-->

Register Now to SAVE BIG & Join Us for Enterprise AI World 2025, November 19-20, in Washington, DC

Devices Big and Small Can Learn What We Need Them to Learn

Article Featured Image

A few months ago, I noticed a non-trivial flaw in the setup of my home garage in Colorado. My driveway is tilted downward at approximately 10%. It is a lovely driveway despite settling ground underneath cracking the concrete about a dozen times, but nonetheless, it is a beautiful addition to my home. The problem is that the Colorado winters bring snow and ice, and my house sits within the umbrella of a neighbor’s house, producing a long shadow. Therefore, the driveway retains a significant sheet of thin ice over an appreciable number of winter and spring days.

This arrangement has resulted in several orthopedists substantially increasing their bank account balances at my expense. Despite a lifetime of some degree of athletic prowess, I fall on the ice a lot. To alleviate this problem, I acquired several powerful, serialized electric mats to melt the ever-present ice. This solution worked great! Except for one issue: The mats blow away constantly. The wind in my quaint little gated community in Denver’s Front Range experiences frequent Derecho-level winds, and despite the mats’ substantial weight, the winds, using maneuvers of classical physics, pick them up, unplug them, and toss them around the neighborhood. This forces me to later retrieve them from the adjacent streets, usually unbroken.

To solve this problem in a low-tech manner, I purchased approximately 20 heavy bricks from the local Big Box hardware store. That solution worked great! Except for one issue: Now, the automatic braking system in my 2018 High Altitude Jeep Grand Cherokee activates when detecting the bricks as I pull out of my garage. Oh yes, the lights and alarms work perfectly, constantly warning me of impending doom as the vehicle backs up into the 3-inch-high bricks. My next solution is to simply, using all the possible calm I can muster, step on the gas, overruling the automatic braking system. I make use of the superb powers of my all-terrain, four-wheel drive super vehicle by driving right over those bricks. I do that about a dozen times each week.

In early March, I attended the NVIDIA GTC conference at the San Jose Convention Center. I successfully obtained several interviews to be used in a special episode of The Don Sullivan Show podcast. The specific episode, “Don’s Excellent Adventure at NVIDIAGTC 2025,” includes interviews of the many different vendors and technologies being presented.

I was amazed at the positive atmosphere that has been absent from so many tech events this decade. That atmosphere was enhanced by the many robots present inside the pathways of the show floor. These “creatures” were roaming the floor. While displaying impressive awareness combined with seemingly creative consideration, the robots would say hello to the conference patrons or even converse and interact with the crowd.

The robot dogs were truly astounding. Rolling over, begging for a treat, panting, and doing most of what “real” dogs would do, they all became fixtures of the futuristic show floor. But wait, the future is now. Because it wasn’t only creature robots that were being demonstrated, it was every type of automated and self-directed system that any vendor present wanted the sophisticated audience of this amazing show to witness. Now I understand why GTC is referred to as the Super-Bowl of AI conferences. I’ve been to this event many times, but this event displayed a different and superior level of sophistication.

Then as I turned a corner, I saw the solution to my driveway problem. Not exactly a pile of $3 bricks, but much cleverer: an automobile that could learn. Not a self-driving coupé but an actual car that was meant for human beings to drive that included a learning capability component. Of course, the systems engineers tasked with delivering demos referred to the capabilities I observed as “artificial intelligence,” but regular readers of my columns and listeners to my podcast know that the tangible technology is the ever-enhancing capabilities of “machine learning.” This vehicle could perceive patterns that occurred within the normal scope of the common driving experience and learn how to respond appropriately or, might I speculate, even effectively.

The example used was a bit more significant than the issue with my home’s icy driveway, although I don’t mean to downplay that perilous situation. The illustration offered included a street walkway that was constantly covered in debris of any size. The claim was made that the automatic braking systems present in earlier vehicles would often see the walkway and misinterpret the debris to constitute a person, animal, or some other obstruction.

However, if the vehicle experienced that same set of circumstances with sufficient frequency, it would “learn” and be able to simply circumvent the faux obstacle. This is groundbreaking stuff, pun intended. I’m not sure I’m ready to trust the 2026 version of my present vehicle to the point where I’m going to commit $75k to upgrade yet. Soon, however, I’m quite confident that a new version of my now severely outdated Jeep will, after a few encounters, learn that the bricks holding down my electric mats are no match for the power of the Jeep and simply ignore them and roll right over them.

Since in my extensive shopping, I haven’t seen any 30-lb electric mats that can overcome the Front Range winds, this may present an elegant answer to my innate physical awkwardness. Although maybe I should pay a contractor to embed electric wires in the driveway itself. Or maybe I should just shovel the driveway sufficiently every time it snows so the ice is never a problem. I’ll have to think about it.

EAIWorld Cover
Free
for qualified subscribers
Subscribe Now Current Issue Past Issues