Wearable AI

Evaluating the practicality and potential of AI-integrated devices

????????? Picture yourself at a cocktail party or a business gathering, encountering someone you know but whose name eludes you. A voice, audible only to you, softly reminds you of their name and your most recent encounter.

I've tried to replicate the feature article and most of the content from my email newsletter in the LinkedIn Newsletter, but if you want the additional bonus content, including the Prompt of the Week, my AI Tools recommendations, and other bonus content then I suggest you subscribe via email to The Artificially Intelligent Enterprise and get this newsletter in your inbox every Friday.

Thank you for subscribing; it means a lot to me. Enjoy this week’s edition.

This vision represents the AI future I want, but how distant is this reality? Read on and find out how wearables may change the way we live.

Sentiment Analysis

Human-in-the-Loop AI Workflows Before Agentic AI Workflows

Lately, I’ve been thinking about the reality of where we are with AI and the promise of where we are going. I liken our progress in AI to the word-processing days of computing.

Imagine the day we got rid of whiteout and were able to make mass searches and replace misspellings on a CRT screen (not even an LED!). That’s where we are.

A few decades later, we sent hundreds of properly formatted emails a week, shared information on websites and social media networks, and did most of our shopping and research online. It was an exponential leap. It took a while, but AI is moving even faster.

Today, the promise is that we type a simple phrase into a chatbot like ChatGPT and get something perfect, from a business plan to a blog post. It’s not generated content, but it’s not just generated text, it has accomplished a sophisticated objective.

That’s not reality...yet.

Where we are in the age of human-assisted AI generation or as I call it Human-in-the-Loop AI.

Where we are going is the world of agents where our own Jarvis (Ironman reference) or Rosy the Robot (the Jetsons reference for those of us with more grey hair). But for now, I think we are at the stage where we need to use the tools that give us the most significant advantage as part of a human-in-the-loop toolchain.

I riffed on this during the week on LinkedIn and created a video about how I use AI to create articles. The video's prompt is also this week’s prompt. Here’s my example of a workflow.

https://youtu.be/xEnOOP2vrp4

AI TL;DR

Generative AI News and Tips

• Alibaba's large language model, Qwen - The model from Chinese company Alibaba is currently ranked as the number one open source model on Hugging Face's LLM Leaderboard, outperforming competitors such as Meta's Llama 3 70B and Mistral AI's Mixtral 8x22B.
• Stable Diffusion 3's Disastrous Launch - Stable Diffusion 3 was a big step back over previous models. It specifically struggles to generate correct human anatomy, a shortcoming some users think is due to Stability AI filtering adult content out of its training data for this model.
• Amazon hires founders away from AI startup Adept - Adept, a startup developing AI-powered “agents” to complete various software-based tasks, has agreed to license its tech to Amazon and the startup’s co-founders and portions of its team have joined the e-commerce giant.
• How responsible “agentic AI” can supercharge your startup -Although this article was aimed at startups, I think the takeaways work for any business.

Feature Story

AI Wearables: Revolution or Hype?

Imagine glancing at your watch for a real-time conversation translation or having your glasses whisper a reminder of an upcoming appointment and provide an overview of the details you need to remember. These scenarios, once the stuff of science fiction, are inching closer to reality with the advent of AI wearables. But as tech companies rush to strap artificial intelligence to our bodies, a crucial question emerges:

Are these devices the next giant leap in personal computing, or just another overhyped gadget destined for the junk drawer?

From the ambitious Humane AI Pin to the subtly tech-enhanced Ray-Ban Meta Smart Glasses, the market is suddenly awash with devices promising to bring AI out of our phones and onto our person. These wearables aim to seamlessly integrate artificial intelligence into our daily routines, offering everything from hands-free digital assistance to augmented reality experiences. But do they truly deliver on this promise, and more importantly, do we need them?

The Current AI Wearable Landscape

AI wearables are rapidly evolving, with new devices hitting the market dizzyingly. Here's a snapshot of some key players:

• Humane AI Pin: This clip-on device is a personal AI assistant featuring a camera, projector, and various sensors. It's designed to provide hands-free access to AI capabilities and potentially replace your smartphone.
• Ray-Ban Meta Smart Glasses: These stylish specs incorporate AI features like object recognition and real-time information access through voice commands. They blend fashion with function, addressing one of the key criticisms of earlier smart glasses.
• Rewind Pendant: This neck-worn device records conversations and uses AI to extract insights, raising intriguing possibilities and privacy concerns.
• Samsung Galaxy Ring and Apple Watch Series 10: While not yet released, these devices from tech giants Samsung and Apple are expected to incorporate advanced AI features for health monitoring and personal assistance.

Use Cases: Promise vs. Reality

The potential applications for AI wearables are exciting:

• Real-time language translation
• Instant information retrieval and fact-checking
• Personalized health monitoring and recommendations
• Enhanced productivity through voice-activated task management
• Augmented reality experiences for both work and leisure

However, the reality often falls short of the hype. Many current AI wearables rely heavily on a smartphone connection, limiting their standalone capabilities. Battery life remains a significant hurdle, with many devices struggling to last a full day of active use.

Moreover, the practical benefits of having AI constantly at your fingertips (or on your face) are still being debated. Do we need instant access to information at all times, or does this constant connectivity risk further fragmenting our attention and eroding our ability to be present in the moment?

Learning from Past Failures: The Google Glass Cautionary Tale

No discussion of AI wearables would be complete without mentioning Google Glass, the high-profile failure that serves as a cautionary tale for the industry. Launched in 2013 with much fanfare, Google Glass promised to revolutionize how we interact with technology. Instead, it became a case study of how not to introduce a new tech product category.

Google Glass failed for several reasons:

1. Privacy concerns: The device's prominent camera raised fears about unauthorized recording.
2. Limited functionality: Early versions offered few practical applications to justify their high cost.
3. Social stigma: Wearers were often viewed negatively in public, earning the unflattering nickname "Glassholes."
4. Poor battery life: The device struggled to last a full day.
5. High price point: At $1,500, it was prohibitively expensive for most consumers.

The Google Glass experience offers valuable lessons for current AI wearable developers, emphasizing the need for clear use cases, social acceptability, and robust functionality.

The Current Leader: Meta Ray-Ban Smart Glasses

Among the current crop of AI wearables, the Meta Ray-Ban Smart Glasses stand out as one of the most promising and capable devices. Here's why:

1. Familiar form factor: By partnering with Ray-Ban, Meta created a device that looks like regular sunglasses, avoiding the social stigma Google Glass faces.
2. Practical features: The glasses offer helpful features like hands-free photo and video capture, audio playback, and AI-powered object recognition.
3. Integration with existing ecosystems: They work seamlessly with Meta's social platforms and AI assistants.
4. Reasonable battery life: The glasses can last several hours of active use, making them practical for daily wear.
5. Competitive pricing: While not cheap, they're priced similarly to premium sunglasses, making them more accessible than many other AI wearables.

Challenges and Limitations

Despite their promise, AI wearables face several significant challenges. These all apply to the new AI-enabled RayBans:

1. Privacy and security: The always-on nature of many AI wearables raises serious questions about data collection, storage, and potential misuse.
2. Social etiquette: As with smartphones, society still decides when and where to use AI wearables.
3. Dependency concerns: There are valid worries about becoming overly reliant on AI for tasks we once performed ourselves, potentially atrophying specific cognitive skills.
4. Technical limitations: Many AI wearables still struggle with issues like battery life, processing power, and connectivity.
5. Regulatory hurdles: As AI wearables become more prevalent, they will likely face increased scrutiny and regulation, particularly around privacy and data protection.

The Rabbit R1, Smoke and Mirrors

I was fascinated when The Rabbit R1 launched at the Consumer Electronics Show (CES) earlier this year. I instantly ordered one that was promised by April. It’s well past April, and it just got here. Also, reviews have panned it repeatedly as all hype and no substance. It’s sitting unopened on my desk; my excitement has dropped since the release and seeing the reviews.

The Rabbit R1 was supposed to be a virtual assistant in a box. The battery has been reviewed repeatedly, and the battery life is supposedly bad. It can’t set timers or alarms. It lacks haptic feedback like an iPhone. Also, there’s no back button, and the straightforward interface lacks functionality. Even though it has a touchscreen, it’s not consistently used throughout the device.

What was supposed to set Rabbit apart was that it included a large action model (LAM), which allowed you to control your applications with voice commands. However, there are a minimal number of apps you can use on this device:

• Music (Spotify and Apple Music)
• Generative AI (Midjourney and Suno)
• Rideshare (Uber)
• Food (DoorDash)
• Up-to-date search with Perplexity
• Search with Wolfram Alpha
• Looking up reviews on Yelp

You could already do these things on your smartphone with a better interface and battery.

The Humane AI Pin, Unique UI but Error Prone

The Humane AI Pin has faced many criticisms from reviewers and tech enthusiasts. Users must tap on their chest to activate it, which can be awkward, especially for left-handed users who might inadvertently touch their nipples (according to posts on Reddit).

It relies on a laser projector that displays information on the user's hand, which is difficult to read in bright sunlight. The required tilt and pinch gestures are challenging to master and frustrating.

The AI pin is often slow to respond to commands and can be inconsistent, making it unreliable for regular use. The device tends to overheat, causing discomfort and necessitating frequent cooling periods. It has a short battery life and requires multiple daily charges, exacerbated by the need for additional battery boosters.

The AI Pin does not sync with existing smartphones and lacks integration with popular apps like WhatsApp and iMessage. Users must pay a $24.99 monthly subscription fee in addition to the $700 purchase price, which many find unjustifiable given the device's limited functionality. Despite assurances from Humane about data privacy, the always-on nature of the device raises concerns about constant surveillance and data security.

In addition, the famed YouTube product reviewer Maqurest Brownlee said the Humane Pin was the worst product he’s ever reviewed.

https://youtu.be/TitZV6k8zfA?si=NQ6oAMMpVJYtGH08

The Future of AI Wearables: Standalone or Smartphone Companions?

A key question for the future of AI wearables is whether they'll evolve into truly standalone devices or remain tethered to our smartphones. Most AI wearables rely on smartphone connectivity for advanced processing and full functionality. This limits their appeal as genuine smartphone replacements.

However, as processing power improves and AI models become more efficient, we may shift towards more capable standalone devices. The ultimate goal for many in the industry is to create an AI wearable that can fully replace the smartphone, offering all its capabilities in a more accessible, always-on form factor.

Scenario: AI Wearable Enhancing Workplace Productivity

Imagine a project manager, Sarah, overseeing multiple teams working on a critical product launch. She uses a sophisticated AI wearable that integrates with her company's systems and significantly enhances her daily productivity. The AI wearable provides a personalized briefing summarizing the day's schedule, essential tasks, and recent updates. It monitors her health metrics and offers tailored recommendations for optimal health and productivity.

During the day, the AI wearable dynamically manages her task list, prioritizing activities based on urgency and deadlines. It adjusts her schedule in real-time as new tasks or changes arise. During meetings, the AI wearable transcribes conversations, highlights key points, and integrates action items directly into her project management software. It also provides real-time data insights relevant to the discussions. The wearable enhances communication by filtering and summarizing emails and messages, highlighting the most critical ones, and allowing Sarah to respond through voice commands efficiently.

When Sarah encounters a problem, the AI wearable can pull up relevant documentation, past project data, and expert recommendations. It can also connect her with colleagues or experts for immediate advice via augmented reality (AR) overlays. Using predictive analytics, the AI wearable helps Sarah foresee potential project risks and suggests mitigation strategies, improving decision-making accuracy.

To help Sarah maintain focus, the wearable monitors her work patterns and suggests breaks, mindfulness exercises, or physical activities based on her productivity and energy levels. It tracks time spent on various tasks, providing insights into how Sarah spends her time and suggesting ways to optimize her workday. The wearable summarizes accomplishments, pending tasks, and a preview of the next day's priorities, ensuring Sarah stays on top of her responsibilities.

Below is a demonstration of Google Astra that starts to tease out this reality, though it’s in its very early stages. It’s an open source platform for AI agents that would likely complement a new category of wearable AI devices.

https://www.youtube.com/watch?v=nXVvvRhiGjI

A Wearable Alexa? Not yet.

AI wearables undoubtedly have the potential to change how we interact with technology and the world around us. The ability to access information and AI assistance seamlessly, without reaching for a phone, could genuinely enhance productivity and quality of life for many users.

However, we're still in the early stages of this technology. Current devices, while impressive, are primarily smartphone companions rather than replacements.

The Meta Ray-Ban Smart Glasses represent a step in the right direction, balancing functionality with social acceptability, but they are far from being a must-have device for most consumers.

As the technology evolves, we're likely to see AI wearables become more capable, more discrete, and more integrated into our daily lives. But this evolution also brings challenges that extend beyond mere technical hurdles. As a society, we'll need to grapple with questions of privacy, social norms, and the appropriate role of AI in our lives.

For now, AI wearables remain a promising but unproven technology. They offer a glimpse into a potential future where our digital assistants are always with us, ready to help at a moment's notice. But whether that future is one we truly want or need is a question we're still in the process of answering.