The AI Daily Brief: Artificial Intelligence News and Analysis - 7 Takeaways from Apple's AI Announcement at WWDC

Episode Date: June 11, 2024

Apple made its much-awaited announcement about AI today. NLW covers everything that was said, and how Apple is trying to distinguish itself from competitors. ** Join Superintelligent at https://besupe...r.ai/ -- Practical, useful, hands on AI education through tutorials and step-by-step how-tos. Use code podcast for 50% off your first month! ** ABOUT THE AI BREAKDOWN The AI Breakdown helps you understand the most important news and discussions in AI.  Subscribe to The AI Breakdown newsletter: https://aidailybrief.beehiiv.com/ Subscribe to The AI Breakdown on YouTube: https://www.youtube.com/@AIDailyBrief Join the community: bit.ly/aibreakdown

Transcript
Discussion (0)
Starting point is 00:00:00 Today on the AI Daily Brief, we finally got the long-weighted Apple AI announcement. Did it live up to the hype? Let's discuss. The AI Daily Brief is a daily podcast and video about the most important news and discussions in AI. To join the conversation, follow the Discord link in our show notes. Hello, friends, quick note, today, because everything is all Apple excitement, there is no headlines edition, just the main part of the episode. We will be back tomorrow with our normal approach to the episode. And I also didn't want to interrupt the flow of the observations and the rapid reactions with an ad.
Starting point is 00:00:40 So here's my one quick shout out for Super Intelligent, the sponsor of today's show. If you want to learn about AI to go use all these tools that Apple is about to bring to the masses, come check us out at B-Super.a.i. It's fun, fast tutorials paired with step-by-step how-toes, all in a really engaging community context. And if you use code Apple, just for the next couple days, you will get 75% off your first month. Again, that's B-Super.AI and Code Apple. And with that, let's get into the show. Welcome back to the AI Daily Brief.
Starting point is 00:01:10 Well, friends, today was the day after months and months of speculation. We finally had the Apple Worldwide Developer Conference 24. And the anticipation was that we were going to get a big announcement around Apple's much-questioned AI strategy. Certainly this was what all of Wall Street was waiting for. Analyst Ann Ives this morning tweeted today, Apple will unveil its highly anticipated AI strategy at WWDC and Cupertino, which we believe will kick off an AI-driven accelerated growth cycle on the iPhone and services that will be the narrative of the Apple story for the
Starting point is 00:01:42 coming years, a key moment for Apple. Also, pay attention to this first comment that was on Dan's post from Dissolve DC. Remember when Siri was AI and Apple had the lead and then did nothing with it and it still sucks all these years later? Just keep that in mind as we go through. So what we're going to do today is talk about seven observations from the event, which should give you a pretty good picture of what was going on. The first is that Apple is definitely testing how much they cannot play into the AI marketing hype. The most notable way they did that was making us wait forever for these announcements. Former founding team member at OpenAI Andre Carpathie wrote at 150 p.m. Eastern Time, if you tuned into WWDC to see what Apple is doing with AI, we're all probably
Starting point is 00:02:23 thinking the same thing around now 50 minutes into it. dot, dot, dot. He followed up with, I am also exhilarated to learn that you can now change the color of your icons and that you can choose any color you want right before we look at how we deploy state-of-the-art AGI to a few billion devices. The joke being, of course, that they went through literally everything else in their lineup, including things like changing the color of icons, before they got to this much-anticipated announcement. Indeed, it was 65 minutes when all was said and done before they actually said the word artificial intelligence. And even then, of course, this being Apple, you knew they had to rebrand it, and what they called it was Apple Intelligence. Now let's talk from here, though, about the second interesting observation,
Starting point is 00:03:02 which is around the next line here, AI for the rest of us. This phrase AI for the rest of us didn't make an appearance until almost the very end of this presentation. We're talking about an hour and a half, hour and 40 minutes in, but it was 100% the theme of this event. One of the things that makes AI so interesting right now is that on the one hand, you have constant mainstream media articles about how disruptive it is, how it's going to destroy entire categories of jobs, how it's going to rewire the economy, full stop. And I don't think that any of those things are wrong if you zoom out far enough. However, in the short term, they're extremely distracting from the way that AI is actually rolling out into the world right now, which is through a series of small but
Starting point is 00:03:44 significant and time-saving wins. When you use generative AI like chat chip ET or mid-jury to turn something that used to take you a half an hour into something that takes you 10 minutes, Well, if you get those 20 minutes saved every day over the course of a standard work year, you're talking about something like two and a half weeks of actual time savings. Apple is not only acknowledging that that is the reality of AI, but they are leaning all the way into it. I don't think a single example they gave was some frontier use case that you simply couldn't do in advance of AI,
Starting point is 00:04:12 with the possible exception of some of the image generation stuff, but even that I'll come to in a moment. Instead, it was all just faster, simpler ways to do the things that we've always done with iPhones. get the weather, figure out directions. The difference is that there was now an AI engine underlying all of that that was going to make it simpler and much, much faster. As you'll see, this will be the theme throughout this presentation,
Starting point is 00:04:34 but the first thing we have to talk about, the first manifestation we have to talk about with this, is, of course, Siri. When all was said and done, it really is still all about Siri. None of this was unexpected. Apple writes in their official announcement, with richer language understanding capabilities, Siri is more natural,
Starting point is 00:04:51 more contextually relevant and more personal, with the ability to simplify and accelerate everyday tasks. It can follow along if users stumble over words and maintain context from one request to the next. Additionally, users can type to Siri and switch between text and voice to communicate with Siri in whatever way feels right for the moment. Siri also has a brand new design with an elegant glowing light that wraps around the edge of the screen while Siri is active. Which, by the way, again, if you're wondering how significant Siri is to the overall announcement, it is that new Siri glowing light action that's happening right here on the cover page of the event on Apple. Now, what matters, though, about this Siri announcement is not just that Siri is getting better, although it is.
Starting point is 00:05:27 It's that Siri can take action everywhere. For example, they write, with on-screen awareness, Siri will be able to understand and take action with users' content in more apps over time. For example, if a friend texts a user their new address and messages, the receiver can say, add this address to his contact card. They continue with Apple intelligence, Siri will be able to take hundreds of new actions in and across Apple and third-party apps. For example, a user could say, bring up that article about cicadas from my reading list, or send the photos from the barbecue on Saturday to Malia. Another one that they highlighted in the presentation was asking when Mom's flight is landing,
Starting point is 00:05:59 which Siri goes into email to get the flight details and then references them with real-time flight tracking to give an arrival time. Halima Rasa, he said, did Apple just casually introduce a real large action model powered by OpenAI? We'll come to the chat GPT integration in a minute, but what they're referring to
Starting point is 00:06:15 is the idea of the rabbit device, which was premiered a couple months ago, which introduced this concept of a large action model instead of a large language model. The difference is that Siri is natively integrated into this operating system, so apps can all plug into this. Sully Omar tweets, you hear that? Siri can take actions on your phone.
Starting point is 00:06:33 This is the LAM you want. Now, a really interesting thing about this is that, yes, even though they've opened up the possibilities of using chat as an interface as well, really, Apple is making the same bet on a central app that is largely voice-controlled, orchestrating the entire suite of your digital experiences, in the same way that it seems that OpenAI and Google are making a similar bet.
Starting point is 00:06:55 This is a significant change that they are imagining, rather than having to navigate into whatever application you wanted to use, which has been the norm on mobile since the iPhone was launched, and which of course followed a human-computer interaction paradigm that was borrowed from desktops, now we're imagining a world where everything is done by a digital assistant. A lot of the destiny of all of these companies is going to rest on this bet. Fourth big note from this presentation, the OpenAI deal is real.
Starting point is 00:07:22 About an hour and 43 minutes into the presentation, Sam Altman tweeted, very happy to be partnering with Apple to integrate ChatGBTGBT into their devices later this year. Think you will really like it. One thing that wasn't exactly clear from the presentation
Starting point is 00:07:34 is exactly where ChatGBT BT ends and Apple Native AI begins, but one area that is explicit about where the difference lies is that Chat Chabit will be called upon by Siri when there's a query that goes outside of what Siri can do. They write Siri can tap into Chatubt's
Starting point is 00:07:49 expertise when helpful. Users are asked before any questions are sent to chatchapT along with any documents or photos, and Siri then presents the answer directly. This is obviously a huge deal for OpenAI. It puts them at the center of the iPhone operating system, and more specifically the center of the big bet of the new core interaction that Apple is betting on through Siri. But it also keeps Siri as the main brand that humans are interacting with. There will for sure be people who are interacting with chat GPUT via Siri who don't even really process that it's some separate application. Now, that's exactly what Apple wants from a user experience standpoint, but it does represent a bit of attention in a competition.
Starting point is 00:08:26 ChatGPT, I think, is also trying to be this default interface for your entire computing experience. Then again, OpenAI has now aligned themselves not with one, but two of the three biggest companies in the world to make sure that they are not left behind because they don't have access to native distribution. They are natively powered into the world's biggest enterprise suite via Microsoft, and now they are plugged in as well to the entire Apple mobile. suite. It will be interesting, of course, to see how the love triangle between OpenAI, Microsoft,
Starting point is 00:08:54 and Apple develops, but for now it's a pretty massive update for the company. Still, there was no Sam Altman on stage, no guest appearance. The whole thing was entirely produced. And indeed, I didn't include this as one of my seven observations, so I'll just make a note of it now. This could not have felt less like some of the other presentations that we've seen. Open AIs was the most intimate, a bunch of people doing some live demos, with a group of employees in the room looking on. Google was real humans in an auditorium, even if many of the demos or most of the demos were canned and filmed in advance. Apple was just a pre-produced video. This could just be personal preference, but it was sort of a turnoff to me, but it's also how
Starting point is 00:09:30 Apple has been doing things more and more recently. And so in some ways it might have been more of a break with what we would have expected had they gone back to the old actual presentation model. Now moving on to the fifth observation and staying on chat GPT for a moment, this shouldn't surprise us coming from Apple, but it is very clear that they are saying that product experience matters much more than state-of-the-art technology access. Or at least they're arguing that state-of-the-art technology is nothing without a good product experience. One example of this is, of course, what I was just mentioning about Chachapit being buried inside of Siri, but I think an even more profound example is Image Playground. Image Playground shows that Apple is willing to reduce people's
Starting point is 00:10:11 agency and decisions in the name of reducing complexity. This feature, I think, more than anything else than we saw, reinforced the idea that this is AI for Normies. So what do I mean by that? Well, a couple things. First of all, if you're used to using Mid Journey or Dali, where you have a prompt that gives you the full breadth of the English language to try to describe the style that you want your image to come back in, image playground will be jarringly on the opposite end of that spectrum. Users can create images in seconds just like they could with any other image generator, but they have their choice of just three styles, animation, illustration, or sketch. Again, they have gone from the full world of options that you get with any other image generator
Starting point is 00:10:51 to just three. This might prove to be really smart if it makes it a much more accessible experience to normal people if they actually start to use it inside apps like messages where it lives natively now. But it really is a remarkable statement on how Apple thinks about AI adoption. Relatedly, there is no prompting, at least it's not the main interface. Instead, what you do with Image Playground is you choose from a range of concepts and categories that include themes, costumes, accessories, and places.
Starting point is 00:11:20 So the example they gave is party, chef, and cat, and the combination of those preset options creates a party and cat chef with a few different example to choose from. Now, you can add a linguistic description to try to get it more specific, but that is an incredibly minimized feature. Instead, what they focused on was the integration with your contacts and with your photos, so that if you're talking to someone, you can use a picture of them as the basis for a cartoon image. Apple is really betting on this idea of personal context. It was the theme that they said probably most in the entire AI presentation. And part of that personal context is the people you know. Sixth observation is that Apple is threading some really difficult needles. As I just mentioned,
Starting point is 00:12:02 personal context was basically the most discussed value proposition. of artificial intelligence, or Apple intelligence, rather. But personal context means giving access to personal data. Apple is not oblivious to this, and in fact, we're pretty aggressive about the privacy fud when it comes to AI. Not only did they focus on device processing, which everyone expected, but they introduced something new called private cloud compute, that when an AI request requires more processing capacity than the device has,
Starting point is 00:12:31 it can use this special newly purpose-built cloud that are powered by Apple Silicon and they say set a new standard for privacy as a way to deliver against that request. A lot of people are really excited about this concept. It's interesting to see Apple trying to figure out how to not just do on-device sort of edge computing, but also extend that type of privacy to the cloud. But at the end of the day, they are still asking users
Starting point is 00:12:53 to give them access to a huge amount of personal information. They've committed to the idea that your data is never stored, that it's used only for your requests, and that there are verifiable privacy promises. But as Andrew Curran pointed out, A big part of the value of Siri's new capabilities is that it has on-screen awareness. In other words, Siri will be able to see your screen. This is not all that dissimilar from Recall, the very controversial feature that we had Microsoft
Starting point is 00:13:17 announced a few weeks ago. It's not that Apple can't thread this needle, and anyone who even wanted to be real about privacy was going to have to try to do this, or something like it. But it's also not easy. Last observation before we talk about what the community thought. About 50 minutes in right when Andre Carpathy was tweeting about how everyone was kind of wondering where AI was. Apple started talking about its new iPad calculator app. I thought at first they were joking and they were just about to go into AI. But no, they actually gave
Starting point is 00:13:44 more time to calculator and notes on the iPad than basically anything else in this entire presentation. Linus Eckinstrom writes, Apple, a $3 trillion company finally ships a calculator to the iPad. Not sure I should laugh or cry or both. There were many memes going around, like a picture of Tim Cook, with a remix of Apple's famous slogan this time saying, think more of the same. Still, others pointed out that actually this might have been the most state-of-the-art technology presented in the entire two hours of this event. Marquez Brownlee writes, okay, you know what, that's sick. Math notes equals write down a math problem with Apple Pencil and the app solved it immediately.
Starting point is 00:14:19 Also, when you erase one number and change it, it changes the answers automatically. Aaron Levy from Box writes, iPad calculator is actually pretty nuts. And Darrow Boshenjo, who is often a pretty skeptical dude, writes, Three years ago, Craig Frederigi told Marquez Brownlee that Apple hasn't built an iPad calculator because they want to build one that's distinctly great. They just did, and this is honestly like magic. I'm filled with childlike wonder.
Starting point is 00:14:42 I almost feel like they dropped this one little thing here just to let us know that even though their focus is so clearly about as they put it AI for the rest of us, right? AI for Normies. And that that means being focused on really simple, clear, useful use cases that come in day in and day out. that they still have the chops to do some amazing things with state-of-the-art technology as well, and that it's probably not the last we've heard from them. Yes, I got all of that from a calculator and a notes app. So what did people think?
Starting point is 00:15:14 There was definitely some skepticism. I saw a couple tweets along the lines of, well, that was two hours of my life that could have been a tweet. But by and large, it's very clear that Apple still has a lot of goodwill and that people are really buying this AI for the rest of us idea. Steven Sinovsky of A16Z writes, this integration is so exactly right. It shows that a chat app was a short-term demo UI.
Starting point is 00:15:35 Combined that with all the context on the operating system and the privacy, and this is all how AI can really come to life the right way. Kelsey Hightower writes, Apple is reading the room on AIML. Personal context is how I want to think about my data. The last thing I want to do is send my data anywhere. Logan Kilpatrick, formerly of OpenAI, now of Google, Siri becoming 10 to 100 times more useful
Starting point is 00:15:55 is genuinely going to change so many people's lives. really excited for Apple to get to deliver this experience to their customers, including me. Andre Carpathy wrote a really thoughtful response to all of it, saying, actually really liked the Apple intelligence announcement. It must be a very exciting time at Apple as they layer AI on top of the entire operating system. A few of the major themes. One, multimodal I.O. Enable text, audio, image video capability, both read and write.
Starting point is 00:16:19 These are the native human API, so to speak. Step two, agentic. Allow all parts of the operating system and apps to interoperate via function calling. Step 3. Frictionless. Fully integrate these features in a highly frictionless fast, always on, and contextual way. No going around copying, pasting, information, prompt, etc. adapt the UI accordingly. Step 4. Initiative. Don't perform a task given a prompt, anticipate the prompt suggests and initiate. Step 5. Delegation hierarchy, move as much intelligence as you can on device, but allow optional dispatch of work to cloud. Step 6, modularity, allow the operating system to access and support an entire and growing ecosystem of LLMs.
Starting point is 00:16:54 And step 7, privacy. We're quickly heading into a world where you can open up your phone and just say stuff. It talks back and it knows you and it just works. Super exciting and as a user, quite looking forward to it. To the extent there was any critique, it was that this is not Apple's native technology. But ultimately, Apple has always been making a bet that the combination of distribution with an existing install base and actual useful integration into those experiences that people have already opted into was going to carry the day.
Starting point is 00:17:23 To the extent that you think that the LLMs powering these systems might become commoditized, this has to be some of the strongest evidence we've yet seen. Overall, it's going to be a really interesting time. Will Apple actually bring AI to the masses? Only time will tell, but they're sure going to take a big shot. That, however, is going to do it for this AI Daily Brief. Appreciate your listening or watching as always, and until next time, peace.

There aren't comments yet for this episode. Click on any sentence in the transcript to leave a comment.