The AI Revolution is fueled by equal amounts of hype and fear. (I asked Gemini to illustrate ChatGPT controlling your job and this is what it came up with.)
They have the same root, the idea that computers will take away your job. The hypesters say this like it’s a good thing. The rest of us fear they’re right, and there will be nothing for us to do.
I think we can cut through the bullshit by describing what AI programs are. They’re trained databases. AI uses collections of data, hopefully curated, with pointers inserted so they have ready answers to the questions users will ask.
Microsoft, Google, ChatGPT and the rest claim to use the whole Internet as their database. While this is impressive, it’s misleading and proving impractical. Large Language Models don’t scale their intelligence as they get bigger. It’s not like Moore’s Law. The fourth version isn’t 16 times better than the first one. Vendors are starting to figure this out. All those monkeys aren’t writing Shakespeare.
The action in 2025 will be in training smaller databases. Some will contain personal data, so they can go into the virtual world and perform tasks for you. Others will also contain personal data but will be designed to impersonate you. Both of these programs are called agents. They let some of what you know and want to be available into the market all the time.
Where’s the Excitement?
The databases I’m most excited about, or at least interested in, are those which include machine data. They are designed to model a hospital, a school, a battlefield, or a political campaign. Every company with assets, both physical and human, can build and use trained databases to improve their operations beyond what their bureaucracies might be able to manage on their own.
Think of an oil refinery, having all the data on pressures needed to prevent an accident. Imagine a hospital with all the data needed to make sure your operation will go well. Think of a city being able to manage its traffic lights and construction projects, or a police agency managing both cameras and officers. Back in 2019 I referred to this as the Machine Internet, but there are people involved in every scaled system, and they’re in the database too.
AI moves us from the Internet of Things to the Internet of Systems. AI can do what no person, or group of people, were able to do as effectively as before.
Then there are AIs that improve the productivity of the creative class. They don’t necessarily replace people, but they make people more productive. They improve your writing, they do the first draft of your computer programs, they create visualizations around whatever you can imagine. We assume this will mean they don’t need imagination. But this is nonsense. No database program can go beyond the data inside it or the programming that runs it. It will be, by its nature, derivative. It’s a box. It can’t think outside itself. You’re not a box. You can.
Replacing People
As I was writing this, I saw hundreds of jobs that have disappeared. Bank tellers, travel agents, fast food and retail cashiers, insurance agents, and inventory managers to name just a few. What happened to those people? Some are working at warehouses or driving orders to customers. Some are visiting customers at their places of business. Unemployment remains low, and most Americans are making money.
The American economy is doing much better than the economies of our trading partners, who aren’t as automated. We have lower costs of doing business thanks to AI technology. We’re more efficient. We capture that value and the dollar is strong. This gives us something we can sell, at incredible markups, that will still provide value to the buyers.
It’s fear that’s our problem. Fear of AI, fear of our adversaries or of job loss, fear of Trump or of a world without him. We really do have nothing to fear in the world of AI but fear itself.
The trouble is that fear is a very powerful thing, and it we just focus on what the hypesters are telling us, we will succumb to fear.
Train your data and face the salesmen who are scaring you. People remain in charge until we choose to give up that control.
