Big Week

Welcome back to the AI Geekly, by Brodie Woods, brought to you by usurper.ai. This week we bring you yet another week of fast-paced AI developments packaged neatly in a 5 minute(ish) read.

TL;DR WWDC wonders; GenAI video thunders, MSFT blunders

This week we have an eclectic mix of stories for you. Headlines have been dominated by news of Apple’s latest AI releases at its WWDC conference this year. Privacy and security remain top priority while the company fires a salvo of impressive new bespoke AI capabilities, perhaps the most integrated offering seen to date, in a device many already own (but will probably have to upgrade, but at least now you have a good reason!). GenAI video generation is back in the news with Luma’s new Dream Machine (Dream Weaver belongs to the Gary Wright estate, sorry), which democratizes access to video generation capabilities via text prompts. Is it going to usurp the major film studios? No… not yet… just remember, this is the worst this technology will ever be. It only gets better from here. Finally, we’ll take you through one of the classic blunders (no, not starting a land war in Asia) as Microsoft takes its ready, fire, aim approach to releasing new big brother capabilities for its new Copilot+ PCs.

WWDC GOAT
Perhaps its Greatest of All Time

What it is: In classic Apple style, the company showcased its ability to “Think Different”, unveiling the fruits of 1+ year of effort to build-out its AI offerings at this year’s WorldWide Developers Conference (WWDC) via its announcement of its own flavor of AI dubbed Apple Intelligence (because, of course). Apple Intelligence features a suite of AI-driven capabilities integrated across its mobile operating systems (iOS 18, iPadOS 18) and desktop/laptop OS, macOS Sequoia, and more. Key highlights include the integration of ChatGPT with Siri, advanced photo editing tools, and AI-generated emojis. The event also featured updates to VisionOS, enhancing the Vision Pro headset by opening up the developer kit to allow greater customization (we should note here that every AI Geekly since early February has been written exclusively using the Apple Vision Pro —tech pundit haters be damned!)

What It Means: Apple has had to play a bit of catch-up in the Generative AI (GenAI) space relative to Google, Microsoft and Amazon. With no formal GenAI partnerships until today (with OpenAI) the company has thus far focused its efforts on releasing open-source models it hopes will be adopted by the developer community to advance its tech gratis. So far we’ve seen a few devs take them up on that offer with some interesting implementations of on-device LLMs whereby quantized models have been squeezed into iPhones (an impressive feat given the limited hardware of mobile devices and the high-spec requirements of most AI models). By partnering with OpenAI, Apple aims to enhance user experiences through personalized intelligence and context-aware interactions (read: a smarter Siri with contextual understanding across apps). This move positions Apple as a serious contender in the AI-driven tech landscape, leveraging its hardware and software ecosystem to deliver (presumably) seamless, intelligent user experiences. We’ll admit we found the demos impressive.

Why It Matters: The integration of AI across Apple's product lineup represents perhaps the greatest surface area to-date for broader exposure to generative AI across the population. With millions of iOS devices globally, as Apple begins to roll-out the new capabilities to its user base, many millions of users will experience ChatGPT and GenAI tools for the very first time (today, you still have to visit chat.openai.com , but with the release of iOS 18, ChatGPT comes to you!). For developers, the new AI capabilities offer powerful tools to create more engaging and contextually aware applications. For consumers, these advancements promise a more connected and efficient digital experience, reinforcing Apple's reputation for innovation and user-centric design. Once again, Apple is playing to its strengths. We applaud the company for taking something as complex as GenAI and breaking it down into a presentation that is understandable by all.

One more thing: Hard to believe, but the word is that the OpenAI + Apple deal that brings ChatGPT to Apple devices includes no monetary compensation. OpenAI hopes the exposure will lead to increased subscriptions to its Pro tier offering, as does AAPL (it collects 30% of all app store sales). Hats off to Tim Cook for negotiating this Jobsesque sweetheart deal.

Lights, Camera, Text Prompt!
Luma Labs leads the way to open video generation

What it is: This week we’ve got some new challengers in the increasingly competitive AI-generated video space as Luma AI launched Dream Machine: a free, browser-based AI text-to-video generator that produces high-quality, realistic videos from text prompts. This new tool rivals exclusive models like OpenAI's Sora and Kuaishou's Kling, which are currently accessible only to select testers. Dream Machine generates five-second videos in about two minutes, offering an intuitive interface for creators. We’ve taken it for a spin ourselves, check-out our video of a robot playing with a toy here (note the video may be a bit choppy due to high-usage of the site from its newfound popularity)

What It Means: GenAI video generation has just gone public (an A-IPO?); Dream Machine democratizes access to advanced AI video generation, giving everyone on the planet (with an internet connection) the ability to generate high-quality (albeit short for now) video content for free. Unlike Sora and Kling, which remain exclusive, Dream Machine is open to everyone, allowing a broader range of users to experiment and create. We expect this to prompt peers to expedite their own public releases to stay competitive.

Why It Matters: Dream Machine's accessibility and impressive capabilities mark a significant step forward in AI video technology. It’s not the shiniest, the best performant, or the longest, but it is available to everyone, which is a huge leap forward for the space. While it still has limitations, such as occasional unnatural movements and blending issues, its potential for rapid improvement through open-source development is promising. This democratization of AI video tools could lead to new creative possibilities and innovations in digital storytelling. As long-time readers of the Geekly know, we are big proponents of democratization of AI via Open Source as a check on the power of big tech and government. By making these tools widely available, we reduce the wide capabilities gap between the haves and the have-nots.

Total Recall
Microsoft recalls its Recall feature due to criticism

What it is: Microsoft found itself in hot water this week over its Recall feature, part of the new Copilot+ PC offering (previously covered in the Geekly). Recall captures regular screenshots of user activity, storing them locally to augment MSFT’s AI copilot by providing deep context through access to previously viewed content. Initially, the feature was set to be enabled by default, but following significant criticism, Microsoft has decided to make it opt-in and require biometric authentication. The company also announced that the database storing these screenshots would be encrypted to enhance security because, shockingly, it wasn’t before.

What It Means: The decision to make Recall opt-in and to encrypt the database addresses major privacy and security concerns raised by experts and users. Critics rightly argued that the feature, as initially designed, could be a "dangerous magnet for hackers" and a "privacy nightmare" due to its potential to expose sensitive information. Wow guys, just wow. Everything I do on my computer captured in snapshots saved in an unencrypted folder that just sits there waiting to be stolen; is this your first day? Perhaps then it’s no surprise that MSFT was raked over the coals this week by the U.S. House Homeland Security Committee reviewing the company’s cybersecurity shortcomings which have led to several high-profile leaks.

Why It Matters: The controversy surrounding Recall highlights the ongoing tension between innovative technology and user privacy. It’s precisely because of missteps like this that Apple is favored as the most privacy-centric tech company. We didn’t mention above, but Apple Intelligence, and indeed AAPL’s entire approach to GenAI is privacy-centric. Where possible, all AI compute is done locally on device. Where not possible, data is transmitted to newly-built privacy-focused servers owned and maintained by Apple, and running its own proprietary M-silicon chips in a supply-chain hardened environment (read: limiting infrastructure backdoors that may slip in from unscrupulous manufacturers and government entities). Privacy and ethics are critical elements of consideration in our quickly evolving AI world. AAPL presents a master class in doing ethics and privacy correctly while MSFT is just the opposite. Glad they’ve rolled-back their aggressive Recall moves but they score no points with us —their approach underscores their core beliefs: privacy and ethics are second fiddle to shiny new things. Caveat emptor.

About the Author: Brodie Woods

As CEO of usurper.ai and with over 18 years of capital markets experience as a publishing equities analyst, an investment banker, a CTO, and an AI Strategist leading North American banks and boutiques, I bring a unique perspective to the AI Geekly. This viewpoint is informed by participation in two decades of capital market cycles from the front lines; publication of in-depth research for institutional audiences based on proprietary financial models; execution of hundreds of M&A and financing transactions; leadership roles in planning, implementing, and maintaining of the tech stack for a broker dealer; and, most recently, heading the AI strategy for the Capital Markets division of the eighth-largest commercial bank in North America.