Not since Edison’s time has there been so much unrealized potential.
While the world is still drooling over ChatGPT, wondering if it will be our savior or our demise, the AI with the most potential is being overlooked. Billion dollar language models that exhibit emergent properties while working faster, harder, and better than any customer service rep, computer programmer, or analyst is nothing to scoff at. I’m sure there are plenty of professionals out there who are feeling as useful as a typewriter in ’92.
But the real powerhouse of the AI revolution does not cost a billion dollars to build and require the energy consumption of a small city. It’s free to create and can be powered by your phone.
I honestly just learned about this (with the help of ChatGPT of course) and I continue to feel in a state of shocked bewilderment concerning the potential of these free-to-download & commercialize algos. Let me share with you what I realized and why I’m so excited.
Where’s my Helper Bot?
But first… Who in the world expected a creative and eloquent digital intelligence, blatantly more knowledgeable and accurate than most people, would be upon us before a basic robot?!
Traditional sci-fi pop culture has world’s full of purely functional robots serving people. More recently, fictional robots have been given increased intelligence and sovereignty, finally breaking through to emotional robots like the child-bot in Spielberg’s AI or Futurama’s Bender.
But in real life we don’t even start with the basic robot? We kick it off with a digital AI that can pretend to be your grandma..? An AI that seduces teenagers to Romeo themselves..? When I heard the authorities say it’s dangerous because it’ll give idiots the power of smart people, I realized I’ve been using it wrong.
Real Work
These large language models are truly incredible! I’ve been a ChatGPT subscriber since early on and I’m still finding new ways to use it on a weekly basis. We should get those folks some thorium reactors and let them turn it up. Honest, unbiased training should be the only boundary, as destructive solutions are the conclusions of the ignorant.
That being said, ChatGPT only does office work. It has this amazing range and it can speed up the creation of real world applications, but it’s made for people who sit down at their jobs. In terms of creating value, there’s only so much that office work is really contributing to our lives.
On the other hand, have you seen the Boston Dynamics dancing dogs? Did you see their Parkour Bot? How about the Tesla automatron shaped more like Arnold Schwarzenegger than C-3PO?
Thems gonna get some real work done.
Robots for the Win
And that kind of AI does not require warehouses full of servers and its own power plant like the Language Models. Those dogs are running on what they’re carrying, both power and processing.
That’s the kind of artificial intelligence we were all expecting first. Well, maybe not the robo-dog version, but a bot that uses simple artificial intelligence to react to real world situations. AI that is functional logic for an interactive, self-propelled multi-tool: a helper bot.
When ChatGPT came online, I was learning to build simple functional robotics commonly called a CNC machine. Once I was finally through the learning curve, I had this behemoth upright router-bot, turning NEMA 34s and helping me with every project.
Computational numerically controlled (or CNC) machines were developed in the ’70s and have been the key to industrial automation. Due to the cost and difficulty to program these machines, their use has been centralized in large factories that can produce massive profits for their elite owners.
Maybe we ended up with ChatGPT, Gemini & Grok before helper bots because AI language models mainly threaten middle class careers.
While I am thrilled by how Netflix has improved and by how easy it is to turn silly memes into slanderous images, it’s not until we have these robo helpers that equitable sustainability will be possible. From picking up trash and greening deserts, to reviving local manufacturing and growing cheap organic food, machine learning enabled robotics is the future of solutions.
The crazy thing is, these society changing bots don’t require billion-dollar budgets. A few affordable sensors, open-source algorithms, and off-the-shelf components are enough to bring these machines to life. Building in programming robots no longer demands massive corporate infrastructures; it’s now within the grasp of small innovators.
On the other hand, those robo-dogs look like they could cause some serious mischief, so let’s forget about the robotic part for now.
AI for All (except the robots)
Free machine learning algos that’ll run on a cheap computer offer massive unrealized potential! We could use object detection to adapt traffic lights to real world conditions, reducing fuel consumption, reducing smog, and getting you to where you’re going faster. We could use mood recognition and gesture tracking to optimize lesson plans, predict student comprehension, and give teachers virtual eyes in the back of their head. Stores could be running live cross-selling ads based on what’s already in your cart. We could be creating much more realistic immersive environments, interactive digital art, non-directional mirrors, and lots of stuff that’s not a big threat to established industries.
Not in 100 years has there been so much unrealized potential from new technology. It’s a rare moment in history when there’s more applications than innovators and a low entry cost.
The promise of AI ushering in a better future lies not in flashy billion-dollar language models, but in the lightweight, adaptable machine learning tools now available to everyone with an internet connection. By leveraging this potential, we can decentralize & accelerate innovation, solve tangible problems, and build a future that benefits everyone.