RethinkX

Cheaper, Faster, Better: How converging technologies in robotics and AI present enormous opportunities for innovation

Written by Bradd Libby | Sep 25, 2024 7:52:38 AM

By Bradd Libby

Disruptions come from a convergence of technologies. Cars, for example, were born primarily from the internal combustion engine and bicycle parts. Look at this photo of Carl Benz’s 1885 Patent Motor-wagon, the first commercially available automobile. It is made of the same parts as bicycles of the day—tubular steel, solid-rubber spoked tires, and a chain-and-gear drive. To invent the car, you do not need to reinvent the wheel. You just need to find a way to put existing bicycle parts together, along with an engine, in the form of a motorized carriage.

Benz’s 1885 Patent Motor-wagon, Source: Wikipedia

Now, new technologies that are rapidly decreasing in price and increasing in capabilities —artificial intelligence (AI), lithium-ion batteries, infrared cameras, lidar laser range detectors, actuators, 3D printing and more—are all converging simultaneously, unlocking entirely new uses we are only just beginning to glimpse.

Better, faster, cheaper

New technologies are often rapidly adopted if they deliver some value that is radically better than any incumbent offering. Benz’s first car delivered two-thirds of a horsepower; later versions had larger engines that delivered 1.5 horsepower. Owning one was little better than owning a horse. Only about 25 were ever sold. But starting 20 years later, Henry Ford’s Model T was equipped with an engine that delivered 20 horsepower, 10 times better than a pair of horses. He sold 15 million.

We are now seeing similar performance multipliers in the use of drones for everything, from pesticide spraying to window cleaning. These machines could reshape their respective industries as quickly as the car industry disrupted the horse.

The South Florida Sun-Sentinel reports that “Broward [County, in southeast Florida] is now using a drone to spray for mosquitoes in hard-to-reach places.” What would take a guy “out there for two to four days” without a drone, the article reports, “could be treated in four hours by drone,” about one-quarter to one-eighth of the time, and at about half the cost. That sounds like a Model T to me.

Similarly, companies like Skyline Robotics are tackling the challenge of maintaining skyscrapers with their window-cleaning robot, Ozmo. By combining AI, machine learning, lidar (laser-based range finding) and advanced robotics, they’re addressing the issues of labor shortages and worker safety in a traditionally dangerous job. And they claim their technology “delivers results three times faster than traditional methods.”

Not just better, different

The convergence of technologies doesn’t just make things happen cheaper and faster—it creates new possibilities and expands markets to new customers. By combining 3D imaging, machine learning and precision robotics, the makers of the Perceptive robot dentist claim it can perform procedures faster and more accurately than human dentists. IEEE Spectrum reports that “a dental procedure that typically requires several hours and multiple visits” can instead be completed by a robot “in minutes in a single visit” with “better results.” Such a robot could expand the market for dental services, not only by being faster, but also serving a large segment of the market that is currently underserved: people who are too ashamed by the quality of their teeth to expose themselves to the potential judgment of a human dentist.

New York-based Aescape is positioning its personalized massage therapist similarly—a device that could meet a currently unsatisfied demand for a service, rather than replacing human therapists. Speaking to Fast Company, the company’s founder said, “Aescape’s goal is not to replace massage therapists, but rather to act as a complement to a space that he says is drastically underserved.” That is, massage on demand, when needed by the customer, not based on when massage therapists have an available slot in their calendar.

But few technologies have the potential to expand markets as much as humanoid robots. A robotic home servant has been a staple of science fiction for many decades, but now we are on the cusp of seeing units commercially available for consumer purchase. Companies like Norway’s 1X are developing humanoid robots like their NEO that can safely interact with humans. This is made possible by the combination of high-torque, low-speed motors, advanced sensors like lidar, and AI. These robots and others like them are not just more advanced industrial arms—they’re a new category of machine capable of working alongside humans in factories and offices, and helping around the home. Who would own a dedicated dishwashing machine if their humanoid robot could wash dishes by hand instead? Moving to an apartment with lower rent might sound really tempting if a team of robots could quickly and cheaply pack your belongings and help you unpack them at your new place.

In a recent blog post, “How much should we invest in humanoid robotics?”, RethinkX’s director of research Adam Dorr presented a simple humanoid robotics ROI calculator that shows that under very reasonable assumptions about the price and productivity of humanoid robots, the impact of their labor on GDP could be enormous.

These technologies—like industrial drones, dedicated high-skill robots and general-purpose humanoid robots—are just becoming enabled now due to recent breathtaking drops in cost and increases in performance in the underlying technologies of AI, lithium-ion batteries, lidar and more.

RethinkX co-founder Tony Seba saw years ago where these technologies were headed. Speaking in 2016, at a time when lidar units just a few years before had cost tens of thousands of dollars, he said, “The next generation [of lidar] is going to cost $90 and it’s going to be the size of a postage stamp. You’ll be able to use it on your iPhone.” Apple began including lidar on their devices just a few years later.

Now robots like 1X’s NEO can use lidar to operate safely in home environments. And cheap lidar is enabling the autonomous taxi service that Waymo is expanding rapidly in terms of customers served. Waymo recently stated that they are serving more than 100,000 customers per week, a number double what they reported just a couple of months ago.

(Tesla, which uses cameras instead of lidar to enable their cars’ self-driving features, will perhaps unveil a robotaxi service in a highly anticipated scheduled event in October.)

Survival of the fastest

The improvement in the performance-to-cost ratio of all of these underlying technologies in recent years has been astonishing, enabling longer range, longer-lasting, faster, more capable drones and vehicles. But for many of these machines, the key metric is not the amount of cargo it can lift, the range of the vehicle or its top speed: it is how quickly the developers can iterate and adapt.

The ongoing conflict in Ukraine has become a real-world laboratory for a new, horrific world of drone and robotic warfare. Both sides have leveraged commercial off-the-shelf components and open-source technologies to quickly develop and deploy new drone capabilities.

Take, for example, the evolution of first person view (FPV) drone aircraft. Early models were essentially remote-controlled toy helicopters. But the integration of advanced sensors and AI has rapidly transformed them into versatile tools. FPV are not just better reconnaissance tools; they’re a new class of weapon entirely, capable of dropping hand grenades on unsuspecting enemies as precision strikes at a fraction of the cost of traditional munitions.

With low-cost thermal imaging cameras, it is now easier for drones to spot enemies at night. And we have seen drones with AI-enabled Identification Friend or Foe technology that automatically recognizes enemy and friendly uniforms.

And these newly developed drone technologies might find other uses during and after the conflict.

Euromaidan Press reported on the ST-1, “an autonomous mine-detecting drone that scans 1 hectare in 4 hours—4x faster than humans.”  This sounds a lot like Broward County, Florida’s mosquito-fighting drone, just applied to a different pest.

But new battlefield technologies are not limited to drones. People are putting the pieces together, quickly, and in radically new ways using cameras, 3D printing, even microphone networks. Ukraine has deployed thousands of microphones paired with mobile phones to detect incoming drones. Dubbed “Sky Fortress,” the system was not only cheap to deploy, but has proven highly effective at its task.

Are you looking for trouble?

The rapid convergence of technologies in robotics and AI presents enormous opportunities for innovation. Advancements in any underlying technology often catalyze progress in others: AI improves the design and control of robots, while better sensors and actuators enable more capable AI systems. 3D printing (of not just plastic pieces, but now also complex metal parts, like rocket motors) allows for rapid prototyping and customization, accelerating the entire development cycle.

This interconnected nature of technological progress, driven by self-reinforcing feedback loops, means that advancements are likely to occur at an increasingly rapid pace. The key is to find problems—or “trouble”—that these new capabilities can solve. Are you looking for trouble?

The convergence of AI, advanced sensors, lithium-ion batteries, actuators and 3D printing is unlocking new possibilities in robotics and drones. From the battlefield to the dentist’s chair, these innovations are not just incremental improvements but fundamentally new approaches to solving problems. By looking for “trouble,” innovators can drive the next wave of disruption. As these technologies continue to evolve and interact, we stand on the brink of a new era of robotic capabilities, with implications that will radically reshape industries, warfare and daily life. 

What does the disruption of labor mean for you?

Understand the speed of change. Drive the next wave of disruption.