September 21, 2026
It’s Time To Stop the Negative Blind Bashing of AI
The harshest criticism of AI is now coming from the people with the least understanding of what it is, and that’s a bad sign

By Joe Procopio
4 min read
I knew this was going to happen.
Back in 2023, the biggest cheerleaders for AI seemed to be the people who understood it the least. The ones pushing for AGI, the ones talking about AI replacing humans in every imaginable capacity, including personal relationships, and the ones wanting to get in on the early dumb money like they saw happen with NFTs and meme stonks.
How'd that work out? Those chatbot burritos produce any billionaires yet?
Now in the back half of 2026, it's almost the opposite. I'm seeing outright bans on generative AI in the workplace, the wholesale buying-in to the AI-took-your-job layoff excuse, and hordes of Rivian-driving, iPhone-clutching data center protesters.
You know how much AI is in that R1 there, champ? It's not a little.
Look. I'm not naive, and I'm not defending AI. I just wrote three posts in a month, declaring the generative AI age to be over, pinpointing exactly how the slinging of AI slop is going to come back to bite people, and recommending the amount of trust you should be putting into an LLM (Spoiler Alert: Zero).
I'm with you, naysayers, but let's just make sure we're pushing back for the right reasons.
You Can Still Hate AI
In a vacuum, there's no reason to hate AI.
The way it was rolled out? Visceral rage.
On the heels of the death of crypto, AI looked like the next get-rich-quick scheme. But a quick peek under the hood proved it was not a con, then it became a get-rich-for-real scheme, but for Big Corporate. Soon, anyone with an ounce of technical know-how was looking at the latest advancements in LLMs and being like, "Oh, hey cool, they figured out a way to put some UX around the models and WAIT WHY IS THAT GUY MARRYING A CHATBOT AND NOW EVERYONE IS FIRED?!"
Like that.
But then it settled. This isn't my first rodeo. I remember Flooz. Tech settles. It's settling.
So I ask you remaining naysayers, do you really want to die on FORTRAN Hill? And maybe you do, but don't complain when that vibe coder rolls out 30 apps worth of junk while you're still manually building relational database schemas by hand.
I get it.
Overkill Is Built In
If I want to add 2 + 2, I use a calculator. But I don't anymore. I use a spreadsheet, because if I do it there I'm not locked into the calculator ecosystem and I can do a shitload more calculations, add words, and send it to someone. Well, now we're using LLMs to add 2 + 2 and there are two problems.
One was that LLMs weren't very good at basic addition, because their packagers started on the natural language UI end, but that's been fixed to a certain extent and will continue to improve. The second is a problem of overkill. In a philosophical sense, too many people are using LLMs to do the equivalent of adding 2 + 2. And their "4" gets buried under 1000 words of slop.
This will also get fixed.
When I Say Generative AI Is Over, I'm Not Kidding
Or being hyperbolic. I'm serious.
The history of resultant set output went from binary lights to numbers to charts to words. That's a natural progression. We're not universally pushing back against the words. We're pushing back against the context for which AI-generated words are a misguided choice, and we're pushing back against the UI "demo mode" nature of every insight being buried under an avalanche of words to "impress," which it no longer does, and to "hide" its own factual flaws, which too many people have been caught regurgitating.
This is going away, because humans want knowledge, not words, and also the mistake of trying to use AI-generated words in human writing context is becoming ever more clear.
Look, we were making fun of all the flashing lights 50 years ago in Airplane! Or maybe Airplane II. I know Shatner did it in Airplane II.
So at some point in the near future, bashing AI becomes screaming at a spreadsheet for doing too much math. The more we use generative AI out of context, the more the model-makers give us more words to go with our words, the more it will get rejected. I literally watched clients learn this in the early 2010s. I have no doubt it's going to happen again in the late 2020s.
But AI Is Here To Stay
Like calculators. Like spreadsheets. Like mobile phones. The utility is too rich.
It's the processing power. Separate the quagmire of misguided generative AI implementations from the processing power that makes it work. Think of generative AI as it rightly is, a "demo mode" of all that processing power being used in the absolute wrong context most of the time, and you'll see the value underneath it.
Fight for that value, but fight against the flaws by being clear about the flaws.
Look, mobile phones are a great thing. I couldn't live without mine. But I choose not to doomscroll. I sleep in a room separate from where I charge my phone. And I'm still shocked when I go out at the sheer number of people living their lives in this world with their face buried in their phone. Eating meals together, crossing the street, driving.
I'm shocked and saddened by this, but I'm not calling for the end of mobile phones. Because having a computer in my pocket is much better than not having it. I feel like that's true for everyone. I'm not "calling for the end" of doomscrolling because you don't change human nature by mandate. You change it by providing and championing an alternative.
Now, that processing power isn't going to suddenly revert back to 1990s levels any time soon. And we're at a point where we get to decide the eventual outcome of how that processing power is used.
Use it wisely.
Doomscrollers are gonna doomscroll. Sloppers are gonna slop. Terrible CEOs are going to fire thousands of people and blame it on a technology.
I keep saying "AI is spreadsheets" and I keep getting dragged for saying it, but from a processing-power-to-utility perspective, I think I'm dying on the right hill. Stop condemning and blaming the spreadsheet. Take the time to learn what this processing power can do and use it for the right reasons.
I'm going to keep trying to separate truth from fiction with this AI stuff. Now would be a good time to join my rebel alliance email list, for a heads up on my posts.