Fable Engineering #7: The Pivot We Didn't See Coming
From home robots to cognitive augmentation, and why it makes sense.
Hey friends,
Six months ago, we were all-in on consumer robotics. We had 250K+ social media reactions, 150+ early adopters on a waitlist, and a clear conviction that physical intelligence belonged in every home.
We explored everything. Companion robots for kids. DIY robotics platforms. Elderly care robots. Premium vacuum cleaners with personality. We talked to hundreds of people. We surveyed 1,500 parents. We studied the competition obsessively.
And nothing came close.
Now we’re building something completely different. This post is about what happened, what we learned, and why this pivot isn’t an abandonment. It’s a liberation.
The Vision Was Never Robots
Here’s what we need to say upfront: the vision was never really about robots.
From the beginning, what excited us was the idea of a memory OS. A system that remembers everything. That compounds over time. That unlocks personalized experiences no generic AI could ever provide. The more it knows you, the more useful it becomes. The more useful it becomes, the more irreplaceable it feels.
Think about what this actually means. Today’s AI tools are stateless. Every conversation starts from zero. You explain your context, your preferences, your history, over and over again. But a memory OS would be different. It would know that you’ve been thinking about a career change for three months. It would remember the name of that person you met at the conference. It would notice that your energy drops every Wednesday afternoon. It would connect dots you didn’t even know existed.
We called this the Third Brain. Your first brain is biological: creative, intuitive, but limited in memory and easily overwhelmed. Your second brain is your smartphone: infinite information, always connected, but constantly demanding attention. The Third Brain would be different. It observes without interrupting. It remembers without you having to. It acts on your behalf because it actually understands the context of your life.
Companion robots seemed like the natural path to get there. A robot that follows you around your home, always present, always capturing context. It could see what you see, hear what you hear, and build a rich understanding of your daily life. It would become an extension of your personality. Starting in the home made sense. That’s where people spend most of their time. That’s where the most personal moments happen.
So we went deep on home robotics.
The Graveyard We Kept Walking Past
But every direction we explored led us back to the same uncomfortable truth: consumer robotics has a body count.
Anki raised $182M from top-tier VCs. Sold 6.5 million robots. Cozmo was a holiday bestseller. They still went bankrupt in 2019, burning $50-60M per year against $100M in revenue. The market was smaller than everyone expected, and there was no recurring revenue to sustain the business.
Amazon Astro had a $1,000+ price point, Amazon’s full resources, and their entire distribution machine. Launched 2021, discontinued 2024. Even Amazon couldn’t crack it. The core problem: users didn’t know what job they were hiring it for.
Jibo. Vector. Kuri. Companion-first robots consistently fail. Cuteness isn’t enough. Single-purpose robots become closet decorations within weeks. Combined, these companies raised hundreds of millions and all ended the same way.
We kept finding ourselves in an impossible positioning problem:
Pure utility leads to commodity margins, competing against iRobot and Roborock’s scale
Companionship-first leads to the graveyard
Premium design + utility was already being executed by well-funded players
Every angle we tried, someone had either failed at it spectacularly or was doing it with 100x our capital.
We Got Locked in the Solution Space
Here’s the mistake we made: we fell in love with the solution instead of the problem.
The problem we cared about was context engineering. How do you build a system that truly understands someone? That captures the raw material of their life, their conversations, their ideas, their patterns? That compounds this understanding over time into something genuinely useful?
But instead of staying focused on that problem, we got pulled into the weeds of robotics. Navigation algorithms. Motor controllers. Battery life. Manufacturing costs. Obstacle avoidance. SLAM mapping. We spent months thinking about how to make a robot move gracefully through a living room, when the real question was: how do you capture and structure the context of someone’s life?
Robots were one possible vehicle for context capture. But they came with enormous baggage. Hardware complexity. Capital intensity. Long development cycles. Regulatory concerns around cameras in homes. The need to raise tens of millions before shipping anything.
And fundamentally, a robot that stays in your house only captures a fraction of your life. The most important conversations happen outside the home. The best ideas come during commutes, walks, meetings. You might spend 8 hours at home, but the richest context often lives in the other 16.
We were building an elaborate delivery mechanism for something that needed a much simpler form factor.
Why Wearables Make More Sense
Once we stepped back from the solution space, the answer became obvious.
If the goal is a memory OS that captures context and compounds over time, you need something that’s always with you. Not something that stays in your house. Not something that follows you from room to room but can’t leave the front door.
A wearable.
It goes where you go. It captures the conversations, ideas, and moments that matter, wherever they happen. The coffee meeting. The phone call on your commute. The idea that strikes you during a walk. A home robot misses all of this. A wearable catches everything.
It’s also simpler hardware, which means faster iteration and lower capital requirements. No motors. No navigation. No obstacle avoidance. Just audio capture, on-device processing, and a beautiful form factor that earns its place in your daily life.
And critically, it solves the context engineering problem directly, without the overhead of mobility and physical presence. The Third Brain doesn’t need legs. It needs ears.
We’re now building a wearable AI device designed for cognitive augmentation. Intelligence that learns your patterns, remembers your context, and acts on your behalf. The memory OS we always wanted to build, in the form factor that actually makes sense.
The Opportunity Is Massive
This isn’t just a better path to our original vision. It’s a better market.
Cognitive effects are the new network effects. For two decades, the strongest moats in tech were built on network effects. Facebook won because your friends were there. Amazon won because of logistics scale. But a new kind of moat is emerging. One that doesn’t require millions of users. It requires deep understanding of one user at a time.
The product that knows you best becomes irreplaceable. Not because switching is impossible, but because switching feels like losing part of yourself. Oura keeps 95% of its users. Spotify Premium retains 85%. Not because of network lock-in, but because leaving means abandoning a system uniquely shaped to you.
Our device builds this kind of moat from day one. Every conversation captured, every context learned, every pattern recognized. The longer you use it, the more valuable it becomes. And that value is yours alone. No one else can replicate the understanding we build with you.
The timing is right. AI is finally capable enough to do something useful with continuous context. A few years ago, you could capture everything but couldn’t make sense of it. Now, the intelligence layer can actually understand, structure, and act on ambient information. Large context windows. Better reasoning. Real-time processing. The technology caught up to the vision.
The form factor is ready. Wearables have crossed the chasm. AirPods normalized always-on audio devices. Oura and Whoop normalized health wearables. The market is primed for a wearable that augments cognition, not just tracks biometrics.
The business model works. Subscription-friendly retention. Smaller capital requirements than robotics. Faster validation cycles. Weeks, not years. And wellness/augmentation positioning commands higher multiples than productivity tools.
What We’re Bringing Forward
The robotics exploration wasn’t wasted. Everything we built transfers to this new direction.
Our partnership with Futurewave, the Belgian design studio behind products for Renault, Sony, and Nokia. They’re now designing the wearable with us.
Our privacy-first architecture. On-device processing was always core to our philosophy. It applies even more strongly to something you wear all day.
Our obsession with design that earns its place in your life. Technology that belongs rather than demands. That principle is form-factor agnostic.
And our deep understanding of what doesn’t work. Six months of exploring the robot graveyard gave us pattern recognition for failure modes. We know what to avoid.
The Honest Reflection
For other founders navigating similar crossroads:
On solution vs. problem. It’s easy to fall in love with your solution. Robots are cool. Hardware is tangible. You can hold a prototype and feel like you’re making progress. But the solution should serve the problem, not the other way around. We spent months optimizing for the wrong layer. The moment we stepped back and asked “what problem are we actually solving?” the path forward became clear. If you’re deep in execution and feeling stuck, zoom out. Make sure you’re still pointed at the right destination.
On sunk costs. It’s painful to walk away from work you’ve invested in. The prototypes, the research, the partnerships, the identity you’ve built around a particular direction. But the worst outcome would have been committing to a vehicle that couldn’t reach the destination. Pivoting isn’t failing. It’s refusing to let past effort dictate future direction. Everything we learned about context capture, privacy architecture, and designing technology that earns its place in someone’s life still applies.
On timing. Consumer robotics may have its moment eventually. But the memory OS, the context engineering problem, the Third Brain? That moment is now. AI is ready. The form factors exist. The market is primed. Better to ride the wave that’s here than wait for one that might never come.
On knowing when to walk away. We could have kept exploring consumer robotics for another year. There’s always one more concept to try, one more user segment to test, one more angle that might work. But the evidence was overwhelming. Every path led to the same structural problems. Sometimes the bravest thing is admitting you’ve learned enough to change direction.
What’s Next
We’re completing hardware development with Futurewave, the Belgian design studio that’s been our partner since the robotics days. Building the memory OS layer that captures, structures, and compounds context over time. Designing the intelligence that turns raw conversation into actionable understanding.
The first version will focus on professionals who live in meetings and conversations. People who lose ideas between calls. Who wish they could recall what was said three weeks ago. Who want an AI that actually knows their context instead of starting from zero every time.
If that sounds like you, we want to hear from you. We’re looking for early adopters who believe the future of AI isn’t just smarter models, but smarter context. People who want technology that amplifies their memory and frees their attention, rather than fragmenting both.
The vision hasn’t changed. A system that remembers everything. That compounds over time. That unlocks experiences no generic AI could provide.
We just found the right vehicle.
We’re not building robots anymore. We’re building the Third Brain.
Until next time,
Pierre-Louis, CEO
PS: We will be in Las Vegas from the 5th to 10th of January for CES. Reach out to meet and test our first demo!