The Problem(s) with AI
Many years ago I worked as a computer specialist in a Radio Shack store. For some unknown reason, they expected me to sell computers! There was an interesting problem. We’d get a new catalog with new machines that I’d have loved to sell, but we wouldn’t get them into our store for months. By that time, we’d have another catalog with newer computers. Only occasionally was I able to sell a fairly current machine.
The Pace of Government
It may not be obvious, but this leads me to recall a military proverb that goes something like this: “Generals always prepare to fight the last war.” It’s not just generals. It’s all of us.
I could extend this proverb to this: Governments are always fighting over solutions to past problems.
Mentioning “government” reminds me of a call I got in the mid-80s, while I was in the Air Force. It was from an FBI agent. She wasn’t investigating me. She was investigating fraud conducted through computer Bulletin Board Systems (BBS), and one of the perpetrators had calls to my BBS on his log. They wanted to know if I had been defrauded.
I recognized the name, but had found his attempts to break into my system quite laughable. As I recall, I had logs of him trying to sign in using words in my introductory paragraph on the login screen as passwords. None of those were my password, of course.
But the memorable thing was the end of the conversation. Having asked all the written questions she’d been told to ask, she wondered if I had a few more minutes. “Sure,” I said, all cooperation. “What is a Computer BBS?” she asked.
She was investigating fraud that was carried out on a system she couldn’t even identify!
Now in her defense, she seemed quite intelligent, and this was before PCs were all that common. She also had simply been assigned to ask a series of prewritten questions. But she nonetheless illustrates the government often being behind in tracking technology.
We thought technology was moving forward quickly back when I was working at Radio Shack. People couldn’t keep up with the pace of technology.
We were wrong. Technology was moving very, very slowly back then. “Slowly,” of course, being a relative term.
Regulating AI
Now we see calls for various regulations on AI development. The question is, what should be done?
My experience with government and with bureaucracies would suggest that the government is unlikely to be all that successful in making AI development any safer. The problem is that the people who know enough to understand the necessary controls are likely to be working in AI development and not in some government office planning how to control the process.
At the same time, AI does present some serious challenges and does have the potential to do great harm. Most new inventions have that potential, especially if people are not careful with them, or use them without understanding what is going on.
There’s a great story, almost certainly apocryphal, that the first two automobiles in Ohio somehow found one another on hundreds of miles of dirt roads and collided. Everything has its hazards. Technology has hazards. This one probably originated in at ad in the 1960s, but I suspect we could invent one similar for the first rock used as a hammer.
So should we try to do something about AI? If so, what?
What we might try is not unnecessarily encouraging it.
Let me note here that I use AI. I’ll allow AI to proofread this post before I post it. Once it has done that, I’ll check carefully to make sure that it didn’t get too intense and change my style of writing. I use AI to generate marketing text because it honestly writes hype way better than I can. Then I always tone the hype down. But it’s easier for me to tone it down that to generate it at the desired level myself.
What I mean by encouragement here is specifically community or government action that makes it cost less for companies to build AI. This may seem like an odd thing to say.
Capitalism and Consequences
Let’s take a look at how capitalism works. Someone has capital (that’s money and other valuable stuff) and they want to invest it in something. That something can be a new invention, or a new way of production of something older, or perhaps filling an unfilled niche in the market. The capitalist, who has those resources, invests them in producing the stuff (of whatever variety) and then tries to sell the stuff to others. If they’re successful, they earn money. If not, not so much.
Now let’s look at the one thing people are always bringing up about socialism. It’s not really the defining characteristic, but it’s a strong human tendency: The desire to get something you haven’t earned.
We usually hear this in stories of social welfare, how lazy people are collecting large quantities of “our” money in order to spend it on unnecessary stuff, and besides, it was our money, so why are they getting it? We are always unhappy that someone else gets money that they didn’t earn.
We’re generally a great deal happier if we get a windfall, something we didn’t earn. In that case, good fortune has smiled on us, the universe has rewarded our diligence and skill with some extra stuff that was not the result of all of that.
Businessmen are just as subject to the desire to get things they haven’t earned as is any poor person. They just do it at a much larger scale. They especially like to get this in the form of special treatment from the government. It’s often not best for them from a PR point of view to try to get grants, though those do happen from time to time. More commonly they get this in the form of some sort of tax relief. Build your building in our city and we’ll forgive some portion (or even all) of your property taxes for a certain number of years.
Even better camouflage is provided by getting regulatory relief. Joe Local who is trying to start a business with a few 10s of thousands of dollars has to meet a set of standards and the time taken to get permitting and so forth, but an expedited process can be provided for the larger business because of all the employment he’s going to provide, and of course taxes that can be collected on those new employees.
There have been studies done showing that these projects don’t produce those benefits, such as those done by the Brookings Institute, at least not in the amount projected , but that’s not the point I’m trying to argue here. I’m back to the function of capitalism.
One of the great things that capitalism does is that it punishes foolish investment. People who don’t know what they’re doing tend to lose money they invest. Good investors know how to decide on good investment. If you don’t have the government there to fix it for you, this tends to weed out the field so that those with wise investment practices make money and those that don’t, don’t!
When you start offering tax and regulatory incentives, what the government is doing is providing unearned income to the investors in a project. You can restate that as the government (local, or federal) that invests is taking part of the risk, and making the decisions on the value of that investment. When the government takes part of the risk, they also alleviate the risk for the investors.
And thus they blunt one of the great counter-incentives of capitalism: The penalty for making unwise choices with your money.
Of Capitalism and AI
How does this apply to AI?
I’m glad you asked. AI is an invention. There are things it can do well and there are problems with that. I sincerely doubt that the government can or will effectively regulate it. I think they may try. I think they might even save us from some problems. But the government is very bad at predicting the future. They’re going to use the solutions of the past to solve problems that are cast in terms of the past. They’ll miss some of the most important issues.
But what we can do in our capitalist country is enforce some capitalism. Don’t invest as government entities. Make those who are developing these program pay their actual cost. That includes all the costs of covering the externalities (the impacts on the neighborhood and environment) of things like data centers.
Some might call this anti-business. I’d rather say it is pro good business and anti bad business. It means that you are asking capitalists to take their own risks with their own resources. This might slow things down, and if it does, that is all to the good. It leaves open the invention and potential, but does not drive it faster than it should.
I would suggest that the only government investment should be for the needs in terms of the new technology for military and security purposes.
In addition to slowing things down, it is very likely that better ways of creating data centers might well be invented, such as the possibility of putting them under water or in space (each of these suggests negatives to me, but they can be explored), due to the cost of dealing with the external impact of the projects in our communities.
And in Conclusion
Capitalism is very much about investment. It can be corrupted or destroyed as easily by removing the penalties of bad decisions as by throttling the collection and investment in the first place. Let’s be very capitalist about AI investment.
After Concluding … for More Info
In this little book, Elgin Hushbeck discusses labels such as socialist, communist, and fascist. If you’re offended by my use of socialism and capitalism, or think I’m off-base, it might be a good place to start investigating.
I don’t have a cover image, but a great deal of what I say about bureaucracy I learned from Ludwig von Mises in his delightful (to me!) little book with the really creative title Bureaucracy. (That link is an affiliate link to Amazon.com, representing a little bit of time I invested in the hope of future returns!)
Here are two videos on AI by Elgin, author of the book above:
And here is a full hour conversation, in which we do discuss AI briefly, but we discuss political labels even more.
(The featured image was generated by Adobe Firefly using the Firefly 4 Ultra engine, based on a prompt generated by Gemini. The post excerpt was generated by Gemini. The entire post was checked and edited by Gemini, but all final choices were mine.)
