. When this announcement was initially made, my hope in Apple was briefly lost. But after they explained it more, and after I did some of my own research, I have to say that I think I’m more okay with it than I was during the keynote. Private Cloud Compute sits on top of the same Apple Silicon, with the same built-in privacy and security functions in individual user devices. Additionally, Apple is committed to the same “feeder function” model thats built locally to maintain user privacy. Not only that, Apple is committed to preserving privacy even in the administration of these Siliocon-based servers themselves as outlined in this blog post regarding secure Private Cloud Compute design.
Across Apple’s AI research papers and announcements, there is a sentiment of aggressive protection of privacy; this is good. Apple emphasizes the fact that data used for AI purposes will never be stored in a server or anywhere else off of your device, not even for logging or R&D purposes. This is also good. Many other AI-providers host the bulk of their AI capabilities on massively powerful servers that they control, meaning that security researchers are left with no un-biased way to verify protection of privacy on those devices. Apple on the other hand wants to make that a priority. We’ll see how that shakes out in practice, but at least the commitment to transparency is there.
Overall, I must say I’m well pleased with this architecture. Their commitment to safer AI seems very clear, and as both a daily Apple user and IT professional, I look forward to taking advantage of this new technology in my personal and business life. But enough of the serious talk, I think there are some very cool features coming out in the upcoming OS updates that I want to share my opinions on, however valuable they may be.
“Finally!” Features
There are several features that I’m so happy are finally reaching the Apple ecosystem. I’m certainly upset that some of this functionality is late in hitting the modern market, but I’m mostly just excited that Apple users will now have some of the features they’ve been asking for for so long.
iOS: Home screen icon and app customization are finally here! That’s all I have to say about that.
macOS: There are a lot of cool things coming out in macOS Sequoia too, but the one with my loudest “finally” remark was its windows organization feature. It really had me wishing I didn’t shell out $10 for Magnet earlier this year. I should’ve known things were bound to change.
Across the Apple ecosystem, you will be able to place passwords over any application that you want, meaning your sensitive business information, or the notes app where you store your kids’ Christmas gift ideas, will no longer be subject to prying eyes! This was a finally feature for sure, but Apple added to the functionality of securing privacy by removing that information’s system indexing, meaning that data from password-protected applications won’t be available in other applications such as Spotlight Search. I like this, and it is a big win for privacy.
Cool New Features
iPhone Mirroring is a feature that enables Mac users to remotely access their iPhone with an on-screen emulation. It is a fully functional feature allowing audio passthrough and a complete set of navigation gestures. It is really cool in concept, and I completely understand the accessibility advantage with not having to get up and walk 15 feet to get to your iPhone, but I’m also left wishing that there was a grander contribution to this major update. Especially considering how disastrously expensive the latest MacBook models are. Also, the security professional in me has me wondering how catastrophically this could break. Don’t get me wrong, I have faith that the wizard developers at Apple have put the effort into making this a secure, open-air remote connection. But I’d hate for airport travelers to get locked out of their phones as a degenerate hacker hijacks their iPhone, leaving it in a locked state as everything in their private lives gets exposed. I have to look into it more, but I’m hoping to see some kind of Apple ID sign-in or passcode requirement to prevent free access to this when it’s rolled out.
There are some stellar releases hitting iPadOS in this next update. I just wish I was still in college to take advantage of them! I can’t tell you how many multivariable calculus problems I could’ve used Apple’s new Math Notes feature on. The ability to comprehend and dynamically process math problems in the way it was demonstrated blew me away. This begs the question though, how complex can my math problems get? I mean, they demonstrated it’s capabilities to perform rudimentary algebra and graphing, but will it solve differential equations? If not, will it someday?
Genmojis are about to change my whole world! I can’t tell you how blown away I was when I saw this announced. It’s such a simple concept now in the age of AI, but I can’t wait for it to be a built-in component of my Apple experience. My excitement about this doesn’t really come from a personal use case though, but rather from a web development case. I enjoy using emojis from time to time in my web dev work, but I can’t tell you how many times I’ve wanted there to be an emoji that would compliment the topic or the page well and I’ve turned up with nothing. But that’s no more! …hopefully? I have to wonder how compatible these custom icons will be across devices. There is already a noticeable difference between emojis on Windows and Mac, so will we see this become a universal compatibility issue? I certainly hope not. It won’t stop me from using them though!
Many of the same emotions were felt for Image Playgrounds. I just think this is such a cool concept and I can’t wait to try this out. There is too much functionality in this feature alone to discuss in this post, but if you haven’t watched the keynote or seen shorts on social media about it, do yourself a favor and take a look. I think that if Apple Intelligence is truly as private as they say it will be, there is no stopping how much I’m using this. I apologize in advance to everyone I’m closely connected with, especially my wife!
Thank you for taking the time out of your undoubtably busy day to read this recap of WWDC 2024’s keynote! I was super excited to watch it and if you haven’t seen it yet, you really should. That techie side of you will be really happy you did!