Breaking the Humanoid Robot Delusion (www.computerworld.com)

Silicon Valley startup 1X this week unveiled NEO, a consumer-focused humanoid robot billed to automate housework and act as an in-home assistant. The company published a glossy demo of NEO folding laundry, cleaning, fetching items, holding conversations, and even dancing, and opened pre-orders ($20,000 early access or $499/month), with U.S. shipping slated for 2026 and broader rollout in 2027. NEO’s hardware specs include a tendon‑driven skeleton with high‑torque motors, a soft 3D‑lattice polymer body, 22‑DOF hands, a 66‑lb mass, 154‑lb lifting capability (55‑lb carrying), and quiet operation (~22 dB). Its software claims include natural conversational interaction, visual recognition for objects/ingredients, memory of preferences, and ongoing learning via updates and remote expert sessions; it supports Wi‑Fi, Bluetooth and 5G. The announcement is significant less for a technical leap than for what it reveals about industry hype: independent reporting shows the demo was largely teleoperated, not an example of robust autonomy. That mirrors prior “showroom” events (e.g., Tesla Optimus) and highlights core research gaps—generalizable perception, planning and dexterous manipulation across home variability require vast task‑specific training and remain unsolved. For researchers and buyers this is a cautionary moment: teleoperation and data collection are legitimate development steps, but marketing should not conflate puppeteered demos with deployable autonomy. The episode raises questions about transparency, consumer risk, and realistic timelines for bringing true household humanoid robots into everyday homes.