The development of general-purpose robots capable of performing household chores has been a long-standing goal in the field of robotics. Such a robot would need to be versatile, adaptable, and capable of handling a wide range of tasks and environments commonly found in American homes. Achieving this level of capability remains a significant challenge.
Will a general household robot capable of performing household chores to a high level of reliability be developed before January 1st, 2030?
Resolution Criteria:
This question will resolve to "YES" if, before January 1st, 2030, a general household robot is developed anywhere in the world and has been publicly and credibly documented to have:
Demonstrated the ability to autonomously navigate and operate within a variety of residential environments, including:
a. Identifying and avoiding obstacles, such as furniture and pets.
b. Maneuvering through doorways, hallways, and multi-level spaces.
c. Adapting to different floor types and surface conditions (e.g., carpet, tile, hardwood).
d. Recognizing and safely handling fragile objects, such as glassware or delicate appliances.Exhibited proficiency in performing a comprehensive range of standard household chores, including at least 5 of these complete chores:
a. Cleaning tasks, such as vacuuming, sweeping, mopping, dusting, and tidying up clutter.
b. Laundry tasks, including sorting, washing, drying, folding, ironing, and putting away clothes.
c. Dishwashing tasks, like loading and unloading a dishwasher, or washing, drying, and putting away dishes by hand.
d. Cooking tasks, such as meal planning, ingredient preparation, cooking, and serving.
e. Maintaining indoor plants, including watering, pruning, and repotting.
f. Basic pet care, such as feeding, grooming, and cleaning up after pets.Shown the ability to adapt to user preferences and instructions, including:
a. Learning and adjusting to individual household routines and schedules.
b. Following specific instructions regarding cleaning methods, food preparation, or pet care.
c. Recognizing and responding to verbal and non-verbal cues from household members.Demonstrated a high level of reliability and safety while performing these tasks, with:
a. Consistently high-quality results that meet or exceed the performance of most humans performing the same tasks, within 500% of the average time it takes humans to perform these tasks.
b. A low rate of errors, accidents, or damages to the home or its contents. More specificially, it is required that, when given full instructions, the robot can fail no more than 5% of the time on average while completing a designated chore. A fail counts as any error that would require human intervention to fix, as otherwise the chore would not be completed to even a minimum level of satisfaction.
c. The ability to recognize and avoid potentially dangerous situations, such as electrical hazards or fire risks.
The development must be accompanied by independent reviews, testimonials, or high-quality case studies documenting the robot's performance in real-world residential settings, demonstrating its ability to perform tasks consistently and effectively, with a high level of satisfaction among users.
I will use my discretion when resolving this question, possibly in consultation with experts, to ensure that the criteria are met and that the general household robot is indeed capable of performing standard household chores to a high level of reliability.
Related questions
Looks like we're not too far away (Figure 2 announcement): https://x.com/adcock_brett/status/1820792697315348640
@RemNi They don't believe that a reliable and general household robot will be developed in the next five years
@Tumbles but only one of these needs to be developed, and it could use a very significant amount of compute for the demonstration of its capabilities. This question is very generous.
@RemNi It also has a lot of specifics, and the people developing such a robot wonโt be developing it to the specs of this question, so it would probably need to be even more broadly capable than as described in a number of other ways
Nobody has yet built an appliance-level-reliability home device with three or arguably even two actuators in series. The robotics and software challenges are so far beside the point compared to the service and maintenance problems and environmentals (aerosolized grease, diversity of climates, etc.).
Current robotic systems die on issues like dirty sensors, manipulator surface durability, and even cable wear. We are a long way from even starting to have the sorts of problems that we are now (appropriately) excited to be starting to solve in software.
I actually think the only bottleneck to this is software, a human operator controlling a robot is already capable of nearly all these tasks. Self driving cars have the same physical mobility as humans driving, itโs just the brain controlling it is dumb and canโt generalize as well as a human so they havenโt taken off yet.
So I find it surprising that this market sits much lower relative to the optimistic AI predictions on this site.
If I was a robot manufacturer, Iโd be targeting industry first. The edge cases for a house are harder: manoeuvring around without falling on the kid thatโs trying to trip you up. There are thousands of factories and workplaces that can soak up the supply of robots from a company trying to ramp up. For how many years? And the data from these environments might be necessary for generalization to households.
Robot with Teleoperation
Video: https://www.youtube.com/watch?v=HaaZ8ss-HP4
Site: https://mobile-aloha.github.io/
@Lorxus It learns by teleoperation, but then can do the learnt task autonomously.
The market has until 2030, that's plenty of time for improvements.
I've come up with a set of rules that describe our reactions to technologies:
Anything that is in the world when youโre born is normal and ordinary and is just a natural part of the way the world works.
Anything that's invented between when youโre fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.
Anything invented after you're thirty-five is against the natural order of things.
- Douglas Adams, The Salmon of Doubt
This one looks promising: https://www.therobotreport.com/apptronik-unveils-apollo-humanoid-robot/
It's weird for me to be in the tech pessimist crowd on this site, as I'm in the optimist camp in almost every in-person social group.
The development of the type of robot envisioned in the description would be industry- and life-changing. Unless the cost were truly exorbitant, such a robot would quickly become a staple in every physical industry. The incentive to invent such a robot is immense.
But a positive resolution contemplates incredible advancements in hardware and AI capabilities between now and market close, in a format (robotics implementation) where we historically have seen snail-paced innovation. It seems the closest parallels are Roomba, custom manufacturing robotics, and self-driving cars....and it's not even clear when self-driving cars will be economically viable.
For such an undertaking, investors will likely need proof of blowout financial success for smaller-scale robotics technology before pouring investment capital into more ambitious projects. We're only a bit over 7 years away from market close, and it seems extremely unlikely product cycles can turnover to create "Capable Household Robot" between now and then.
@CarsonGale
> It's weird for me to be in the tech pessimist crowd on this site
Most are betting "no" here, and the actual probability seems to be mostly just Catnee, and we all know he is a chaotic cat :3
A robot that had lots of manipulator arms but no locomotion capabilities would still be very useful for some types of household chores, and might be easier and safer to commercialize than a fully mobile robot.
I'm imagining that you have a device that you manually wheel into your kitchen or laundry room, set up your dishes or your laundry or whatever within it's reach, and switch it on, and then it starts grabbing and sorting things within its reach without its base needing to move around much, if at all.
To me, that would be a "reliable and general household robot" and a revolution in household robotics, but it wouldn't meet the criteria in this question due to lack of locomotion.
I feel like the criteria are all reasonable and should count towards the market except one: the ability to navigate multi-level environments. I'm not really convinced that robot manufacturers will put much effort into solving a problem that to me looks fairly complex, so I think the market should resolve to yes even if that specific point is not met.
@NikitaBrancatisano A household robot unable to move itself up/down stairs, etc would be less practical. It just depends on what the author intends -- is the spirit something that roboticists are actually likely to develop this decade, or is the spirit meant to be a very high bar?
To me the other criteria heavily imply that it's meant to be a very high bar.
@Jotto999 Agree. Also - I actually expect that navigating multi-level environments is way easier than say, washing dishes.
Here's Boston Dynamics' CEO talking about it https://youtu.be/cLVdsZ3I5os?t=5692