[HN Gopher] The Robotic Dexterity Deadlock
___________________________________________________________________
The Robotic Dexterity Deadlock
Author : shmublu
Score : 65 points
Date : 2026-02-27 19:51 UTC (3 hours ago)
(HTM) web link (www.origami-robotics.com)
(TXT) w3m dump (www.origami-robotics.com)
| amelius wrote:
| What is the power draw of these joints?
| Zigurd wrote:
| Surgical robots, and robot pianos both exist. Neither employ
| humanoid hands. This all just illustrates how humanoid robots
| are, in multiple dimensions, going down technology rat holes. In
| some cases better solutions already exist without looking
| humanoid. In other cases, the humanoid form factor fails to
| address problems like a high center of gravity in a device that
| needs to not fall on grandma while helping her around the house.
|
| I continue to be amazed that the wrong form factor keeps being
| pursued. Though I suppose I shouldn't be too surprised given the
| parade of failed "AI devices."
| haarts wrote:
| You are right. If the hand is doing a specific task, better
| morphologies are likely. But that's not always desirable. The
| canonical example is of course the household. I don't want X
| robots, I want 1. And I don't want to change anything. Robot
| hand!
| fatbird wrote:
| Not to mention that the world is very widely designed to be
| manipulated by hands: doorknobs, handles, container sizes. A
| unique door opening appendage isn't going to do much good
| around your house.
| leonflexo wrote:
| Similar to how we are seeing LLMs shoved into spaces where
| existing ML was already doing well and better suited.
|
| Not to dismiss the value of LLMs in those cases as an
| interface/interpretation layer.
|
| If grandma goes into the windowless surgery factory, I just
| want the best bots working on her. There is value in having Dr.
| Bot the replicant give me the face-to-face status updates. We
| are not breaking out those layers as much, anymore, as the
| focus becomes minimizing FOMO.
| program_whiz wrote:
| I think one major draw to human-like for factors is the reuse
| of existing ecosystems and tools. If you have human-like
| grasping, you can reuse tools and utensils for human hands,
| otherwise, you need custom attachments. If you have human-like
| legs you can navigate stairs, wear pants for customization, and
| possibly operate a car or bike.
|
| Its a bit like choosing JS / python -- of course performance is
| inferior to a compiled language with highly tailored code, but
| they are flexible and have an ecosystem that might do 99% of
| the lifting for you.
|
| But in isolation, I agree with your idea that specialized
| robots with form fitted specifically to task will likely
| outperform a more generalized solution in a specific domain of
| behavior, the more generalized will likely outperform in
| flexibility and reusability (e.g. capable of reusing the human
| ecosystem).
| nkrisc wrote:
| I think it's less about tools and more about the spaces that
| humans operate in.
|
| You don't need a human-like hand to hold a tool made for
| humans. As an extreme example, you can make a robot operate a
| power drill with strap to hold it and a servo with a small
| bit of wood to operate the trigger mechanism.
|
| But for a robot operating in a space made for humans there
| certainly are some physical requirements which are based on
| the human form: maximum volume and clearances, stairs,
| fragile fixtures that can't be operated with too much force,
| etc.
|
| Ever walk through some over-crowded antique shop where you
| need to twist and lean your body to avoid knocking into
| thing?
| amluto wrote:
| There are a whole lot of tools intended for human use that I
| would use much more effectively if I could rotate my wrist
| repeatedly in the same direction.
| mikepurvis wrote:
| But that's a superset of human functionality, aka even
| worse.
| numpad0 wrote:
| Reuse tools to do what? Touch type on a mechanical keyboard?
|
| What makes human hands especially suitable for e.g.
| assembling a phone or installing a door handle onto a car?
| tryauuum wrote:
| > Touch type on a mechanical keyboard
|
| yes. do you think it's safe to just plug usb into some hole
| and type? the safest option for a robot is typing with
| fingers
| makeramen wrote:
| I see it as trying to apply the bitter lesson to robotics.
| Specialized robots will always have their place, but humanoid
| ones can take advantage of all the design interfaces that
| already exist in the world for humans.
|
| Similar to how claude code gained so much traction in terminal
| by just leveraging the command line interface that already
| exists for humans, no need to invent a domain specific MCP to
| just run shell commands.
|
| I agree with you that it's far from the most efficient approach
| for specific tasks. But the analogy would be that you also
| generally don't want to use LLMs to do something you can "just"
| write a script for... that doesn't make LLMs useless though.
| beau_g wrote:
| Many overactuated, purpose built robots (like surgical robots
| and pianos) exist, and have existed since the Unimate, and work
| great in certain situations. The problem with all of them is
| they are very expensive, often extremely large, and single
| purpose or very narrow purpose (and even if they are narrowly
| multipurpose, require tons of setup to get to work for each job
| they are intended to do).
|
| I personally am not bullish on 1:1 human hands either, but IMO
| the question shouldn't be $100k 2 ton Kuka arm vs biped with
| hands, it's overactuated robotics (build it from the floor with
| hard coded operations) vs underactuated (build it from the
| contact point of the work backwards with ML and sensors). We
| shall see which form factors prevail, but the type of robotics
| development posted here seems like the way forwards regardless,
| an ecosystem of small, power dense, reliable, accurate QDD
| actuators will lead to many general purpose robot applications.
| I recognize I am not using underactuated vs overactuated in
| their strict definition here but if you are familiar with
| robots I think you'll understand where I am coming from as far
| as a robot design ethos.
|
| I will say though in designing robots of this type without
| necessarily being bound by trying to make a robot look like a
| human, I have often found myself accidentally recreating human
| arm DOF in a round trip way, it does just end up being well
| packaged beyond the "world designed for humans" talking point.
| Maybe hands will end up being a similar situation.
| amluto wrote:
| A humanoid human will fall over too if pushed into a
| sufficiently awkward corner. It's a fundamental problem with
| things that aren't statically stable and need active
| stabilization.
| ortusdux wrote:
| Website down? "This deployment is temporarily paused"
| re wrote:
| Website is offline, archived snapshot:
| https://web.archive.org/web/20260227201321/https://www.origa...
| arjie wrote:
| I couldn't read it except on archive.org. Here's the link to save
| people trouble
| https://web.archive.org/web/20260227201321/https://www.origa...
| eyeino wrote:
| Nothing scarier than a Vercel bill.
| analognoise wrote:
| It's not X. It's Y.
|
| Multiple times, over and over.
|
| We need to stop with the AI stuff.
| roywiggins wrote:
| Even the title follows a common LLM pattern (The X
| Problem/Issue/etc).
| kristjansson wrote:
| While it's got some clear LLM patterns, the content seems novel
| enough to be worth the squeeze. That or I'm far enough outside
| of my Gell-Mann amnesia bubble that I can't see the slop
| airstrike wrote:
| Plus a bunch of other elements that are dead giveaways.
|
| In this day and age, I wish people would ask any model OTHER
| than ChatGPT to rewrite their shit. At least we'd get a
| different flavor of slop.
| modeless wrote:
| I think they did a search-and-replace to turn em dashes into
| semicolons in an attempt to hide the AI. Weird usage of
| semicolons.
| k1ko wrote:
| Once I became aware of this AI slop pattern, I can't stop
| seeing it everywhere.
| furyofantares wrote:
| Tip for those who want to skip shit like this, excessive
| headings glued together by bullet points is quicker to spot,
| especially since the headings almost always start with "The".
|
| I now scroll any AI-adjacent article I see and just read
| headings and if I see this I know what I'm getting into:
|
| The Dexterity Deadlock
|
| The Problem
|
| The Geometric Curse
|
| The Sim-to-Real Gap
|
| The Structural Gap f([?])
|
| Seeing It in Motion
|
| The N^2 Impedance Mismatch
|
| The Chaos Term [?]chaos
|
| The Information Wall
|
| The Weakest Link
|
| Why Manipulation Needs Better
|
| What We Built
|
| From 288 to 15
|
| Does It Work?
|
| Hardware Validation
|
| Robot Hand Landscape
|
| The Take-Home
| mort96 wrote:
| Oh my god, it's absolutely chock full of AI-isms. Almost every
| sentence is a list of 3 items, often _nested_ lists of 3 items.
| roenxi wrote:
| Deciding whether A is an X or a Y is a really basic part of why
| we're all communicating. Suspicion of em dashes is one thing,
| but once you start getting nervous on seeing "It's not X. It's
| Y." then you're just going to get paranoid.
|
| The fundamentals of an LLM is to statistically match their
| output with the corpus. The tics they have are really common in
| natural human usage too.
| knollimar wrote:
| Did you read the article? It's all AI tells. The tone may as
| well be a fax machine
| Animats wrote:
| That's a nice piece of motor engineering. It's well known that
| high ratio gearboxes for robots are a headache. Back driveability
| doesn't work, and tiny teeth are fragile. Comments on this go all
| the way back to Feynman writing about his time spent engineering
| automatic gunnery aiming systems in WWII.
|
| This new discovery is that gearbox problems mess up a machine
| learning system. It's trying to track gearbox noise and is using
| up all its learning capacity on that. This discovery means that
| robotics people can tap machine learning funding for motor and
| gearbox development. Robotics labs used to be really low-budget
| operations. No longer.
|
| What you really want is a direct drive motor, but those have to
| be large-diameter. They can be flat; that's a pancake motor.
| That's too large for fingers. So their compromise moves partly in
| that direction; the rotor is flatter, torques are higher, speeds
| are slower, and gearbox ratios are lower. As they point out,
| reflected inertia is the square of the gear ratio, because the
| gear ratio gets you both going out and coming back. So this is a
| bigger than linear win.
|
| Good back-drivabiilty means much less risk of gear breakage on
| overload. Some of the academic designs, such as harmonic drives
| and series elastic actuators, have huge gear ratios in a small
| space. That's OK for prototypes but not production. As I've
| mentioned before, "you cannot strip the teeth of a magnetic
| field", a line from a GE electric locomotive salesman around
| 1900. If an overload forces a motor backwards, nothing breaks.
|
| Would have been nice to hear more about the motor design. That's
| the real achievement here. There are CAD tools which understand
| electromagnetic fields now, so strange motor geometries are not
| as much of a trial and error and experience process as it once
| was. It's also respectable for an EE to work on rotating
| machinery again. That field matured around the 1960s, and until
| computers took over motor control, didn't change much.
| AceJohnny2 wrote:
| (I am not a mechanical engineer) Are capstan drives a solution
| to some of these gearing problems?
| Animats wrote:
| They don't back-drive well. The whole point of this hand
| design is to back-drive the contact forces into the motor,
| where there's force control. They're somewhat bulky, too.
|
| Key concept: force-based motor control works quite well.
| Preserve that property through the gear train and force-based
| hand control works.
| LoganDark wrote:
| > They don't back-drive well
|
| What? An ideal capstan drive can be backdriven perfectly
| fine. You only run into problems once it stops being ideal
| (e.g. built out of heavy parts, high gear ratio, etc.)
| Animats wrote:
| It's the high reduction ratio that's the problem. If you
| build a 200:1 capstan, it's not going to back drive well.
| And it won't be anywhere near ideal.
| kstrauser wrote:
| "This deployment is temporarily paused" is the crappiest "this
| account has gone over its budget" error page ever. Does that mean
| the site's down, or is that some meta-joke like "you've reached
| the end of the Internet"? Quick, explain to my non-technical
| friends what a "deployment" is. They're not trying to go on a
| deployment; they're trying to look at a website.
| 827a wrote:
| In this case, I'm not sure it matters what it says or how your
| non-technical friends interpret it. The site is down. Why it is
| down doesn't change the next thing casual viewers will do
| (close the tab).
| mort96 wrote:
| But it _does_ matter to their opinion of the site: is it down
| because the author took it down, is it down due to a
| technical problem, or is it is it down because the hosting
| provider took it down?
|
| "This deployment is temporarily paused", if anything, sounds
| like the people who put the site up took it down again. That
| sends the wrong message.
|
| Personally, if my hosting provider took my post down, I'd
| want them to make that obvious to my visitors. Or at the very
| least make it look like a technical issue. Not make it look
| like _I_ took it down.
| kstrauser wrote:
| > The site is down.
|
| Is it? The title is "The Robotic Dexterity Deadlock". For all
| I know, it's a joke about what deadlock looks like for
| robots, showing what could be interpreted as a deadlock in a
| webserver. At a glance, I can't tell if the site is down, or
| if it's up and correctly showing its very short message.
|
| So, yeah, in reality, I'm 99% sure it really is an error
| message. That's only because I've seen similar error messages
| in the past and can infer how to interpret it.
| ACCount37 wrote:
| You can redesign every single actuator, sure. Or you can design a
| wrist camera and use it to close the loop. Or just run enough
| imitation learning + "RL IRL" and fold the imperfect real world
| actuators into the policy implicitly.
|
| Robotics doesn't have a single silver bullet - the design space
| is vast and underexplored.
___________________________________________________________________
(page generated 2026-02-27 23:00 UTC)