Like electronic scientific calculators, personal computers, and CAD systems – generative AI is just another new tool in our toolbox.
I am an ancient engineer. Or, perhaps a better term would be a Methuselan engineer. “Ancient,” while not completely inaccurate, might suggest something set in stone, unchanging, while “Methuselan” conveys the sense of a long journey of adaptation and evolution. It implies continuing survival and endurance shaped by the march of technology, and always endeavoring for continuous advancement.
When I began my engineering studies in college, computations were a manual ordeal. Addition and subtraction were done by hand, multiplication and division by slide rule, and trigonometric functions and logarithms came from lookup tables. Then, during the summer between my freshman and sophomore years, the Hewlett-Packard HP-35 electronic scientific calculator became an affordable reality. A year or two later, Texas Instruments introduced a competing model, and just like that, the engineering profession ditched the slide rule and hand calculations. The electronic calculator brought speed, precision, and fewer errors – and it freed us from carrying cumbersome paper handbooks filled with trig and log tables.
The productivity boost was undeniable, but it came at a cost. We lost some of the mental math fluency and problem-solving tricks that once defined our craft. Was that entirely bad? What we gained in efficiency, we indeed traded in and gave up some of our other capacities. At the time, our professors cautioned us not to blindly trust the 10-decimal-place precision our calculators provided. With slide rules, we worked with two or three significant digits, forcing us to think thoroughly about the magnitudes of our calculations.
From those days in the 1970s, the evolution of technology has been staggering. We’ve moved from mainframe computers to minicomputers, from microcomputers to powerful devices that fit in our pockets. The progression from Intel’s 8-bit processors to Nvidia’s graphics chips capable of super-speed advanced computations is astounding. Today, we hold in our hands smartphones more powerful than IBM’s house-sized computers of the 1970s, and potential access to the entire Earth’s store of knowledge through the internet and web-based applications.
Artificial intelligence has been part of this evolution for some time. Early advancements focused on pattern, image, and voice recognition, but what is new and driving the evolution today, is that AI has become “generative.” Where once computer software could be taught to recognize the image of a cat, now it can generate one – even of a cat with green and pink fur – if we so desire. Hence the name of one of the more popularized systems – ChatGPT, where GPT stands for “generative pre-trained transformer.” This shift brings immense potential but also there have been concerns about “hallucinations,” the sometimes unintentional creation of misleading or fabricated information. It’s a cautionary note, similar to that of not blindly trusting the 10-decimal-place precision of our electronic calculators.
My first experience with generative AI, specifically with ChatGPT, was reminiscent of the first time I held the HP-35 calculator in my hand. There was a sense of awe – a recognition of newfound power and possibility, coupled with the thrill of imagining what might be achieved. Productivity, creativity, and quality seemed poised for another leap forward. But then, how to implement generative AI’s fullest capabilities and what comes next?
This brings us to an award-winning science fiction novel from 1992: China Mountain Zhang by Maureen McHugh. Its protagonist, Rafael Zhang, is a design professional – a “construction engineer,” as she describes him, though the term is slightly off. In one passage, Zhang describes the design of a simple beach house, assisted by an advanced “system” as follows: “I envision a huge expanse of windows… I expand, the system becomes my own memory. I fall through. I feel my mind’s boundaries... the system is there for me, a part of me. To modify the house, I only have to think it and it is so. It hangs there. I am outside it, seeing the long portion of the house that is the kitchen and the great room, off the kitchen the steps down to the beach… the bedrooms are beyond the kitchen, higher to take advantage of the uneven terrain… and I think that this Western building needs a tile roof. Blue Chinese tile. Soften the variation in the roof height, and the roof becomes a wave.”
While this passage still remains a flight of science fiction fancy, history has shown that yesterday’s science fiction often becomes the science fact of today and tomorrow. I’ve seen the progressions of information technology time and again in my Methuselan journey.
And this, I believe, is the long view of generative AI: Like the electronic scientific calculator, the advent of personal computers, Apple’s graphical user interface replacing IBM’s DOS, and CAD systems replacing hand drafting – generative AI is just another new tool in our toolbox, albeit of a much greater promise and to be rolled out with some caution. But there is a present and soon-to-be future where AI systems can merge seamlessly with our creativity, further enhancing – not replacing – our abilities as engineers, architects, designers, and AEC project and firm managers.
Sam Liao , Ph.D., MBA, PE is principal consultant at Strategics, LLC, and adjunct faculty member at the University of New Hampshire. Contact him at sam.s.c.liao@gmail.com.
