When we were researching the fictional AI/Automation company, Polestar and characters for our book, The AI Analyst, we knew we could take some liberties as it was well, a fictional mystery. However, we wanted it to be somewhat realistic so I went back to my 2016 book on automation, Silicon Collar and extrapolated from that the likely “art of the possible”
For Silicon Collar, I had explored automation in settings in over 50 industries from advertising agencies to waste management. While many of them are white collar focused, many have blue collar and trade jobs. The BP use case in the book had focused on robotic crawlers, drones and autonomous vehicles on land and water to monitor pipelines and equipment in remote and risky places. Republic Services had focused on how a robotic arm on the garbage truck and sturdy bins for every house on the collection route had shrunk dangerous jobs of dangling at the back of the truck and emptying hundreds of stinky, heavy, un-standardized bags of garbage. We looked at how UPS telematics were allowing its drivers to maintain an impressive safe driving record, many without an accident over a million miles. How infusion pumps at the medical center at UCSF were helping deliver the right dose of the right drug at the right time for patient safety. How robotic harvesters were doing the picking jobs in many orange orchards and drones were used for spraying vineyards. How computational chemistry advances at Texas A&M using supercomputers were allowing simulations of many lab tests. How Kiva robots were doing the dull and physically tiring work of picking and ferrying items for the Amazon warehouse employee. How robotic butlers and drink dispensers were freeing up hotel employees to be responsive to more complex guest requests.
The book explored exoskeletons, autonomous vehicles and many other forms of automation in other digital and physical world settings. That with updates for progress over the past decade allowed us to paint Polestar as not having just a software or silicon focus, but one which blends today’s GenAI with humanoid robots, drones, UAVs etc. Polestar also has an expansive definition of its verticals as each of the 800+ occupations the Bureau of Labor Statistics tracks. Their solutions automate limbs, eyesight and other human faculties, not just cognitive skills. Lots of STEM disciplines, not just coding geniuses in action.
There is a scene in the AI Analyst book where Barry Roman, the CEO of Polestar visits the GE Global Research Center in Upstate New York and at lunch, he finds himself surrounded by chemical engineers, astrophysicists, and nuclear scientists. This was, of course, back in GE’s glory days, and the Center had over 1,500 technologists, many with PhDs covering every STEM discipline. He challenges his executives to build him one of those centers. He tells them: ‘I want us to be able to boast about Nobel Prize winners on our staff too.’
I will be honest, since most of the enterprise software vendors we know and love have only focused so far on white collar focused copilots and back office AI agents I wondered how many years it would take for that blending of GenAI and humanoid robots – our fictional scenario – to become reality.
Actually, it has been happening at a rapid clip. It just has a new moniker – Physical AI
Jensen Huang of NVIDIA talked about Physical AI at his GTC event in February and more recently expanded on his vision here
Elon Musk as he rolls out his Optimus humanoid robots says the appetite for them will be “insatiable”
Mary Meeker, in her usual prolific style has several examples of physical AI across several industries in her new report on AI trends:
“Self-driving fleets like Waymo’s and Tesla’s Full Self-Driving (FSD) beta are no longer science projects confined to test tracks –they’re revenue-generating deployments, logging millions of driverless miles with increasingly autonomous software loops.” and
“In defense, companies like Anduril are redefining what defense looks like –shipping autonomous drones and counter-intrusion systems with AI in every edge node, not just the command center.” And
“ In agriculture, companies like Carbon Robotics are putting AI into the dirt – using computer vision to eliminate weeds without herbicides.”
But most impressively, Marc Andreessen who famously wrote “Why Software is Eating the World” (which in my opinion was wrongly interpreted by many that software (and by extension GenAI) alone matters and that as a corollary, hardware and other infrastructure is trivial) recently said “The Next AI Revolution is Physical”. In this video he also talks about the coming reindustrialization of the US and the related physical AI opportunities.
All this did not happen overnight. Japan has been experimenting with humanoid robots to automate elderly care for decades. You could argue the auto industry has been evolving autonomy in vehicles since cruise control was first introduced in the 1950s. However, the use cases keep exploding. Ukraine has redefined warfare with its use of drones. Autonomous cars are not being sold to us as consumers but are starting to take off as robotaxis.
Blending agentic AI and robots and other physical automation – drones, UAVs etc. - widely expands the vertical reach. As Polestar finds out every one of the 800+ occupations the BLS tracks can now be given tools to turbocharge talent performance.
The opportunities could be humongous especially if the US starts to re-industrialize. It won’t be old-school factories, farms, hospitals, office buildings. And we may find rather than fretting about losing IT and coding jobs we are creating a whole new variety of jobs in other STEM disciplines.
I titled the book Silicon Collar because I said 9 years ago, irrespective of our occupation “we are no longer white, blue or brown collar workers – we are all Silicon Collar workers since technology is reshaping all our workplaces”.
Get ready for the next wave – all of us across industries and geographies are poised to become “superworkers”.