Turns out, Talking Humanoid Robots Are Still Kind of Creepy…

Talking Humanoid - Lily

Throwback to Lily v0.1—our humanoid robot that could introduce itself, answer questions, and strut confidently through the halls.

In this video (from a year ago), Lily is smooth. Responsive. On it. Most of the time.

But here’s the thing I didn’t anticipate:

As Lily roamed the hallways of my 40-story building in downtown Miami (yes, we actually launched in Miami! 😊), I quickly realized that a humanoid robot saying “Watch out, coming through” or asking people for directions wasn’t exactly charming. It was… unsettling. People froze. Some bolted. A few looked genuinely concerned for their safety.

So yeah—I turned it off.

But that wasn’t the only oops moment.

A lot of Lily’s time was spent searching for stuff—apartment numbers, trays left in hallways, light bulbs that needed changing, debris by the pool. And honestly, having a humanoid slowly scan things like a confused tourist wasn’t the best use of time.

That’s when it clicked: humanoids (and robots in general) need better spatial awareness. Instead of each robot awkwardly mapping out an entire space on its own, why not build large-scale spatial models that they can just plug into?

So we pivoted. Now, we’re focused on building real-time spatial intelligence for complex spaces—so robots don’t have to figure it all out from scratch.

Lesson learned: humanoid robots with voices freak people out, but giving them large scale spatial awareness instead of just legs? That’s the real unlock.

(Watch the video and let me know what you think!)

Also, keep your eyes peeled on www.augustmille.ai for some really exciting news coming up!!!