Artificial Intelligence has been in my head and heart more than 50-years!

I've been following artificial intelligence since before most of today's AI tools existed.

My graduate-level AI course at NC State was in 1973. My BS in Computer Science came the same year. My MSEE in Digital Systems followed in 1980. I've watched this field develop from punch cards to neural networks to large language models running locally on a Mac Studio sitting on a desk in Santa Fe.

What that long view gives me isn't nostalgia — it's pattern recognition. I've seen enough technology cycles to know the difference between a tool that changes everything and one that changes the pitch deck. The current wave of practical, local, privacy-first AI is the real thing. I've been waiting for it for fifty years.

Technology is always the Driver…

Yep, I’m among the guilty who said “Real Soon Now” back in 1973. We honestly thought we knew how, when, why. We didn’t.

The missing pieces in 1973 were staggering in hindsight. First, we needed the raw infrastructure: a global high-speed internet to make the world’s data accessible, coupled with an exponential drop in storage costs so it could actually be kept. The real bottleneck, though, was compute. Early AI ran on machines that would be outclassed by a modern thermostat (my degree was mostly on IBM360 which occupied a 10,000 ft2 building); it took the accidental discovery that video game graphics chips (GPUs) were perfect for the massive parallel matrix math neural networks require. Without GPUs—and later, wafer-scale chips and specialized AI accelerators—training on internet-scale data would have taken centuries, not days. So, yes, we thank you GAMERS for demanding advanced 3D chips to make those games fun - made initially by NVIDIA - started in 1993.

Just as critical were the conceptual shifts in how we program intelligence. The field had to abandon the elegant, top-down symbolic logic it favored in my CS511 days and embrace the messy bottom-up approach: statistical machine learning where systems learn their own rules from raw data. This required new algorithmic breakthroughs—backpropagation, transformers, reinforcement learning from human feedback—but also a philosophical change, treating AI as an engineering discipline built on pattern recognition at scale rather than a quest for a concise set of logical axioms. The 50+ year delay wasn’t just waiting for faster chips; it was waiting for AI experts to realize that intelligence is more about learning from petabytes of noisy data than about symbol manipulation in neat, human-readable code.

What is REALLY coming next?

Over the next decade, AI and robotics will revolutionize healthcare by shifting from reactive treatments to highly personalized, predictive care utilizing integrated biological data and microscopic precision. This technological leap will extend into education, where universally accessible AI tutors and immersive environments will transform learning into a continuous, embodied experience woven into daily life. At the macro level, AI's unparalleled modeling capabilities will enable precise climate interventions and force a shift from geopolitical conflict to mandatory global cooperation by making the true, interconnected costs of aggression transparently obvious. Simultaneously, autonomous robotics and optimized energy systems will drive the marginal costs of essential human needs—like housing, food, and power—toward near zero. By liberating humanity from menial labor and artificial scarcity, this era could unlock unprecedented creative and relational potential, provided we possess the collective wisdom to distribute these gains equitably.

Want to keep up with my day-to-day following of AI and the technology that enables it?
Zip over to: AI-All in the main menu