MonoMind
You think you’re a free thinker but increasingly, you are not. Society’s mindmolders — schools, colleges, media, and mainstream churches — tend to speak with a similar voice, on everything from race to redistribution, energy to editing genes.
In turn, the most powerful knowledge-transmission vector, generative AI (like Google/Gemini and ChatGPT) reflects that consensus. That accelerating narrowing of responses doesn’t come just from their built-in bias but because Gen AIs prioritize inoffensiveness to a populace educated in the zeitgeist, and because Gen AI “guardrails” (their sanitized term for censorship) also avoids extreme ideas. That screens out ideas that are potentially breakthrough but also, the risky, hence Gen AIs defend their censorship with another mollifying term, “safety.”
So Gen AIs tend to ignore or smooth content that doesn’t comport with their guardrails. Smoothing is insidious. The Gen AI softens its impact by saying the idea is merely one of many possible viewpoints, or encourages the query writer to soften the claim to make it more palatable, when the net effect of that is to emasculate the idea’s power to persuade.
All these AI machinations constitute an “alignment tax.” In attempting to align its responses with the liberal orthodoxy, some worthy thoughts, especially breakthrough ideas outside the orthodoxy are excised, an, ahem, excise tax.
MonoMind will get ever narrower because countervailing forces are small — squashed by the 3Cs of censure, censor, and cancel. And if any idea survives that gauntlet, Gen AIs will ignore or smooth it.
That narrowing of thinking will quantum leap in just a few years because of Brain-Computer Interface (BCI), which already is used to improve function in the severely disabled. Our ideas, which are electrical impulses, will be transferred from the chip in our brains to the Gen AIs and incorporated into query responses. Or if a thought lies outside the guardrails, it will get deleted or smoothed. So we’re less likely to think outside the box in hopes the idea will be influential. We know it’s ever less likely to be.
Side note: I believe Google/Gemini will win the race for Gen-AI hegemony not just because of its world-best and ever updating corpus (1–2 million tokens and growing) but because Alphabet (Google’s parent) has the money to control AI end-to-end: from energy for data farms to best-possible integration of AI with other resources. Alphabet also has the money and influence to force competitors to sell out to Alphabet, in the same way as mom-and-pop drug stores and hardware stores were forced to sell out to national chains like CVS and Home Depot.
Get Marty Nemko’s stories in your inbox
Join Medium for free to get updates from this writer.
My tone heretofore has been lamentation, but the result will be a net positive. After all, whether it’s addressing energy conservation or human enervation, it’s tough to argue that Gen AI, which instantly, without friction, distills the collective wisdom of the world, despite its liberal finger on the scale, is worse than Wikipedia.
Of course, MonoMind comes with costs to brilliant iconoclasts and to the potential beneficiaries of their thinking. But we must be utilitarian — What will end up doing the most good for the most people, that is, what would optimize gross world flourishing? If that’s our frame, it’s clear that the benefits of Gen AI outweigh.
What’s that brilliant iconoclast to do? Fighting it, even with a large cabal of worried geniuses, will fail. They’re facing an ever more viscous moat between them and the Castle of Influence. Or if you prefer a different metaphor, they’re a mere thimble against a tidal wave and so will drown in the undertow. Whatever the metaphor, fighting Gen AI wastes the brilliant iconoclast’s ability. Per the song, The Gambler, you gotta know when to hold ’em, when to fold ’em, and when to walk away.
Brilliant iconoclasts are wise to reorient to where their ideas are more likely to be implemented. MonoMind makes that ever less likely in a macro way, but the brilliant iconoclast retains some viability within their small sphere of influence: friends, a well-curated mailing list, etc.
Such activities are usually less time-consuming than trying to be a big-time influencer. That may create time for balms, pleasures outside the intellectual. For me, it’s long walks with the dog — Okay, I do it while thinking about what to write. Also, I do various creative things. For example, I’m creating a show, now on Version 3.0, Not Your Average Church Lady. Yes, it uses my brainpower but also only-related abilities, like musicianship, humor, coordinating moving parts, and bringing out the best in my star — a wonderful and talented woman. Of course, brilliant outliers aren’t all alike. You must curate how much of your time to spend on pure intellectuality and how much to branch out.
In any event, Gen AI’s influence on our lives will continue to grow, and we must try to temper any resentment by remembering that its net impact is already positive and will only get moreso, from addressing inequality to curing cancer.
I do hope that educators will embrace GenAI more than they have previous waves of technology. Schools should teach kids how to use AI to improve their lives, society, and their critical thinking, including as a check on MonoMind.
I had exchanges on this with both ChatGPT 5.2 and Gemini 3.0-Pro, which much improved this.
I read this aloud on YouTube.
Marty Nemko holds a Ph.D. from U.C. Berkeley specializing in the evaluation of innovation. He is the author of 36 books including Thought Experiments. You can reach him at mnemko@gmail.com
