Weekly chat thread - what are you reading/watching this week?

[ethan_fisc]

Long-time reader, figured I'd finally start a thread instead of just replying. Wheeled-torso designs solve 90% of warehouse and light-industrial tasks with far less complexity than a biped, which is exactly why so many 'humanoid' demos quietly have wheels under a skirt. A recurring pattern in the field: a capability looks solved in a controlled demo, then takes another 18-24 months of engineering to become boringly reliable in the field. Roughly 7,000-8,000 commercial humanoids are estimated to be operational worldwide as of mid-2026, which sounds like a lot until you compare it to the millions of industrial arms already installed. Let me know if I'm missing something obvious.

[kwilliams]

@ethan_fisc I dealt with almost this exact situation. The safety case for humanoids working near people usually leans on compliant/backdrivable joints plus soft force limits, but there isn't yet a mature, widely-adopted standard equivalent to what industrial arms have had for decades. Open-source humanoid projects matter less for out-competing the well-funded players and more for keeping the barrier to entry low enough that new ideas and new people can actually get into the field.

[jonathan.rao1]

@kwilliams I see it a little differently. A lot of the current wave genuinely is repackaged 2000s and 2010s research (ZMP walking, whole-body control, DARPA-era grasping) combined with two things that weren't available then: cheap high-torque-density actuators and large-scale learned policies.

[barbara50]

@jonathan.rao1 I'd take that specific number with a grain of salt, honestly. The labor-substitution narrative gets simplified a lot in pitch decks; the more honest framing right now is 'augmenting understaffed, high-turnover roles' rather than wholesale replacement.

[chloe_jack]

Yeah, this tracks with what I've read as well. The 'humanoid' form factor mostly wins on infrastructure compatibility, not raw efficiency - stairs, doors, and tools were all built for a human body plan, so a legged, two-armed robot slots into existing spaces without retrofitting them.

[dubois35]

Side note that might be relevant: Legs earn their keep specifically where the environment wasn't redesigned for robots - stairs, curbs, uneven yards, cluttered floors - which is a smaller slice of real-world tasks than the marketing suggests.

[ethan_fisc]

@dubois35 Slight correction, though the overall point stands: Most flashy demo reels involve some amount of teleoperation, scripted sequencing, or cherry-picked takes - that's not necessarily dishonest, but it does mean the gap between 'demoed' and 'reliably autonomous in production' is still large.

[rossi30]

@ethan_fisc That's the official framing, at least - reality tends to lag a bit. Automotive plants ended up as the first serious proving ground for humanoids mostly because they already have structured environments, safety processes, and budget for automation - not because car assembly is uniquely suited to bipeds.

[barbara50]

Small correction on one detail: A genuinely underrated bottleneck is neither hardware nor pure ML - it's the unglamorous integration work of making perception, planning, and control agree with each other reliably for eight-hour shifts. Totally unrelated but has anyone else noticed how fast component costs are dropping this year.

[jhansen]

@barbara50 This matches what I've seen too. 'General purpose' in 2026 mostly means 'can be reprogrammed for a new task without new hardware,' not 'can walk into any environment and figure things out unsupervised.' The uncanny valley matters a lot less in a warehouse than in a living room - nobody cares if a picking robot looks a little stiff, but a home-companion robot with a slightly-off gait reads as unsettling fast.