WWDC For the rest of us
If you’re reading this, it’s too late! This article was released exactly one week ago. To get all of our articles as soon as they’re available, consider subscribing to our newsletter at www.caju-creative.com/newsletter
June 14th, 2024 - CaJu’s Fresh Picks 03
WWDC for Marketers, Producers, Analysts, and Armchair Futurists
WWDC For The Rest of Us
Key Announcements and Their Impacts on Marketing, Production, and Data
A Word on Spectacle
Theatrics vs. Content in WWDC
Apple TV+
Streaming Services Comparison: Netflix vs. Apple TV+
Vision Pro
Challenges of Optical Computing and Augmented Reality
iOS 18
Control Center, Privacy Features, Satellite Messaging, and Apple Pay Updates
Audio and Home
Machine Learning in Audio Enhancement
iPad OS
Professional Use and Math Notes
Mac OS
iPhone Mirroring, Safari Enhancements, and Page Highlights
Apple Intelligence
Privacy Concerns and Implications for Communication and Creativity
Conclusion
Philosophical Considerations and Future of AI in Daily Life
WWDC For The Rest of Us
This week was Apple’s annual World-Wide Developer’s Conference which more and more every year becomes less for software developers and more for shareholders as Apple unvails slews of new proprietary UI, UX, and Development tools that disrupt the way millions of people use billions of devices around the world. Small changes to iPhone user interfaces can equate to the destruction or creation of entire industries so I always find it helpful to keep a watchful eye on the conference.
This year was the essential WWDC. There were big announcements and small announcements, and I don’t want this to turn into a recap or a full review. There are plenty of other shill YouTube channels for that.
No, what I’d like to do with the newsletter this week is to go a step beyond what Apple has announced, to contemplate the various ways these announcements may affect our “pillar” industries; namely Marketing, Production, and Data. How can some of these new software changes impact your workflows and how will it impact your customer’s content consumption pipelines? Let’s dive right in.
A Word on Spectacle
I want you, in the coming years, dear reader, to be aware of when millions are being spent to hypnotize you. A great example would be the constant theme of this WWDC crossing into action-movie territory. It all looks very cool. It all looks like it’s geared towards getting a wider audience interested or sharing memes and clips. Ultimately, this is a virtual software announcement. No amount of tongue-in-cheek should change that impression for you. Between this and Canva’s recent rap show, it feels like we are losing content for form as HBO’s Silicon Valley slowly becomes prophetic. Do not get lost in the flash.
Come on baby, don’t fear the details.
After today, I’m going to try not to talk about Apple for a while.
Okay. In chronological order, everything from WWDC 2024 we think will matter to our readers, here we go.
Apple TV+
You likely don’t produce shows that are getting picked up on Apple TV+ so I won’t spend time talking about Apple’s new TV UI or show lineup, but I did want to use this as an opportunity to quickly discuss content preferences between streaming services. Some take radically different approaches to content production that you can learn from.
Netflix is not simply a production company, they’re a data company. They produce a large amount of content because a higher quantity allows for a greater number of data points to track. That means they can better track and predict audience preferences quantitatively, but it also means views are spread thinner across offerings. Compare the wide and vast variety of C+ shows that Netflix offers to the more limited choice offered by Apple TV+. Lower quantity means less data-driven decision-making, but apparently there’s no accounting for taste in massive datasets.
I’ve heard it said that Netflix and most of the other streaming platforms operate the same way buffet restaurants do. Subscription models only work if you have a constant stream of content ready for people every time they come back to the table. It doesn’t have to be the best quality, it only needs to fill the countertop (so to speak). Apple TV+ and Max, on the other hand, are better represented by fine dining. Your options are more limited, but you know that you’re paying for quality.
A way in which this metaphor fails is that Apple TV+ costs less than some of Netflix’s plans.
I’m not going to say which I believe is the better model, that’s admittedly not my wheelhouse, but I would urge those in the marketing, production, and data fields to see which model performs better over the coming years. You can churn out a lot of “meh” content or take a big swing at fewer, higher-quality pieces of content. Both approaches are valid, but both may not work for your purposes.
Vision Pro
Again, not much to say here, but maybe it’s worth talking about why I don’t have much to say.
Optical computing is not taking off as fast as Apple had hoped. I’ve always held that augmented reality is bogged down by two factors: looking like a dork and cost vs function. Firstly, there is a social stigma against paying for a device that is that expensive and fundamentally performs no tasks better than a given iPad could. It’s pure spectacle. Similarly, a years-old function on most smartphones is AR guidance on Maps. Hold up your phone, and your walking directions are overlayed on what your camera sees. Nobody does this because, while it’s an interesting novelty, functionally it doesn’t accomplish anything that using your regular maps app couldn’t also accomplish much more discretely. It also makes one look like a dork in much the same way saying “Hey Siri” in public does. A facet of using Siri which Apple has now acknowledged, allowing one to shake their head to respond silently to their virtual assistant in public.
All this is to say that I wouldn’t worry too much about visual computing being the next thing to invest in for your content. The user base for virtual reality devices is so small and niche as of right now that the headsets are designed to appropriate content made for more traditional devices into the virtual space. Perhaps, someday, there will be some merit in investing in 3D advertising and content for augmented reality spaces, but as of right now, it’s more trouble than it’s worth IMHO. However, on the other hand, if someone is willing to drop 4 grand on a headset that does less than their phone does on a fraction of the battery power, maybe they are a mark.
In 2012, in my undergraduate studies, I was told that 3D video was going to be the next big thing. Obviously, it never caught on permanently, but it was enough to get me noticing that these trends come in 3-5 year-long cycles. That is not the only similarity 3D and VR technology has to cicadas.
There were headsets before Vision Pro and there will be again after it goes the way of the iPod. Don’t let it occupy too much of your mind in the meantime.
iOS 18
The new control center is cool. Homescreen personalization is meh and honestly a little surprising given how strict tech companies are with how their branding is positioned. Wanna make your Facebook app bright green? Go right ahead! Perhaps there was some merit in the old Steve Jobsian graphic design philosophy of severely limiting a user’s ability to destroy the aesthetic of their own device. Truthfully, I’m of the opinion that you can never go wrong giving the user more control regardless of how much it “clutters” the UX. Minimalism is great for art museums, but in the realm of technology, it can be a plague.
Text effects are cute.
New and improved privacy protections are good. It shows how Apple continues to position itself as the only large tech company that isn’t trying to sell every part of your personal data to advertisers. Though, again, this may be to position itself as a premium service. Yes, one must pay more for an Apple device, but included in the cost is the peace of mind that your personal data won’t end up scraped and stored in some server somewhere. More on that point later.
Messaging by satellite seems dangerous. I get satellite messaging for emergencies, but I can’t imagine satellites have the bandwidth to deal with millions of devices trying to send animated poop emoji stickers and people trying to get rescued from rock slides at the same time.
New email categorization might make junk mail ads less effective. That’s not a bad thing. If your email blasts are getting caught in smart junk filters, it’s likely because your email blasts are junk.
**Record Scratch** Not this one though, whitelist “CaJu.Creates@gmail.com” today.
There’s no sense in buying someone’s email from a data broker and sending them a newsletter they didn’t sign up for. You’re just going to make them resent you and possibly break the law in the process. Smart email categorization might be a good wake-up call for spamy email advertisers if it’s widely adopted. Everything – every single thing – that you put in front of your customers or audience needs to have some kind of use value for them.
Apple Pay online and Cash to Phone will make it much easier to accept payments in person. This is a big one. This is one of those innocent enough features that has the potential to destroy entire industries and create entirely new social habits. Social habits, for the uninitiated, are the holy grail of marketing. When you try to sell a product, ultimately you are trying to get a customer to change their habits by adopting new ones at your suggestion, a very difficult feat to accomplish. The ability for any two people with iPhones to exchange money simply by tapping their phones together has the potential to create those much sought-after habit disruptions. Anyone working in public events or fintech should keep an eye on this space.
Audio and Home
I’m only including this section to point out the prevalence of the phrase “machine learning” in this announcement video. This is the first time I heard it, but it comes back many more times. For example, they say that Apple TV will now offer an Enhanced Audio option that uses “machine learning” to lower music volume so you can actually hear actors’ voices more clearly during tense scenes with loud soundtracks. This begs the obvious question, who is doing Apple’s sound mixing so poorly that it would need a machine learning patch if not Apple themselves?
Two things here strike me as interesting. The first is that Apple is deliberately using the phrase “machine learning” here to differentiate these smaller features from the more sophisticated “AI” features to come later. This is strange because the phrase “machine learning” has kind of gone out of style lately. Most companies nowadays just say “AI” in its place despite everything that “AI” does being more technically and accurately described as “machine learning.” Secondly, this is interesting because it illustrates the kind of band-aid that machine learning has become lately. No need to mix your TV series well, that could get expensive down the road! Just do a slapdash job and then make an ML model to clean it up in the future.
All in all, very silly. I include it to point out interesting modern facets of machine learning and to illustrate the lengths that companies will go to these days to avoid hiring sound designers. I promise you, it’s always better to have and not need than need and not have when it comes to audio professionals.
iPad OS
I saw a review recently about someone who tried to replace their laptop with an iPad and did it successfully but with a lot of extra hassle. I personally don’t know any professional who works in production who uses an iPad for anything more than a note-taking device. That said, I’ve used an iPad to edit video and audio in a pinch, here’s my recommendations for that: For video, Luma Fusion. For audio, Logic Pro. For photos, as of right now, I use Lightroom but honestly, I’m at my wits end with Adobe so I may try switching and I remember Afterlight being good from years ago. Of course, my biggest tip for editing on iPad is to grow up and buy a laptop.
I didn’t see much else that I believe would be impactful for our readers save for possibly “Math Notes” which could be helpful for anyone working with data. Math notes will let a user draw math equations freehand and then display the answer in a replicated version of the user’s handwriting. Given that it understands variables, it may be helpful in writing pseudo-code or statistical approaches to problems during meetings where data tactics are being discussed. Anyone who has worked with data analysis will know that good and effective note-taking is often where the majority of the job’s difficulty comes from.
I also noticed a small mention of the ability to produce real-time graphs. If that could be integrated with tables in notes, it would be an amazing productivity tool or presentation tool being able to load a dataset and generate histograms or boxplots within seconds. Hey boss, as you were talking, I implemented and visualized the changes you wanted. Here they are.
Yeah, that might feel pretty cool.
Mac OS
Decent amount to talk about here. I’ll try to be brief.
Math Notes on Mac, I think even more useful than Math Notes on iPad.
iPhone mirroring seems like a novelty but could be useful if you’re using your iPhone as a camera on a production shoot and want to use a computer as a monitor. iPhone mirroring also allows anyone who shoots on an iPhone and edits on a Mac to more seamlessly transfer files to and from the phone into your video editing program of choice.
Safari doesn’t have the widest adoption rates, but I will say as a person who only uses Chrome when forced to, Safari is lean and fast and very good with privacy. Now it’s getting even better. If more widely adopted, it could severely impact the amount of data that advertisers and data brokers are able to collect from users online.
The new “highlights” feature which summarizes page content will need to be further studied and I mean that literally. Those who work in web design will need to know exactly how the algorithm decides what is important on a webpage. It would really suck to put a lot of thought into some web content only to have Safari filter it out as irrelevant for a user’s “highlight” of your webpage. Whenever this feature is available and I’ve had some time to work with it, I will follow up with exactly that in a future newsletter.
I work in marketing and even I’m excited for Reader and Viewer to improve the absolutely horrendous, ad-filled web experience on certain websites. As mentioned in the previous CJFP edition, Google’s emphasis on SEO tactics really destroyed the usefulness of vast swaths of the internet. Oh, to see a recipe blog, ad-free.
Apple Intelligence
An LLM in every iPhone. But at what cost?
In brief, Apple is attempting to own the term AI by terming its new approach to large language models as “Apple Intelligence.” It’s being billed as a privacy-first LLM that is small in scale but has access to larger servers when needed, all without selling your personal data down the river. Off the bat, as you can imagine, I’m skeptical.
In their presentation, Apple said privacy must be key in the functionality, this is fine, but it is to say nothing of the public data used to train such models. Questions posed to Siri can be anonymized, but still represent, in bulk, a mass spying operation on users’ preferences. Data anonymized is still valuable and wiping someone’s name off something they own doesn’t stop it from being theirs.
Apple Intelligence will have three main facets: language, images, and actions. To get this out of the way, for Large Language Models to generate anything, they need training data. Apple says it will use your own writing style as training data and rewrite anything that needs writing in your own personal style. That gives me some worries I’ll discuss shortly, but in theory is fine. Apple Intelligence will also, however, summarize any other kind of textual data you receive on your phone be it from a text message, email, or website. It had to be trained on how to do this and Apple has said the training data was a mixture of data they acquired legally, data generated themselves, and “publicly available” data scraped from the internet via their AppleBot crawler. That’s the worrying part. Posting something publicly online is not the same as permitting multi-billion dollar companies to take it and use it for free. This only incentivizes the further firewalling and privatizing of useful information in public online forums in order to avoid scraping by competitors and contributing to the further degradation of net usefulness.
It’s a similar story with images. Apple Intelligence will let you generate images directly on the phone. One can’t help but wonder how many artists on the internet were ripped off to allow for this ability. It could be less than we think, but if that is the case Apple is not being very forthcoming about it.
The “actions” of Apple Intelligence refers to its ability to translate spoken, natural language into instructions for the computer. That is the part I think will lead to genuine breakthroughs in the way we interact with our technology. That has always been the true superpower of GPT models.
A Jetsons-esque personal assistant was always the vision for Siri and now it seems the tech has finally caught up with the idea. From a business perspective, every person having access to a virtual assistant does have the potential to streamline a lot of daily tasks like meeting summarization and scheduling. As someone who takes a lot of notes, the ability to transcribe from audio or phone calls on-device is a powerful feature indeed – one that also has the potential to damage transcription companies like Rev and DeScript if done well. And as a person who characteristically sends very detailed work emails, knowing that I don’t have to rely on the recipient’s attention span, but rather their personal assistant’s ability to summarize my email into actionable items based on their personal context does make me smile.
A fear of mine is the alienation that may come from knowing anything sent to you by one of your contacts was generated and not actually their words. For instance, I receive an email from someone that I suspect has been written by AI. I now feel no obligation to read it. I have my AI summarize it and draft a response. Our AIs are now communicating on our behalf. This reminds me of the philosopher Slavoj Zizek’s idea of interpassivity or media technologies which perform the function of enjoyment for us so we can relax much like a sitcom’s laugh track laughs for us so we’re more comfortable watching stoicly.
The language recommendation features will certainly put companies like Grammarly out of business. But I also fear that nobody will be able to trust a single thing they read online was written by a human being again. As I’ve said previously, AI-generated text tends to be the lowest common denominator, but just good enough for everyday use. True, human-generated text, when done creatively, has the potential to come at a premium, but that’s assuming it’s actually being read. In a future where every piece of writing on the internet is being summarized, what becomes of creativity? There’s that interpassivity again. I use AI to generate my web page’s content, you use AI to summarize it. We are not actually communicating, our AIs are. But I digress.
Anyway, pause on the fearmongering, the industry-related specifics now.
While, as mentioned, I have my reservations with AI-generated images, putting this technology on every iPhone has the potential to seriously devalue AI-generated images by reducing the barrier to entry for nearly everybody. For content producers, this may even represent a cheap and easy way to generate simple assets that one would otherwise get from Creative Commons libraries. I also think that it is about the level of quality it will max out at. If you need anything more sophisticated than a royalty-free vector image, it’s probably best to go to the source: an actual graphic designer.
It’s really not a matter of what local data the models are using, it’s the training datasets that are where the controversy lies. This extends further than Apple and Open AI, Adobe recently got in trouble for their TOS mentioning that they can use anything created using their programs or stored in the Adobe cloud as training data for their GenAI. They did eventually have to walk that back.
AI image generation seems to be something Apple is really leaning into. Strange coming from a company that claims to speak on behalf of artists and creators.
For Siri, I do wonder how much of these “personal assistant” tasks can transfer into the professional sphere. For instance, could I tell Siri to write this newsletter for me? It seems like it would request permission to utilize ChatGPT and then do it. If that’s the case then we’re no longer utilizing my own “Personal Context” or writing style, it would just become a generic ChatGPT text block. I think many of these next-generation features will have implications for a lot of online communications, not the least of which will be cold calls. An AI writes all my cold emails, your AI reads them all. It’s AIs talking to AIs all the way down.
Didn’t think I’d be mentioning Zizek twice in one newsletter – or even once for that matter – but during the height of the COVID-19 pandemic, he got in trouble for suggesting that a positive aspect of the virus was making the average person think philosophically. I think this point was correct. How much did we hear about personal freedoms and government overreach and what we owe to one another? Every person affected by COVID was forced to take a philosophical stance on issues that, just a year earlier, would’ve been simple thought experiments. I believe Apple bringing LLM access to so many of their devices will have a similar effect.
For instance, Apple will now allow you to, with a touch, remove unwanted objects or people from photos. Google for some time has used AI to change their photos like to make your kids smile when they weren’t in any of the photos you took. This not only directly impacts the professional photography industry and (hopefully) Adobe’s bottom line, but it also, on a more philosophical level, forces us to think about what a photo taken on a phone is meant to represent. Is it a rendition of a memory as it was or an impression of a memory as we’d like it to be remembered? Changing a photo to portray a moment differently than it happened is one thing when done for more artistic endeavors, but it seems suspect to enable by default for family photos. Sometimes unideal photos make for the most precious memories, but anyone given the option to make their photos “perfect” will certainly take the opportunity.
What do we owe to one another? Why do you enjoy reading messages from your friends in all their winding complexity? Why do you despise even the shortest work emails? What makes AI-generated art so unappealing to look at? Will that be enough to stop Samsung from putting GenAI in our kitchen appliances? The more the public begins to think philosophically about these issues, the more advertisers and marketers will need to have their fingers on the pulse of the zeitgeist.
Boom. Brought it home.
Using GenAI to make a bedtime story for your child sounds tantamount to negligence, but Apple thought it was worth suggesting as a selling point. If that interests you, my suggestion would be to have your personal AI tell a bedtime story to your child’s AI and while it does, you and your kid can both go read a book together.
Until next time, stay fresh.
- Casey