AI Wearables: Is There a There There?

Today on the AI breakdown, we're asking if there's really any demand for AI wearables before that on the brief certain versions of Llama 3 are expected to drop this week. The AI breakdown is a daily podcast and video about the most important news and discussions in AI. Go to breakdown.network for more information about our YouTube, our newsletter and our Discord. Welcome back to the AI Breakdown Brief. All the AI headline news you need in around five minutes. Today we kick off with the most anticipated drop of this week. Now, of course, this being AI, who knows what might get released? But one thing that we do know for sure is that it appears that some of the smaller versions of Meta's forthcoming Llama 3 models should be coming out this week. This news was first reported last Monday by the information they had it from a Meta employee that Meta was going to launch two small versions of its third generation open source Llama network, which would precede the large version, to be released later in the summer. While initially a Meta spokesperson declined to comment, the next day at a public event, Meta made it seem like, yes, these models were coming, and they were coming very soon. This is exciting to the AI community for a couple of different reasons. First of all, Meta has made it clear that their objectives when it comes to Llama three have shifted. While previously they seemed, at least for the time, content to be state of the art, when it came to Open Source Models, Llama 3 They've promised to be a GPT4 class model at least. And indeed Zuckerberg has said that going forward they want Llama models to be able to compete at the state of the art in general, not just for open source. That, of course makes the big release the largest version that would come out in the summer. But there's still reasons that people are really excited about even these smaller versions. A couple big ones. One, of course, is cost. Smaller models tend to be cheaper, and a lot of what's going on in AI development right now is arbitrage between cost and performance, given how expensive the most advanced state of the art models are. But there's also differences in where small models can run. If you're looking to build, for example, mobile device applications that don't necessarily rely on the cloud or AI PC applications, once again that don't have to rely on Google Cloud or Amazon Web Services or Microsoft Azure, advances in smaller models actually matter more than state of the art when it comes to LLMs. The three versions of LLAMA 2 that we got had 7 billion parameters, 13 billion parameters, and the largest had 70 billion parameters, and the Information has previously reported that the largest version of Llama 3 could have more than 140 billion parameters. Anticipation is heating up. Hyperrite AI CEO Matt Schumer says Llama 3s will start to drop next week. Assuming there's a 7B version, I'm expecting it to far surpass the current Mistral model. Now something I've said on this show before is that up until Mistral started releasing its models, Meta was the undisputed open source champ in terms of dominating where developer attention was with the release of Mistral models, they definitely started to suck some of the oxygen out of Meta's room. And so this is a chance for Meta both to reclaim that open source space, but also to try to push into the general LLM performance space as well. Next up, something at the intersection of AI and geopolitics One of the big questions right now in US policy is whether TikTok is going to be outright banned or, as many are advocating, forced into a sale to a US Owned company for the U. S facing version. One of the people who is trying to orchestrate that type of deal is Steven Mnuchin, former Treasury secretary under Donald Trump. According to the New York Post, Quote, insiders say Mnuchin is pitching a plan to rebuild TikTok's algorithm in the US as the only likely way to satisfy congressional concerns that the app poses a national security risk as well as China's strict export laws that could block a sale. So basically where we are with this situation is that the House has voted to pass a bill that requires ByteDance to divest TikTok within six months or face a total ban. But the devil is going to be in the details. ByteDance doesn't want to give away its proprietary algorithms, even if that's a lot of where the value lies. And so trying to get it ahead of that, Mnuchin and his consortium are trying to find an AI partner that could rebuild that particular capacity. One partner that has been mentioned, although not necessarily in this context, is Oracle. Even before the ban push reached its recent apex, TikTok had committed to storing the data of US and servers operated by Oracle in order to try to head off those issues. This isn't the first time that an Oracle TikTok hookup has been discussed. During the Trump administration, when Mnuchin was Treasury Secretary, Oracle almost bought TikTok then as part of that administration's push to get ByteDance to divest as well. Now, on top of all the concerns and issues with China. Another big question is what's going to clear regulatory scrutiny, particularly when it comes to antitrust. What's on going One of the reasons that an Oracle might be a better candidate than someone like Google Meta or Apple is in fact those antitrust considerations. All of this is very weird, uncharted territory, and frankly has been for some time. Back in 2020, when Microsoft tried to buy TikTok in a similar context, CEO Satya Nadella later said that the aborted bid was, quote, the strangest thing I've ever worked on. Now over in Adobe land, a bunch of new announcements today. First of all, Adobe Acrobat has released an AI assistant that starts at five bucks a month. This is basically PDF intelligence as a service. Now, given how popular PDF reading type features were in the early ChatGPT app store, it makes sense that Adobe's trying to grab a piece of that for itself, given that it owns the core PDF reader in Acrobat. Additionally, however, Adobe announced that generative AI is coming to Premiere Pro. This is of course its video editing software, which is hugely used across basically every industry that touches video at all. Now, in addition to Adobe's proprietary Firefly models coming into Premiere Pro in the same way that they have in software like Photoshop, Adobe also surprised by announcing that it was developing a way for users to integrate third party tools from companies like Runway, Pika Labs and OpenAI. The details have not yet been figured out. Adobe said that they haven't settled how revenue generated by third party AI tools would be split up between Adobe and outside developers. But the fact that they're announcing this at all suggests they're at least a little bit nervous about competition from these new generative AI video tools and want to keep everything in their ecosystem. This is a good distraction from the end of last week when Bloomberg reported that despite a lot of hemming and hawing about their ethical approach to Firefly training, which was sort of pitched on the idea that Adobe was only training their models on images that they had rights to, it seems as though Firefly was also trained on midjourney images, which of course if you're concerned about where midjourney images are coming from, kind of just seems like the AI data training equivalent of money laundering. Still, mostly I am just excited to see inpainting come into video editing, and so I for one am excited to see what Premiere Pro does with AI. That's going to do it for today's AI Breakdown brief. Next up, the main AI breakdown. Here's a story that is all too common. Your business gets to a certain size and the cracks start to emerge. Things you used to do in a day are taking a week. You have too many manual processes. You don't have one source of truth. If this is you, you should know these three 37,025 one 37,000. That's the number of businesses which have upgraded to NetSuite by Oracle. NetSuite is the number one cloud financial system streamlining accounting, financial management, inventory, HR and more. 25 NetSuite turns 25 this year. That's 25 years of helping businesses do more with less, close their books in a day, not weeks and drive down costs. One because your business is one of a kind. So you get a customized solution for all of your KPIs in one efficient system with one source of truth. Manage risk, get reliable forecasts and improve margins. Everything you need to grow all in one place. Listen, I know a lot of you guys are entrepreneurs out there and one of the things that is a reality of entrepreneurship is that almost from the moment that you've started, you have more information in more places than you possibly know what to do with. That gets nothing but harder the bigger that you grow. And frankly, right now with AI, companies are growing bigger, faster than ever before. If you're facing any of these challenges right now, you can download NetSuite's popular KPI checklist, designed to give you consistently excellent performance absolutely free@netuite.com breakdown that's netsuite.com breakdown to get your own KPI checklist. Netsuite.com breakdown Attention AI Breakdown listeners. Consensus 2024 marks the tenth gathering for all things crypto, blockchain and web3. However importantly, this year's agenda will also dive deep into AI driven transformation and the speaker lineup includes the leading minds and innovators at the forefront of this digital renaissance. Don't miss the Consensus AI Summit to cut through the hype to find where true transformation and opportunity lie. Listeners to this show can get 15% off registration with the code AIBREAKDOWN. Visit consensus2024.coindesk.com to learn more. Some of the folks will be at Consensys this year include Guillaume Verdon, AKA Beff Jesos, founder and CEO of Xtropic, as well as spiritual leader of the accelerationist movement, Neal Stephenson, co founder of Lamina one, and Brendan Eich, the CEO of Brave Software. Again, go to consensus2024.coindesk.com to learn more and get 15% off registration with the code AIBREAKDOWN welcome back to the AI Breakdown. One of the things you will have noticed if you've been paying any attention at all to the AI space over the last call it six months is that there has been a huge explosion in the number of different companies that are interested in AI wearables. One that's gotten a ton of buzz, for example, is the Rabbit R1, which was designed in part by Teenage engineering and came out with a very Steve Jobsian introductory video. Then there are all the rumors about the collaboration between iPhone designer Jony I've and Sam Altman, who are apparently out fundraising more than a billion dollars. The reports say that they are talking with Lorraine Powell Jobs and her Emerson Collective, as well as Thrive Capital. And previously we had also heard reports that Softbank was in the conversations as well. Now, when it comes to that device, all we've really heard about its form factor is that it will not look like a phone. Speaking of phones, every phone device manufacturer right now seems to be advertising their new phone as an AI powered phone, and I'm sure that will only increase when we hear about iOS 18 at the Worldwide Developer Conference this year. We've also had big, bold pronouncements that the era of the AI PC is coming. And so all of this is to say that hardware is big when it comes to AI, and specifically this new form factor of wearables is something that a lot of people are interested in. Perhaps the poster child for this is the Humane AI pin. This has been buzzy since last year, when one of the founders of the company showed off some of its capabilities in a demo at ted. For example, there was a live translation that sounded exactly like the speaker. Well, when the pin was previewed last November, it came with what to some was a surprising price tag. The Device itself was 600 bucks and you needed a $24 a month subscription. On top of that, as the Verge put it, this wearable computer promises to free you from your smartphone. But then, echoing basically every review that we've gotten so far, they write, there's only one problem, it just doesn't work. Now, I think that there is a lot of agreement or interest in the underlying thesis behind humane, which is that it would be better if we spent less time screens hunched over this device, furiously typing with our thumbs and ignoring the world around us. So the question for the Humane AI pin in many ways was two part one Are there actually use cases that you don't want to use your phone for? That some device like this would be better? And two is this that device now. I read part of the Verge review last week, but to reread the most important part again, the author writes, I've learned I do lots of things on my phone that I might like to do somewhere else. So yeah, this is something. Maybe something big. I've seen enough glimmers of what's coming that I'm optimistic about the future. But then when that same reviewer asks, should you buy this thing? They write, that one's easy. Nope. Nuh. No way. The AI Pin is an interesting idea that is so thoroughly unfinished and so totally broken in so many unacceptable ways that I can't think of anyone to whom I'd Recommend spending the $6.99 for the device and the $24 monthly subscription. There are a couple challenges that they flag. First, the device itself is a little fiddly that you sort of have to get the position just right. Second, as they put it, the AI pin is also just incredibly unsubtle. When you stand in front of a building tapping your chest and nattering away to yourself, people will notice and everything gets in the way too, then goes through a long list of functionality that the Pin doesn't seem to have. The author continues, the more I tested the AI pin, the more it felt like the device was trying to do an awful lot and the hardware simply couldn't keep up. The battery life, they said, was rough. The projector's user interface was, to use their words, bananas. And ultimately they concluded, it feels like Humane, decided early on that the AI pin couldn't have a screen no matter what, and did a bunch of product and interface gymnastics when a tiny touchscreen would have handled all of these things much better. Kudos to Humane for swinging big, but if you're going to try to do phone things, just make a phone now. Other reviews came out that were very similar. Mark Gurman, who is Bloomberg's Apple whisperer, wrote a piece called the AI Device Revolution Isn't Going to Kill the Smartphone. Gurman had a similar litany of problems with the device. He said it can be slow to respond or fail to recognize the touch gesture necessary to begin listening to a command. It overheats to a worrisome degree and is missing key features like the ability to control timers and calendars. It can't tap into third party apps and has a frustratingly brief battery life. Ultimately, he said, the company acknowledges that the device is a version 1.0, but it feels even earlier than that. The controversy really ramped up, though, over the weekend when Marques Brownlee who has something like 18 million followers on YouTube, published a review called the Worst Product I've Ever Reviewed for now. People had an extremely strong reaction to this. Daniel Vassallo tweeted, I find it distasteful, almost unethical to say this when you have 18 million subscribers. Hard to explain why, but with great reach comes great responsibility. Potentially killing someone else's nascent project reeks of carelessness. First, do no harm. However, a lot of people disagreed with Daniel to an intense degree as well. For example, commentator Sigil writes, yes, a world where reviewers hide their true feelings about a product is so much better. Let's make sure consumers are never informed about the truth. For what it's worth, Marques posted a slightly more nuanced title on Twitter, calling the video Humane Pin review a victim of its future ambition. He said, this clip is 99% of my experiences with the pin doing something you could already do on your phone, but slower, more annoying, or less reliable and accurate. Turns out, smartphones are pretty incredible, but there are occasional small glimpses of a future where this product could actually peel you away from your phone, reduce your screen time and help you live more in the moment. And I can't be mad at that. Sam Sheffer, the head of new media at Humane, wrote an honest, solid review. Marques all fair and valid critiques, both the good and the bad. Feedback is a gift. We reflect and we listen and we learn and we continue building. Broadly speaking, I think there are two camps when it comes to the Humane AI pin. Notably, there is no camp, at least that I can find, that thinks the device is good and really particularly useful in its current format. The optimistic camp, however, is appreciative of the company taking a huge swing, recognizes the fact that totally new categories of technology are extremely hard to get right and take a lot of iteration and experimentation, and generally are glad that people are trying things even if they don't get it right right away. The flip side is that some people are convinced that the entire category is complete vaporware. In a back and forth between Daniel Vassallo, who I quoted before, and Dogecoin founder Shiptoshi Nakamoto, illym2k says his review is the product is literally pointless and bad at what it is setting out to do and doesn't need to exist. Anyone, even mildly smart, already knows that this product is a complete joke and only shows how delusional and out of touch some tech bros are. Benjamin Decracker, who is an AI entrepreneur, writes, haven't seen any evidence that people even want AI wearables. The people enthusiastic about them seem to be the same people trying to sell you one. Meanwhile, the form factors people do seem to want smartwatch, phone, assistant headphones are being nearly totally ignored. Some people have ethical issues. Stuart Reed says, I've said it before, I'll say it again. I absolutely despise the idea of AI wearables watching and recording everything we do. We need less surveillance capitalism, not more. Amit as Investing writes Wow, the biggest tech YouTuber just called the Humane AI pin the worst product he's ever reviewed. I've been on record saying most of these AI wearables are beyond nonsense and the fact that people are buying them is beyond me. It seems most people buying them are tech enthusiasts willing to give new products a shot, not the average consumer. By the way, my bearishness on wearables extends to the Apple Vision Pro. It is absolutely useless and I see zero and I mean zero use cases for it. As much as the company has been trying to put on a good face, the level of critique is clearly bothering people. At Humane, Colin Cornwell writes wild how many people want to see you fail? We'll make it all that much better when we win. Sam Sheffer again quote tweeted that and said, easy to dogpile, much harder to build. Yes, the software is not where it needs to be, full stop. And that's entirely on us. The overwhelming negative sentiment. What happened to optimism? What happened to pushing things forward? Now, regardless of what one thinks of Humane, the reality is that a lot more people are going to try to create some version of an AI wearable before the category is abandoned. Take for example the Pendant, which was just announced today by Limitless, which is a spinoff of the company. Rewind co founder Dan Siroker, who had announced a version of this device last year, tweeted this morning introducing Limitless, a personalized AI powered by what you've seen, said or heard. It's a web app, Mac app, Windows app, and a wearable. So you may have previously heard of this company as Rewind. This is a company that keeps track of basically everything that's going on on your desktop, in your meetings, and giving you the ability to get rich insights from all of that information to be able to go back to exactly where you need to to check something out. Like I said, the pendant they had announced very briefly last year, but the product vision has clearly come together in a different way. This form factor people seem a lot more excited about. It promises a hundred hour battery life. It is actually lightweight. It comes in a variety of colors that make it potentially less highly visible and it even has an interesting approach to privacy. So for example, the way that they show off how it would work in the context of a conversation is in a transcript. They show the pendant wearer, in this case founder Dan Siroker, speaking with a new speaker. With the transcript saying new speaker is speaking but hasn't given consent to be recorded. Dan then responds this meeting is really important to me and I really want to remember your advice. Are you okay if I record and transcribe our chat? The co speaker Sam says sure, no problem and then is identified and actually becomes part of the transcript. Developer Nick Dobos writes Cool touch on AI wearables Auto opt in recording consent using auto Voice ID need to see privacy concerns and solutions evolve now. At this stage I really don't have any idea which if any of these devices are likely to work. I tend to think that when it comes to completely new form factors like this, it's not going to be some one to one replacement use case that makes them useful, but something new that we don't imagine doing now that once we start doing we can't imagine a time when we didn't do it. Could limitless keeping track of everything that you've said or done be that new use case. It's totally possible. It's also totally possible that there's something else that no one is imagining yet. That is the thing. And it's totally possible that these form factors don't amount to anything and that for at least some time to come, the devices that we're already using are the way we interact with AI. I will say in general I am for more, not less, experimentation and so I'm certainly wishing all the builders luck. This is an area of building that I think is going to see more and more attention and more and more entrepreneurs. So I'm sure sure we will be talking about it again. For now though, that's going to do it for today's AI breakdown. Until next time, peace.