Sharpen Your Critical Thinking Skills for 2026: Part 3
Let’s flash back to the years 1999/2000. I’m working in a private bank, watching the dot-com businesses suck up lots of money. I personally have been on the ‘net since I graduated college in ‘95 (yeah I’m old), primarily on the bulletin boards which functioned as chatboards. Using the keyboard to type emojis, listening to the robots scream when you use your dial-up modem.
I’m noticing that some of these businesses are stupid. Like 1800sockpuppets.com. Child, there is no way that business is going to be profitable, because the laws of supply and demand STILL APPLY. And I knew it. Back then. When I was in my 20s and still didn’t know that much about the world.
I kept telling the private bankers and stockbrokers to divest out of tech because it’s a bubble. I’m getting annoyed that the portfolios are still heavily tech-driven.
I complain about this to a stockbroker, who tells me that he wants to divest but the client won’t let him. Another broker phones me up to talk directly to his client about why he needed to diversify his portfolio because said client isn’t listening to the broker.
The larger community is all about these dot com businesses. They’re being celebrated in the WSJ and elsewhere. It’s clear to me that many of these people don’t understand the internet itself and actually think that there’s some new kind of economics that the internet has brought.
Remember what happened? BIG crash, tech stocks nosedive, and many of them bleed out.
Now let’s move on to 2007/8. I’m working at a fixed income software company (which will lay me off in November of ‘08.) I’ve saved up some money and am thinking about buying a house, since the company seems pretty happy with me.
The problem is I live in West LA and housing prices are pretty high. But I could easily afford something in the Inland Empire (interior Southern California where the deserts are.)
Do I pull the trigger? Nope. (Fortunately, since I do get laid off.) Why not? Because SoCal housing is so clearly in a bubble and I can’t do it. I know that we might not be at the top of the bubble, but I still can’t bring myself to invest in the bubble.
Greenspan, the Fed chair, is out here saying there’s no bubble and then there’s little ol’ me saying there is one. (In fairness, he did acknowledge that some housing markets had bubbles.)
I think you remember what happened after that.
POP!
So here we are in the middle of AI hucksterism. It’s the greatest, neatest, and bestest thing to ever happen. All the big tech companies are going all in and of course with their billionaire owners they know things you could never even begin to know, so why don’t you start shelling out money for it now, whydoncha? All the business journos seem pretty happy with it.
Let’s think critically about AI and whether it helps you lead a more fulfilling life
Can I tell you how absolutely fucking tired I am of having AI shoved in my face? I only use LinkedIn for business, but there are ads for AI, people throwing AI in their headlines, people touting their AI prompts, etc.
Do you really want to spend time learning how to prompt robots? I mean, not to kinkshame, but is this what you want to do with your life?
I do want to note here that here in the US we think of AI primarily in terms of large language models or LLMs, and similar image generators. So I’ll be talking about LLMs/generative AI specifically.
Screw all those poets, authors, painters, sculptors, and other artists who tried to capture beauty and truth and other ideals, man. Let’s just cram all that into a computer and hurl whatever shit comes out the other end in front of people’s faces.
(However, other countries such as China focus on AI in terms of robotics and machines instead, leaving humans to their humanity.)
AI is not intelligence, just as social media is not social
AI can’t think. I want to impress this upon everyone, because a lot of AI boosters talk like it can think and act on its own, which it can’t. Do not anthropomorphize the machines, unless you want to be fed into them.
You have autocorrect on your phone (and in the Google doc I’m using to hammer this piece out.) When you type, a suggestion will appear because AI is programmed to discover and replicate patterns. As I typed “multiple”, the suggestion “times” appeared, and I accepted it because that’s what I want.
That’s right, your Gemini, your ChatGPT, your Nano Banana, your Claude, your whatever is basically a souped-up autocorrect. That’s all. It doesn’t think for itself. It has had thousands, more likely millions of datapoints crammed into it and with some probabilistic programming it returns a pattern.
But sometimes it gets the pattern wrong: hallucination. If I wanted a different word after “multiple”, the machine would get it wrong.
The makers of these LLMs have all said that there is no way to prevent hallucinations; in other words, you should expect your AI to give you the wrong answers sometimes. Which makes sense when you remember that it’s pattern recognition. No matter how many levels of training or system prompts the AI is given, or which new model it happens to be, it will never be correct all the time.
When you use one of these LLMs, you’re destroying the environment
Yes, you personally. Because these data centers that house the computers for AI take up a lot of energy and water.
On a planet where water is becoming more scarce everyday. (I personally think we’re going to reach a point pretty soon where desalination will suddenly become profitable because there’s so little fresh water.)
By the way, I have looked up how you can not get that f*king annoying AI overview in Google Search – so if you’re going to destroy the environment you can do it on purpose.
There are two ways (I use Firefox browser): you can either click on the More dropdown (or the three dots if that’s what you see) under the search bar and select Web, or you can just add “-ai” to the last word of your search. Either works.
Guess what? The economics of AI currently don’t make any sense
Now, I have been following a Brit by the name of Ed Zitron, who’s done a deep dive into the financial statements of NVidia (which provides the chips and currently is holding up the US stock market), AI gen companies like OpenAI and Anthropic, and assorted peripheral companies.
In order for the companies to do what they told investors they were going to do, they need data centers to house the computers with the Nvidia chips.
(They’ve all bought pieces of each other, by the way, and you might recall something similar with the mortgages and credit default swaps back in the Recession days. JUST SAYING.)
You know what AI needs massive amounts of, besides water? Power. In order to provide all this power, some of the data centers need power substations. Which, unlike computer models, can’t be some minimally viable product. There’s a lot of connections and equipment involved.
So the fact that none of these substations have been built yet and will almost certainly not be built in time to satisfy AI financial projections… is a tad problematic.
Not to mention that local customers, moms and pops and their businesses, living near these data centers and proposed data centers, need power too. And who do you think is going to get the power if it comes down to it: corporate-owned data centers or you? Hint: not you.
According to surveys, AI hasn’t been a big driver of productivity or efficiency at the offices that have enthusiastically signed on. So although the hype is big, the results aren’t.
How much AI should you be using in your business?
Aside from the economics, maybe AI can take over some social media posts for you. Of course, you’ll need to spend time on the prompt or buy someone’s list of prompts.
After it’s done what you want it to do, you then have to go back and review it to make sure that it didn't say something wrong. (Particularly if you’re a lawyer, because you will get in a lot of trouble if you’re citing cases that don’t exist to a judge or even a jury.)
I will admit that the AI image generators, while still turning out recognizable slop, are getting better. You still see arms coming out of ears and feet with ten toes each, though. But it’s slop. Slop on LI written posts, too is getting pretty recognizable.
There’s already a lot of slop, and unless the bubble pops soon, which I don’t expect, there’s going to be more.
But do customers want your AI-generated posts and articles and thought leadership? Nope.
They want authenticity. They want to hear a personal voice. Almost every service and product now is a commodity: there are hundreds of CPAs in any given metro area if not thousands, and the same with lawyers. Business consultants and coaches are a dime a dozen.
Do you think AI that has flattened and smashed writing will help you stand out? Do you think customers who read the yawningly boring articles and posts that come out of AI will want to work with you?
Or will they want to work with someone they can actually know and learn about? My money’s on this one. Of course, I could be wrong… but almost everyone I talk to is yearning for more human connection.
Recap (tl;dr):
While we’re on the topic of thinking critically… ask yourself if you’re using generative AI because there’s nothing else that can help you do what it is you’re doing, or have you succumbed to the hype?
I’m here to guide you through setting yourself up for productivity success in 2026, and you get to deal with a human. Opposable thumbs and all! Click here to set up your free exploratory consultation call.
Photo by Mathieu Turle on Unsplash.