Human Labor, AI and Robotics
Technological change for better or worse ?
The impact of AI and robotics on labor has been on my mind for years, and the issue I think is more relevant now than ever, pressing even, while it has been tempting to bucket AI under general technological change which is more or less understood, I believe this is somehow naive as AI is proving to be a different animal all together, or rather the promise of AI (AGI in particular) coupled with robotics seems like it could replace all labor, which sounds great as many people would rather not work, until you realize 99% of the world needs a job to survive, then what ?
How much does it cost to hire a human ?
About $10–$20 per day (~$2/hr ), keep in mind there’s a wide range of wage disparity and the bulk of human economical activity ( about 50% or 2 billion humans ! ) is done by unskilled workers on the lower end or below the average cost ( $5/day $0.60cts/hr for garment workers for instance).
So it would make sense from an economical perspective to replace human labor if the same work could be done for less.
How much would a general robotics, general AI worker cost ?
Some factors to consider are how many hours a general robotic worker could function,maintenance costs and downtime and of course initial investment, industrial robotic arms can function for around 10 yrs or ~90,000hrs and about half that if there is no downtime, let’s use 60,000hrs or two 8 human hour shifts in a 10yr period as a reference point, the equilibrium points for different human wages would be:
The price tag for a general robotic laborer then needs to be in the $37,000- $220,000 price range to make economic sense.
Lowering the cost of technology is typically a dificult and long process,
so it seems reasonable to expect general robotic workers will follow suit
and lower paying wages will be safe for some time.Development costs
It is unclear what the development costs will end up being for a general robotic laborer, such laborer if we extrapolate from current technologies could have a software component along with a hardware one, the development of chatGPT was probably in the hundreds of millions, same for early industrial robots like the unimate.
It is most likely though that the Tyrell Corporation (or Wayland Yutani if we are unlucky) of our future will benefit from these early investments, after all Amazon didn’t invent the internet, or ABB ( biggest industrial robotics company ) the industrial robot.
Human labor replacement by technology
Is nothing new, just over 100 years ago about 40% of the US workforce used to be employed in agriculture, these days only about 2%, this dramatic change can mostly be attributed to industrial farming technology, it is estimated that industrial robots replace about 1.6 manufacturing workers, millions of office jobs were “lost” to computerization in the early 2000’s, small mom and pop stores have mostly disappeared due to online commerce, tellers have been replaced first by ATMs and now by apps and on and on.
The promise of AI/AGI plus advanced human like robotics, what I’ve been calling the general robotic worker does seem different in magnitude and coverage, while we’ve been focusing on unskilled labor, semi skilled and skilled labor are also in the crosshairs of AI as demonstrated by current generative and human like chat AIs which paradoxically started from very skilled and difficult tasks.
The aftermath of technological disruption.
Most pundits will claim that either the sky is falling or that the future will be rosy cheeks, in reality and based on previous technological shifts* there are winners and losers and these changes take time, the process goes something like this:
This Creative Destruction typically produces societal gains ( same day shipping on lower priced items ! ) and loses (unemployment, monopolies, misery ), measuring these, especially before the fact is all but impossible but at least we can point to the past to inform our predictions on the future.
*It bears repeating that what makes this technological change
different is the numbers, the dramatic shift in the US workforce can
roughly be meassured in a "few" million displaced jobs over 100yrs, it is
to be seen if these wheels of progress can withstand billions of displaced
jobs in a shorter time period.Where we at ?
Of course this is all in the possible future as we don’t have a general robotic worker yet or the Artificial General Intelligence to drive it, let alone a comercial product, but it’s interesting to note the progress made so far:
You can of course come up with different percentages, here's what I based
mine on:
Humanoid Robots : Atlas, Figure 01, Cyber-01 etc,etc.
AGI : ChatGPT/Llama/Claude etc,etc... Hard to tell if a LLM is a the
correct path to AGI, hence the low score, but I could be wrong.
Integration: Non existent as the previous ones are needed, but warehouse,
home,manufacturing and other integrations do exit.
Commercialization : Like the previous item still very niche and primordial.We are from the government and are here to help !
These are the words that scared ex US president Ronald Reagan and words that still strike fear in many, so how would governments respond to this technological change ? Most likely in a poorly and untimely manner.
The easiest governmental response is to ban and regulate things when societal discomfort is perceived, this usually happens early due to insecurity ( like crypto currencies ) or late due to complacency ( like climate change ), ignorance is usually also a factor.
Banning robotic labor seems a bit of a fools errand though as a country with less unskilled labor could easily adopt it as a competitive advantage.
The correct answer to this and other problems consists of having bigger and
better social safety nets, UBI, just taxation,infrastructure,education and
technological investments, good laws, good enforcement and many other
things that constitute a good representative government and can be seen
as "Organization" rather than "Confrontation". In other words your own
countries response will be as good or bad as your government... Bold and not so bold predictions
- The majority of human jobs are safe for now but automation will gradually encroach on them.
- The shift to a global workplace dominated by robots and AI will take time.
- Unless a black swan company/technology produces the goods, this would create havoc.
- It’s not unreasonable to believe those born in our lifetime will have to compete with robotic workers for most types of jobs.
- Current incumbents might not necessarily be the first to market or even dominate the field of general robotics.
- The first individual trillionaire will be tied to a general robotic worker product and company.
- The general robotic worker will go by another name.
- Countries with lower unskilled labor will benefit the most.
In the end you are responsible for your own job safety, if you believe that it is in danger from automation or other factor, it most likely is, you can learn new skills, move within your industry or even be the one that brings forth said change, hoping that things remain the same or that the government will step in to save your job I think are not valid options.
Thanks for reading !
