MoarrrRrrrobots

[This post originally published elsewhere 2024-10-5]

People love robots, but will they love us back?

The notion of robots plucks at the strings of popular imagination. We write about them in endless fascination, anthropomorphizing them, embodying them with complex human emotions and granting them autonomy and human morphology. Robots fascinate us in a way that few other ideas do, except perhaps faster than light space transport and time travel, both of which seem to always come with robots.

The image of a wholly obedient servant whose sole purpose is to execute one's whims is the stuff of idle porch chatter, whimsical fantasy and feverish nightmares. Generations of storytellers have spun yarns of automatons whirring and beeping in obedient response to lords and overlords. While it once seemed that robots would manifest as people-shaped and metallic, the reality has been more prosaic.

The twenty first century has been a mix of droll disappointment and comical trolling of consumers with new ways to spend money on plastic junk. However, there have been some ideas that were initially dismissed as laughably amateur efforts, but which have since morphed into very real harbingers of things to come. The future seems to follow a few steps behind these clumsy vanguards, slipping in to town quietly after the echoes of our laughter subside. In the first generation, risible and amateurish, second generation offerings have been quite plausible. In many cases, iterative refinements have moved barely-useful things into the realm of the indispensable.

Robot vacuum cleaners and robot lawn mowers - no longer targets of derision

Instead of our robots being named Robby, with two arms and legs, eyes, mouth and nose, they have tended to be one-armed, often with a laser or welding attachment, tasked with small, well-defined jobs on production lines. They have the ability to perform fixed tasks, over and over, with unerring accuracy. That is also their main shortcoming -- they can do only those things that we painstakingly program them to do. If a mouse appears on the car production line and gets in the way of the laser, the door-welding robot can't pause and gently nudge the mouse out of the way so it can get on with the business of car making. The car making robot can't respond to unprogrammed circumstances.

Other robots have shown up that have had more flexible programming than the door welder and the eyeglass lens polisher. Robot vacuums, for example can traverse areas they've not been specifically programmed for. Initially, a minor novelty with niche usefulness, robot vacuum cleaners suffered from some hilarious shortcomings. Their algorithms, while making for good coverage of floors for the purposes of picking up dust and lint, also ensured a good, even spread of things like pet poop and plant soil. As the guffaws faded, manufacturers took note.

Their initially dumb floor coverage schemes evolved with the addition of sensors, room-mapping calculations and other evolutionary improvements. The robots became more useful, adding the ability to self-empty, run longer, gracefully avoid poop, and even mop. They merged into the background of our everyday lives, being included in kitchen layouts, with interior designers adding integrated robot vacuum docks to the list of must-have luxuries, along with wine fridges and pot fillers. They began appearing in movies, with Roombas being driven around by mini Stay-Puft marshmallow men in Ghostbusters:Afterlife.

Having a Roomba rolling around one's floors is no longer a novelty. For many of us, having dusty triangles in the corners of our rooms is a small price to pay for having the rest of our floors manicured by our robots. It's this idea that robots' ability to do somewhat imperfect things endlessly, repetitively, for as long as their power supplies last, that is the key to understanding the shape of our collective future. We don't need cordon bleu robots able to make scrambled eggs, or even be able to pronounce les œufs brouillés. All we need is for them to do dumb things over and over again, getting better at them gradually. Will a robot eventually make a frittata? It's possible, but a robot chef feels like less of a problem than having one capable of packing and unpacking the dishwasher.

Unpacking the problem, unfolding the challenges

Let's take the problem of the clothes folding robot, a favorite of tinkerers and inventors. Laundry has long been a time suck, using unimaginably large percentages of humanity's collective labour hours on the simple business of getting food stains out of shirts, then folding and putting away those now cleaned clothes. The washing machine is a big deal. We forget now, in the twenty-first century, of the struggles women had just to get the laundry done and how only recently were they freed from the pernicious slavery of laundry labour by the Maytag Corporation and the marvel of the fully automatic washing machine. There is a school of thought that suggests that one of the biggest advances in modern gender equality is the now-common expectation that men also do laundry. With the problem of laundry now being less of labour and more of management (i.e. keeping an eye on the machine and emptying it when it's done) we're moving on from the problem of who does the washing to who folds up the laundry and puts it away. Lots of us just don't, and I don't blame you.

Let's consider that mowing the lawn has been a traditionally male-gender role and doing the laundry has been a traditionally female-gender role. Now that doing the laundry has been automated, with focus moving from the job of washing to the second-order job of folding and putting away, can we look around to other jobs that have been automated, freeing us up to look to other parts of the labour cycle for improvements? Unlike laundry, there's no follow-on problem to lawnmowing. The mowing robot just mows and leaves the chopped-up grass on the surface as mulch. Because the mower runs often, the cuttings are short, making for ideal mulch. No bagging here, no dethatch. What is pure tedium for people (mowing twice a week? NFW) doesn't annoy a robot in the slightest. The robot is happy to stop mowing once the rain starts, resuming the moment the grass is ready to be cut again. No more does the male have to watch the skies, waiting for that moment to get out there and mow. Coupled with the transition to battery-powered mowing (gas-powered robot machinery is more risky and maintenance-intensive) we are able to ignore the lawn, finally. This is what robots are meant for!

In the same vein, robot weeders are here. While still in the niche phase of adoption, with various approaches being tested (Blast with lasers! Chop with mini string-trimmers! Slice with rotating cutters!) we will probably end up with different solutions for different problems. Commercial growers will be able to greatly reduce and possibly even completely eliminate pesticide use. With a tireless robot arm pinching off little shoots as soon as they appear, it's going to be hard for weeds to get going. I have no doubt that crafty weeds will work out that the only way to survive is to look like the actual crops being grown, but crafty farmers have also been in this arms race for centuries and I suspect they'll not just give up now.

Going back to domestic problems like laundry, it's natural to look at the folding problem. While washing machines have evolved from the Thor Washing Machine into the fully-automated machines of today, the business of clothes folding seems to be stubbornly resisting automation, for the moment. AI seems to be the route that most experimenters favor, with the random and amorphous nature of a basket of clean clothing seemingly an excellent problem match for a pattern recognizing robot.

A black and white 1920 advertisement for the Thor electric washing machine.

The Thor - the original washing machine

It was simple. The difference it made was not. The first step in liberating women from the punitive drudgery of hand laundry, this was one of the moments when you could glimpse the future winking back at you.

An interesting side note on the Thor was that it was marketed as being both a clothes and a dish washing machine. While I wasn't familiar with the dishwashing function, it's interesting to note that those two functions evolved into different devices. This is likely because a more efficient washing machine makes for a less efficient dish washer, making the divergence of the two functions into separate devices obvious, in hindsight.

While folding seems intractable, consider that it took a hundred years to get from the Thor to today's hyper-efficient front-loaders with sensors and dynamically variable wash programs. Considering that while FoldiMate and Laundroid were the early efforts in the folding space less than five years ago, we can assume that others will appear once the VC money machine swings into action again. It takes a few cycles of money-burning before we get serious.

Less difficult than clothes folding, however, is lawn mowing. Robot lawnmowers have been evolving rapidly, exploiting the same technology improvement curves that have benefited robot vacuums — improved battery capacities, better mapping strategies, smartphone integration and uniquely, with mowers, GPS for boundary mapping. Initially perceived as comedically dumb, mower makers have been plugging away at improvements. Like so many technology-driven product arcs, autonomous mowers have been locked into an arms race, specifically with the problem of theft. Expensive and highly portable and usually left outside in the dark, these are high-value targets. Like car and smartphone theft, the problem is gradually being solved with a combination of strategies like PIN+App control, remote disablement, GPS location reporting and surveillance features.

What's clear from the evolution of robot vacuums and robot mowers is that our robots aren't walking, talking people analogs. They are simple and function-focused, containing just what's needed to do a specific job. They are initially very basic, with marginal or questionable value. They are initially expensive. They are initially able to offer some incremental time and labor savings, making them useful for people who value time savings and convenience and for whom cost is less important. Over time, as the device's effectiveness increases and the costs of manufacturing fall, they become more generally useful and attractive to a larger audience.

Once cultural acceptance is reached, they're part of our lives, like dishwashers and washing machines, with everyone having one. When a seemingly silly product shows up, think about how a few small improvements could turn it into something everyone wants.

Previous
Previous

Developments in the EV space - some second order thinking

Next
Next

Oil power - an anachronism in the making