1.
Humanoid robots.
I feel like something that isn’t super well appreciated in civilian world is quite how quickly humanoid robots will arrive, and quite how good they will be.
See, there are a bunch of humanoid robotics companies now. Here are five robots:
- Atlas by Boston Dynamics – doing the hard engineering graft of movement for years. Look at the vid and the uncanny way the robot stands up.
- Figure – the hot, glossy US startup building general purpose humanoids with a big mission (but actually going after factory jobs).
- NEO by 1X – another gorgeous general purpose humanoid, this one out of Norway and upholstered in plump beige fabric.
- Digit by Agility Robotics – humanoid warehouse robots.
- Walker S1 by UBTECH – I don’t have great knowledge of the Chinese market, this is just one company. A big one though.
The way I understand it, there have been three major challenges with robots in the real world: mechanical engineering, perception, and instruction following.
Engineering has been solved for a while; perception mostly works, though not understanding. Instruction following, including contextual awareness, task sequencing, and safety… that was a work in progress. Solved at a stroke by gen-AI.
So, as of a couple years ago, there is clear line of sight to humanoid robots in the market. Research done, development phase: go.
Now we’re waiting for the shoe to drop.
2.
Robots and nails.
All physical automation gets called a robot now.
Here’s a new fully automated manicure machine (10beauty.co).
All their press refers to it as a robot. But it doesn’t have any legs. It’s a clever box you put your hand inside and it looks like a Bene Gesserit pain device, only instead of fear being the mind-killer, it softens your cuticles.
But this is why the humanoid form is good for robots: you shouldn’t need to develop a dedicated machine for manicures.
The friction for any technology during deployment phase is the build-out for integrations. For robots, integrations = operating in the physical world.
e.g. it will be hard and slow to bring AI-powered automation to warehouses and production lines because the warehouses and production lines will need to be re-tooled. Slow. Expensive.
Unless! Use the human form and you can pick up existing tools and press existing buttons.
At home, why buy a Roomba if your aging-in-place domestic robot assistant can pick up the exact same hoover that you own anyway? And carry it up and down the stairs too.
KINDA RELATED:
This robot hand has fingernails (Robotic Gizmos):
Thanks to this approach, robots can pick up coins and cards and achieve better tactile sensing. These fingernails also assist with dynamic manipulation.
Clever.
Also… I wonder if the fingernails can also break off and then be replaced?
There’s utility in having a sacrificial part of the body, when you’re interacting with the physical world, whether it’s nails on your fingers or the whiskers of a cat. (Cat whiskers are deep-wired tactile sensors that easily regrow when damaged. An intelligent design!)
3.
Economic takeoff.
Back in 2022, factories crossed the robot takeoff threshold.
…a nascent trend of offering robots as a service - similar to the subscription models offered by software makers, wherein customers pay monthly or annual use fees rather than purchasing the products - is opening opportunities to even small companies. That financial model is what led Thomson to embrace automation. The company has robots on 27 of its 89 molding machines and plans to add more. …
Thomson pays for the installed machines by the hour, at a cost that’s less than hiring a human employee … “I’m paying $10 to $12 an hour for a robot that is replacing a position that I was paying $15 to $18 plus fringe benefits.”
Grimace-emoji here.
This is when the market snowball effect kicks in… when cost-benefit is positive, even a little, that increases volume, which decreases cost, which makes cost-benefit more positive, which… and so it accelerates.
Figure publishes its Master Plan:
We believe humanoids will revolutionize a variety of industries, from corporate labor roles (3+ billion humans), to assisting individuals in the home (2+ billion), to caring for the elderly (~1 billion), and to building new worlds on other planets. However, our first applications will be in industries such as manufacturing, shipping and logistics, warehousing, and retail, where labor shortages are the most severe.
And all of this is what I mean when I say people are sleeping on humanoid robots coming into the world.
It’s not going to be like ChatGPT and that explosion into cultural awareness. More like the rise of Waymo and robot cars which are startlingly good and startlingly mundane.
And it might be a year or it might be five.
And before we know it, a ton of people are out of work, on top of all the people who no longer have call centre jobs, and 90% of the commercial artists, and there aren’t any jobs for them to go to, and that fact really, really ought to be the subject of a lot of policy thought, like today.
4.
Who will make clothes for robots?
Not meaning to be flippant but at least some of the future new jobs will be in robot fashion?
I look at those gorgeous upholstered NEO Gamma robots by 1X… NEO’s Knit Suit is soft to the touch and
flexible for dynamic movements.
And tasteful beige knitwear wouldn’t survive a day in my home.
Even those high-capital-cost factory robots are going to have their elegant metal bodies scratched and dented.
So… they need clothes?
Protective workwear, sure. You need something washable and semi-sacrificial for physical work.
But also: the clothes that a robot wears are part of its brand. They feed into how we relate to it, how we remember it, how we feel about it.
Which sounds a lot like fashion?
I’m reminded of how the Apollo spacesuits were designed and manufactured by a commercial underwear brand:
The Apollo suits were blends of cutting-edge technology and Old World craftsmanship. Each suit was hand-built by seamstresses who had to be extraordinarily precise; a stitching error as small as 1/32 inch could mean the difference between a space-worthy suit and a reject. …
A division of the company that manufactured Playtex bras and girdles, ILC had engineers who understood a thing or two about rubber garments. They invented a bellowslike joint called a convolute out of neoprene reinforced with nylon tricot that allowed an astronaut to bend at the shoulders, elbows, knees, hips and ankles with relatively little effort. Steel aircraft cables were used throughout the suit to absorb tension forces and help maintain its shape under pressure.
So I don’t imagine 1X, Figure, Boston Dynamics and the rest will be creating their own robot clothes. Not their wheelhouse. They’ll partner with existing fashion brands.
Which means there’ll be a new need for new fashion designers. Fashion designers with experience in robot apparel, serving well-dressed artificial humans in head-to-toe haute couture.
Start brushing up your portfolios kids.
1.
Humanoid robots.
I feel like something that isn’t super well appreciated in civilian world is quite how quickly humanoid robots will arrive, and quite how good they will be.
See, there are a bunch of humanoid robotics companies now. Here are five robots:
The way I understand it, there have been three major challenges with robots in the real world: mechanical engineering, perception, and instruction following.
Engineering has been solved for a while; perception mostly works, though not understanding. Instruction following, including contextual awareness, task sequencing, and safety… that was a work in progress. Solved at a stroke by gen-AI.
So, as of a couple years ago, there is clear line of sight to humanoid robots in the market. Research done, development phase: go.
Now we’re waiting for the shoe to drop.
2.
Robots and nails.
All physical automation gets called a robot now.
Here’s a new fully automated manicure machine (10beauty.co).
All their press refers to it as a robot. But it doesn’t have any legs. It’s a clever box you put your hand inside and it looks like a Bene Gesserit pain device, only instead of fear being the mind-killer, it softens your cuticles.
But this is why the humanoid form is good for robots: you shouldn’t need to develop a dedicated machine for manicures.
The friction for any technology during deployment phase is the build-out for integrations. For robots, integrations = operating in the physical world.
e.g. it will be hard and slow to bring AI-powered automation to warehouses and production lines because the warehouses and production lines will need to be re-tooled. Slow. Expensive.
Unless! Use the human form and you can pick up existing tools and press existing buttons.
At home, why buy a Roomba if your aging-in-place domestic robot assistant can pick up the exact same hoover that you own anyway? And carry it up and down the stairs too.
KINDA RELATED:
This robot hand has fingernails (Robotic Gizmos):
Clever.
Also… I wonder if the fingernails can also break off and then be replaced?
There’s utility in having a sacrificial part of the body, when you’re interacting with the physical world, whether it’s nails on your fingers or the whiskers of a cat. (Cat whiskers are deep-wired tactile sensors that easily regrow when damaged. An intelligent design!)
3.
Economic takeoff.
Back in 2022, factories crossed the robot takeoff threshold.
Grimace-emoji here.
This is when the market snowball effect kicks in… when cost-benefit is positive, even a little, that increases volume, which decreases cost, which makes cost-benefit more positive, which… and so it accelerates.
Figure publishes its Master Plan:
And all of this is what I mean when I say people are sleeping on humanoid robots coming into the world.
It’s not going to be like ChatGPT and that explosion into cultural awareness. More like the rise of Waymo and robot cars which are startlingly good and startlingly mundane.
And it might be a year or it might be five.
And before we know it, a ton of people are out of work, on top of all the people who no longer have call centre jobs, and 90% of the commercial artists, and there aren’t any jobs for them to go to, and that fact really, really ought to be the subject of a lot of policy thought, like today.
4.
Who will make clothes for robots?
Not meaning to be flippant but at least some of the future new jobs will be in robot fashion?
I look at those gorgeous upholstered NEO Gamma robots by 1X…
And tasteful beige knitwear wouldn’t survive a day in my home.
Even those high-capital-cost factory robots are going to have their elegant metal bodies scratched and dented.
So… they need clothes?
Protective workwear, sure. You need something washable and semi-sacrificial for physical work.
But also: the clothes that a robot wears are part of its brand. They feed into how we relate to it, how we remember it, how we feel about it.
Which sounds a lot like fashion?
I’m reminded of how the Apollo spacesuits were designed and manufactured by a commercial underwear brand:
So I don’t imagine 1X, Figure, Boston Dynamics and the rest will be creating their own robot clothes. Not their wheelhouse. They’ll partner with existing fashion brands.
Which means there’ll be a new need for new fashion designers. Fashion designers with experience in robot apparel, serving well-dressed artificial humans in head-to-toe haute couture.
Start brushing up your portfolios kids.