4.3. Successes and failures of early platforms

We’ve discussed new business models, but there’s also a lot happening with new AI-powered platforms. In this section, we’ll review a few of them, analyzing their strengths and weaknesses. Hopefully, we’ll gain insight into what works and what doesn't in the realm of generative AI.

Meta Smart Glasses

In September 2023, Meta, in collaboration with Ray-Ban, unveiled the successor to its two-year-old smart glasses. The updated version continues to be promoted as an everyday wearable, designed to capture photos and videos from a first-person perspective. Like its predecessor, the new smart glasses come equipped with built-in speakers and microphones.

So, what’s new? Meta has announced that in an upcoming release next year, the smart glasses will get multimodal capabilities. This will enable users to engage with their environment using Meta AI. During the event, the company demoed features like playing tennis while asking the glasses whether an out-of-bounds ball was a fault or not.

This is a basic example of a new software and hardware paradigm: ambient computing. Ambient computing focuses on the unobtrusive incorporation of technology into our environment to automate tasks and improve our daily lives. Imagine communicating with an assistant like ChatGPT that sees what you see and hears what you hear in real-time, continuously. This approach to information is fundamentally different from using apps on your phone and even diverges from the user experience with stationary smart speakers like Alexa, which have limited access to real-world context.

I brought up Alexa for a reason. Although I’m keen to try them, I haven’t yet experienced Meta’s smart glasses firsthand, so my thoughts are speculative. I suspect that even with the integration of a multi-modal large language model, this product may face challenges similar to those encountered by Amazon. I own an Echo smart speaker and mainly use it for basic tasks like setting alarms, reminders, playing music, and checking the weather—nothing transformative. This limited scope of use is one reason why Alexa hasn’t established a sustainable business model, incurring an annual loss of about $10 billion. It was only with the advent of ChatGPT that a mass-market product of this genre truly took off, rapidly becoming the fastest-growing consumer app ever. This raises an intriguing question: Will smart glasses follow the trajectory of Alexa or that of ChatGPT?

Custom GPTs

Speaking of ChatGPT… Do you use custom GPTs?

In November 2023, OpenAI introduced the ability to customize ChatGPT with specific instructions, additional knowledge, and various skills. These custom GPTs can assist in learning board game rules, teaching math to children, or designing stickers. Following this, OpenAI launched the GPT Store, making it accessible to ChatGPT Plus, Team, and Enterprise users. This store offers a selection of popular and helpful GPTs.

I haven't talked much about the store yet, but I did have some thoughts on GPTs themselves at their launch:

Did any of this happen?

OpenAI reports that users have created more than 3 million custom versions of ChatGPT. However, I haven’t come across any that have went viral, say, taking over Twitter in a single night. It seems that these customizations are primarily used for internal workflows—which is exactly how I use this feature myself. Let me show you.

I've developed three GPTs for my personal use: Summarize, Rewrite, and Density.

But they’re not apps, chatbots, or autonomous agents as I anticipated. They are shortcuts. That’s precisely how I created them for my use—I integrated them into my custom instructions:

Treat “/rewrite” as a shortcut for “Rewrite as a native speaker would:”

Treat “/summarize” as a shortcut for "Summarize the following article using bullet points. Keep in mind I have limited time and need a concise, intelligent overview.”

Now, I don’t even have to type the command; I can simply select a custom GPT from the sidebar or, if I'm already in a conversation with ChatGPT, summon any specific GPT using @, similar to mentioning someone in a group chat. This feature is cool and useful since custom instructions are capped at 1500 characters—yet this approach isn’t exactly revolutionary.

I've discussed ChatGPT with my friends who use it for various purposes—some for coding as technical users, and others for more casual tasks. None of them use custom GPTs, likely because they don’t deal with highly repetitive tasks often enough to feel the need—and see the benefit. For example, if you’re a programmer, you don’t really need a specialized GPT; chatting with the base model or using your text editor’s Copilot does the job well enough. (And if you’re a casual, you’ll use ChatGPT to help you draft emails or do homework for you, which the base model does great, too.)

This leads me to believe that custom GPTs may carve out a niche in the enterprise market. Picture a typical company where every team has highly repetitive workflows or tasks they’re looking to automate. These could be shared internally, making them accessible to all employees. Some of these GPTs might also function as knowledge bases. For example, the HR department could upload frequently asked questions about company policies to the platform. This seems like a practical application. While not groundbreaking, it’s a solid product that OpenAI could successfully offer to many companies.

However, regarding consumer-oriented apps, I’m not as convinced.

It appears that even ChatGPT struggles to match the success of its base version. Though the product remains highly useful, the platform doesn’t seem as appealing—not just to me, but likely to the broader audience as well.

Rewind Pendant

Another AI-powered device is Rewind Pendant, a wearable designed to capture and transcribe real-world conversations. The transcriptions are encrypted and stored solely on your phone, aligning with the company’s privacy-first ethos. They also provide features that aim to ensure that individuals are not recorded without their explicit consent.

Rewind is a macOS app designed as a searchable archive for your personal and professional life. It monitors your laptop, so if you’re in a meeting, the app can summarize it for you by listening and watching along with you. Importantly, it avoids becoming a privacy concern by performing most of the analysis locally on your device, without relying on cloud storage. When paired with the Pendant wearable, Rewind becomes a personalized AI that includes everything you’ve seen, heard, or spoken, even when you’re on the go. This represents another example of ambient computing, similar to Meta’s smart glasses, as these sensory inputs are eventually processed by a large language model.

The concept is divisive. Your stance will largely depend on your perspective on privacy; it could either be seen as invasive or beneficial. This paradox echoes the dilemma that Google Glass encountered years ago: while the device offered utility, its nerdy design deterred users wary of social judgment. In the case of the Rewind Pendant, the device looks sleeker, but its core function as a recording machine may make people hate you for wearing it so you face social scrutiny for choosing to use it.

This is currently available for pre-order, so the verdict remains pending until real-world use provides definitive answers. The Rewind app is already accessible, though, so I downloaded it and put it through a week-long trial.

If you look at the use-cases outlined on their website, Rewind appears to be most beneficial for individuals who frequently switch between contexts. For example, if you’re a manager juggling Zoom meetings, Notion comments, Linear issues, and emails throughout your workday, it’s easy to lose track of what was said or done where—making a universal, cross-app search engine valuable. If you’re a salesperson managing multiple daily meetings with different prospects, each requiring note-taking and follow-ups, Rewind can automate this process for you and make your life easier, too.

If you’re neither a salesperson nor a manager, as is my case, the utility diminishes. As an engineer, I minimize context-switching when possible and my work is concentrated in fewer platforms. Despite being technologically impressive, Rewind didn’t prove particularly useful for my specific needs. Drafting status updates for stand-up meetings doesn’t justify a $19-per-month price tag for me.

Based on my experience with the app, I can make some educated guesses about the potential utility of the Rewind Pendant, even though Rewind also promotes personal use-cases on their website. 

If you’re a salesperson who frequently engages in face-to-face meetings, the Pendant could be quite advantageous. Similarly, if you’re a manager coordinating multiple teams and need to keep track of numerous interactions, commitments, and insights from meetings, and you operate from a physical office, the Pendant might suit your needs. Attending a conference? The device will likely prove valuable, too. This could also mitigate the social acceptability issue that plagued Google Glass; using the Pendant in a professional setting might gain tacit approval—especially if all meeting participants see the benefit of automated note-taking.

Perhaps long-term use will change my mind. At the moment, its modest utility hardly seems worth the discomfort of constant surveillance.

Humane AI Pin and Rabbit R1

The Humane AI Pin is a wearable, voice-controlled device with a digital assistant designed to replace your smartphone for various common tasks. Priced at $700, it has received some of the worst reviews in recent memory.

The Rabbit R1 is a standalone pocket-sized device that allows you to interact with AI without the need for a smartphone. Designed like a walkie-talkie, it features a touchscreen, a 360° camera, and an analog scroll wheel. Its main draw is the rabbits, a series of automated scripts designed to handle your everyday mundane tasks. It, too, got eviscerated by reviewers.

The appeal of both device lied not in its hardware, but primarily in its cloud-based AI capabilities. For example, Rabbit utilizes what they term a Large Action Model, trained to understand graphical user interfaces, including the functionality of various buttons and the layout of website content. This means when a user requests something like “order pizza from my favorite restaurant,” the cloud application actually interacts with the Uber Eats interface—provided the user has logged in through OAuth on Rabbit’s website—and places the order.

Spoiler alert: users report that this feature often doesn't work as expected. While Meta's smart glasses may not be as ambitious, they perform their tasks reliably. In contrast, both of these devices promised too much and delivered too little.

Intriguing concepts aside, Rabbit and Humane some questions. Do they need to be standalone devices? The AI Pin was supposedly justified as an always-on device, constantly observing and listening to your environment. R1, however, operates through button inputs. The key here might be its operating system: if it’s designed to interact with various user interfaces directly, rather than through API calls, then this approach could be challenging to implement on iOS and Android due to the extensive app security permissions required. While the technical aspect is impressive, the success of such devices hinges on their practical applications. Previous AI assistants like Google Assistant, Cortana, and Alexa have shown that to persuade people to move away from their smartphones, these devices need exceptionally strong and compelling use cases.

Character AI

Which platform leads in monthly visits from both desktop and mobile in the realm of generative AI? Is it ChatGPT? Not quite. Bing? Think again. Perhaps Google’s Bard? Almost, but not there yet. The crown goes to Character AI.

Character AI lets users to design and converse with virtual characters. Teenagers are using this platform to craft personas of fictional entities, such as Aragorn from The Lord of the Rings, or renowned figures like LeBron James or Elon Musk. Even though ChatGPT might hold more brand recognition, Character AI takes the trophy for user engagement. Reports suggest that users dedicate an average of two hours daily on the platform.

While generative AI has no shortage of use cases or customer interest, it’s struggling to maintain user retention and daily engagement. In terms of one-month mobile app retention, AI-centric apps lag behind established companies. Even when it comes to daily active users as a percentage of monthly active users, generative AI apps have a median ratio of just 14%, well below the 60-65% seen in top consumer companies and WhatsApp’s 85%. The exception lies in the “AI Companionship” category, represented by apps like Character.

The real challenge for generative AI isn’t creating demand; it’s in proving sustained value to convert users into daily members—the impressive user retention of Character indicates that the AI Companionship category might possess one of the most compelling product-market fits within the generative AI industry. In a world captivated by new models, this often goes unnoticed. My personal opinion? Character, flying under the radar of tech nerds, is poised to become one of the biggest winners of this early AI era. It has the potential to be the TikTok of gen AI.