Hacker News Reader

Comment by tootie | Hacker News Reader

The desire for humanoid robots is purely aesthetic and almost certainly suboptimal. Ceiling mounted is actually a much more plausible option. Next most likely is something more like R2D2. A rolling garbage can with an arbitrary number of specialized appendages. Humanoids seem like a real dead end to me.

View on HN

I think something like https://fi.roborock.com/products/roborock-saros-z70 but just doing stuff with picking would more than enough.

Their benefit is they can operate in any space designed for a human without custom modification. No wires to get in the way with an assistant robot that interacts with the existing built world.

Much smaller robots will co-exist in a world with humans a lot more smoothly than anything human shaped & sized.

Yes, once they can be made to quietly jump on to a counter, like a cat.

They need a way to get onto counters, climb stairs, manipulate objects, et cetera. But all of the necessary capabilities do not need to be self-contained. Humans achieve their amazing flexibility through tool use, and so will robots. And in many cases those tools will be other robots. One robot could boost another onto the counter and then they both could get out of the way, unlike a robot large enough to reach up to the counter.

> once they can be made to quietly jump on to a counter, like a cat.

That would be cool. But also a little scary if it has an onboard AI which, based on it's embodiment, decides to develop a cat-like personality. Pretty sure if my cat had a better world model and long-horizon planning, it would kill me in my sleep.

Why? If you're dead, then it stops getting food and pets. If you're dead, it'll eat your face off if your neighbors don't break in first because of the smell.

I'm not so sure. If you could rebuild every human space and tool from scratch, you are probably right. However, we live in buildings filled with stairs. We've got tall cabinets with things tucked in them. We've got lawn mowers and snow blowers and leaf blowers and cars. Making a humanoid robot is a much harder problem if you want a cool demo. If you actually want to deploy robots in billions of homes, it might just be the only answer.

Just have to be compatible?

The xiaomi-robotics-1 concept is but one that seems very doable and very near term.

https://robotics.xiaomi.com/xiaomi-robotics-1.html

If the end-goal is robot maid, aka cooks and cleans for a household, it at minimum needs to handle stairs and uneven ground like gardens.

Something to get the laundry from each room, wash, dry, fold, put away.

Cut and cook ingredients then do the dishes.

Mow the lawn, water the plants and do basic weeding etc.

Maybe even climb a ladder to empty the gutters / clean windows.

A non-humanoid will never be able to accomplish that without dedicated implements.

Rebuilding and replacing tools doesn't sound that big of an issue to be honest.

I'm starting from the assumption the vast majority of people won't want a robot in their home, even just on principle. Then most people will not need it to do every possible action an human does.

For the remaining people in the target, having adapted/compatible cabinets or furniture is probably less of an issue. If someone requires assistance for instance, they'll probably also don't want hard to access shelves or crazy weird stairs.

The same way people have started choosing furniture that can accommodate a roomba(RIP irobot), we'll adjust our newer homes to whatever makes sense when we get the next wave of home robots.

That’s a flawed assumption. Everyone wants a live-in bangmaid. They’re just normally too expensive. If you could buy one for 30k, they would fly off the shelves.

Everyone wants a perfect robot at a low cost.

Except it won't happen, so the amount of people wanting a pretty flawed and compromised yet expensive robot will be pretty low.

We've built zillions of machines to automate tasks previously done by humans (or animals) and how many times did we design something based on humans work? Roughly zeroish? People used to do laundry with wash basin and a river, now it's a metal cube. People wash dishes with a sponge and a towel but the machine is a metal cube. People would cook over burning wood and now it's a metal cube.

The difference is this humanoid is meant to be a bespoke “does everything” robot. The others are very purpose-built/particular. And with good reason of course, an “everything humanoid” is beyond a moonshot in 2026

> almost certainly suboptimal.

I agree from an engineering and cost perspective that matching the embodiment, degrees of freedom and performance to the use case tends to be optimal. But entertainment, fun and inspiration can also be economically valuable use cases - even if less common.

That said, there are use cases which require navigating environments built for humans and the adaptive flexibility to perform unpredictable tasks as a human proxy. And those robots will probably be broadly humanoid shaped.

Ish. Millions of years of evolution and other than snakes and their ilk, land-faring animals have legs. So bipedal with wheels on the feet is maybe optimal. The optimal number of arms I think is three. No "head" is strictly necessary. The interesting question is how do you train a three armed robot when there are few three armed people?

How does R2D2 climb stebs?

ED 209 Can't.

Well if your legs trunnions were coffee cup lids, than stairs would pose an issue too.

https://www.youtube.com/watch?v=dFMJ514lz5w

The ED 209 is still cool, and only outclassed on screen by the T-800. =3

same way Daleks do

We might be at the beginning of a golden age for inventors. Robots of all shapes and sizes, with highly variable price points and specialization. Still, the first challenge for R2D2 is probably stairs.

It’ll be like phones. All kinds of whacky fun designs, before converging on a single black rectangle. Robots will all be flat black rectangles

Interstellar really got it right with TARS then.

The big benefit of humanoids is that they can be trained from captured human actions (egocentric datasets made by humans).

> captured human actions

Or just some dude on remote. (Get in the robot, Shinji!)

Do we really have to capture the humans? Legal says that introduces a whole slew of issues.

> The desire for humanoid robots is purely aesthetic and almost certainly suboptimal

Based on what? Do you think humans evolved into what we are based purely on aesthetic? Of course it may be suboptimal, but given enough time and ingenuity, so could literally everything be. Not sure that should stop anyone.

>Do you think humans evolved into what we are based purely on aesthetic

A modern house is a very different environment to the savannas of Africa.

Humans are def not designed optimally to live and work in a modern house or office. Don’t confuse “survived better than other animals in the Savannah 200 000 years ago” with modern env.

> Humans are def not designed optimally to live and work in a modern house or office

Agreed, those were designed by humans (mostly) for humans. I don't think we as a species "adapt to things" so much anymore, and instead we adapt our environment to us.

> Humanoids seem like a real dead end to me.

We built the world for us, not for the most sensible robot.

So for generalized solutions, a good biped will have some advantages.

Reply on news.ycombinator.com