Build logs are open to everyone. Feel free to share them, use the info, or repurpose the images for your own $SHOW related content
This week was a mix of fixes, experiments, hardware overhauls, and livestream setup testing. The laser cutter went down early in the week, forcing a full tear-down, coolant system check, and sensor troubleshooting. It’s now fully cleaned and back online, ready for merch and precision cutting. Vita’s head got some tweaks, the LED mouth enclosure was flipped for a cleaner look, a mouth light diffuser was added, and we implemented sound wave simulation on the mouth light-strip so it can animate speech in sync with VitaNova’s voice. We ran through multiple TPU “neck condom” prototypes, but print and folding issues led us to shelve it in favor of the blue U-joint for now.
On the electronics side, the I²C bus got tidied up with a new daisy-chain harness and confirmed addresses for the IMU, ambient light sensor, and dual PCA9685s. The plan is to mount the IMU in the torso to keep head movement from polluting motion data. We also identified Pi brownouts linked to the mouth’s power draw, so a bigger refactor is in progress, moving more devices off the Pi rail, and adding a common ground bridge between systems. New tools, including a logic analyzer and oscilloscope, arrived to help diagnose bus noise, track voltage fluctuations, and generally bring a more sophisticated approach to troubleshooting these types of issues. On the compute side, we’re also investing heavily in an upgrade from 16GB to 96GB VRAM with a new Blackwell GPU, letting us run vision models, LLMs, and speech-to-text at full precision and throughput, with plenty of headroom to stack in other advanced models down the line.
Arm prototyping continued, focusing on a shoulder-driven first version with aesthetic bolts and space for future elbow/wrist actuation. Internal debates on proportions and detailing are shaping the next iteration. In the background, the team’s also been getting ready for more outdoor testing.

Vita with the new NeoPixel LED mouth, driven by ESP32 using processed audio waveform data to animate in sync with speech

Servo and sensor wiring routed through flexible spiral wrap for strain relief and cleaner cable management

Vita after a rough night at the workshop 😂




Unified pipeline with live visual context