The end of Manufacturing as we know it, how AI-powered computer vision is going to change everything

The robots in our factories are stuck in the stone age, but computer vision will change that.

In the self-driving space, we see what truly advanced AI-powered computer vision can do to an industry. Jumping in my friend's Tesla and driving through the downtown busy streets completely hands-free was an experience like no other. Their implementation of FSD was the first innovation in the automotive space since the invention of the seatbelt. This is just the first industry to leverage this kind of technology, the next big one will be manufacturing. 

Robotics and computer vision are nothing new to manufacturing, some robotics use stretches back to the 1960s with General Motors first implementing a robotic arm for moving die-casting parts. Computer vision has been around since the early 1980s with the most rapid development coming in the past few years. The current technology in the space is clunky, expensive, and time-consuming. Companies spend hundreds of man-hours and millions of dollars to implement then robotics and automation projects. Robotic systems often struggle with the complexity and unpredictability of real-world environments. Tasks that are simple for humans can be incredibly challenging for robots, especially those involving fine motor skills or decision-making in dynamic settings. Most robots are designed for a single specific task and lack the flexibility to change when needed. Programing robotics is difficult and the skills are highly specialized across brands and industries. To put it simply they’re dumb. The introduction of LLMs, semantic segmentation, and multimodality can revolutionize how things are made as we know them. 

A robotic arm or even a humanoid robot will be able to understand the world around it and adapt accordingly. You don’t have to spend countless hours programming it to know what a certain type of bolt is, present it in the correct orientation, and then have it slowly reach down the specified 12 inches before movie its arm 55 inches and down to screw it in. Now you’re able to tell it to pick up the bolt and screw it in the part. It uses visual data to identify objects and tactile feedback to gauge the force needed. Safety is one of the most important concerns across the industry. No company wants to be responsible for a worker's compensation lawsuit. There are two current solutions to this problem, either cage it off or use a cobot. The best way to deal with this problem is for the robot to be able to see that it’s a human and not hit it. Simplifying the implementation of robotic or automation projects is crucial for the advancement of the industry. Having robot-human interaction take place from a natural language standpoint will dramatically save time and ease of use. This will make them far more accessible and easier to control and repurpose. 

This technology will play a key role in the proxy war that is the reshoring of manufacturing back to America from China. In recent years the United States has seen the economic vulnerability of depending on foreign countries for producing goods. Reshoring manufacturing only works from an economic standpoint with the use of AI in robotics because of the low labor cost overseas. With the tensions rising around the globe and China’s eyes fixed on Tawain, now more than ever the United States needs to make it a priority to be self-sufficient in every way possible. Semi-conductors have been a topic in recent years and for good reason. They’re a vulnerability that America cannot leave unattended for much longer. U.S. Congress passed the CHIPS and Science Act, which allocated approximately $52 billion for the development of semiconductor manufacturing facilities in the United States. Leveraging the new technology we have at hand will be the only way America can compete internationally. 

There is a myth that robots powered by AI will come along and steal 100% of American jobs in the manufacturing sector. This is fluffed up by AI doomers. One of the main problems in the manufacturing industry is the lack of labour which is a leading cause of outsourcing to foreign nations. Americans are unwilling to work dull, dirty, and dangerous jobs in factories. Who’s to blame them when your job is stacking boxes on a palletizer all day or feeding parts into a lathe? This is where the strengths of robotics excel. Running 24 hours a day without stop. A humanoid robot has the flexibility to go from one job to the other based on demand with minimal setup time. Humans will still be required for many aspects of the business. Their roles will adapt to the needs of the times. In the 19th century, 70% of Americans worked in agriculture as of spring 2023 that number is as low as 1.5%. Of course, there is the chance for the darkest timeline and we end up in a world similar to Kurt Vonnegut’s Player Piano. A dystopian near future where robots have taken all the jobs except for a select few engineers and managers to oversee production. But I’m more optimistic about the future than that. 

Just like any other technological disruption, legacy industry players will not be the ones to spearhead change. This is an opportunity for a startup to go zero-to-one with an AI vision-led approach to robotics. Which is easier said than done. Large amounts of capital are needed and the R&D process is long a difficult. Another contender is large tech moving from the world of bytes into the world of atoms. Tech colossuses like NVIDIA, Meta, Tesla, Microsoft, and Alphabet see this potential and are making plays to be the industry leader in it. On December 13th, 2023 Tesla revealed Optimus Gen 2 leaping forward in not just the world of humanoid robots but an AI-integrated approach to manufacturing. Its technical capabilities are astounding already with the ability to handle precise tasks, see the world around it, and be highly repeatable and accurate. “I was surprised that people do not realize the magnitude of the Optimus robot program,” Musk said in regards to the launch. “Those who are insightful or who listen carefully will understand that Optimus ultimately will be worth more than the car business and worth more than full self-driving. That’s my firm belief.” NVIDIA is leading the way when it comes to digitizing existing factories. Their Isaac Robotics Platform helps the development of robots with AI capabilities. It's a comprehensive set that stretches from software to hardware. They’re looking to take advantage of the recent boom from the rise of GPUs needed for LLMs, not just being a supplier of hardware in the industry but a key player in it as well. 

It seems now that we are off to the races in the robotic AI space. Whether it's a new startup or tech giant we all know the name of, the world of atoms is about to change. This inflection point is crucial for the future of America as a leader in the manufacturing world. As the Visigoths are at our gates we need legislation to propel this technology, not stop it. Every decade there is a new technology that comes around which changes economies slightly. Veering them off in different directions. Most change some aspects of our economies but this is going to flip it on its head. This is the first one since the Industrial Revolution that has the chance to have the same impact.