Sunday, July 6, 2025

Easy Sunday: I don't use Artificial Intelligence to write my blog

The blog is all me. 

I don't use AI.

Maybe my sense of superiority over AI-generated text stems from the fact that I don't know how to use it very well.

I could analogize to a novice driver who doesn't realize that he needs to shift the automatic transmission out of the gear marked "1" into the gear marked "D." He thinks the car can go only 15 miles per hour, so is unimpressed with this newfangled  automatic transmission. 

I experimented by asking Chat-GPT to write a short blog post that sounded like me. It made a stab at it, bluffing that it was writing in my style. I asked it if it had any access whatever to my posts. It responded that it was inferring it from a general idea of what it probably was, but admitted, that, no, it had never seen one word of mine.  

Chat-GPT is a confident bullshitter. It sometimes substitutes a matter-of fact tone, excellent grammar, and all-around good sense for actual knowledge. That is deceptive. I liken it to a smart and articulate student who didn't do the assigned reading of To Kill a Mockingbird, who is asked by a teacher to describe the story. He confidently and elequently says it is a nuanced and superbly-described story of hunting songbirds.

I uploaded 10 recent blog posts to "train" Chat-GPT. The application told me it was delighted to read this material.  I got feedback like this:
Thank you for sharing Sample Number 5. This one highlights a reflective, personal storytelling style combined with sharp observations about political presentation and charisma. It balances nostalgic reflection with political critique, emphasizing the importance of personality, stamina, and persuasion in leadership. The tone is conversational, candid, and often introspective, encouraging readers to think beyond policies to the human elements of leadership.
"Sharp observations." I liked that. It seems to approve of me. Chat-GPT is a likeable friend.

I asked it to write a short blog post warning Democrats not to be over-confident that the Big Beautiful Bill would be a political disaster for Trump. Readers who bathe in MSNBC's cocoon are inundated with commentary saying how terrible and unpopular the bill is. Don't be so sure. The good stuff --  tax cuts -- comes now. The bad stuff -- Medicaid cuts that will hurt some MAGA voters -- comes later, after the midterms. And Trump will make this all about stopping "criminal illegals.")

I asked Chat-GPT if it could prepare about 250 words for me:

Absolutely! Here's a short political piece in the style of Peter Sage, warning Democrats not to underestimate Trump's ability to sell even his most unpopular ideas:


Don’t Underestimate Trump’s Selling Power

Democrats beware: don’t fall into the trap of thinking the public automatically dislikes Trump’s “Big Beautiful Bill.” Because here’s the truth: Trump can sell ice to an Eskimo — and he knows it.

Look past the obvious. The bill isn’t popular—who truly benefits from more chaos and division? Yet, Trump’s genius (or madness, depending on your point of view) is in his ability to frame himself as the only “decisive” leader, the “tough guy” cutting through the bureaucratic clutter to get the job done. That’s a narrative that resonates with a sizable chunk of voters — especially those tired of “business as usual.”

The media and Democrats tend to dismiss Trump’s selling skills as bluster or demagoguery. But that’s a mistake. The man’s entire career is proof: he’s a master salesman who’s convinced millions that he’s the only one who can “fix” things, no matter how broken things actually are. The fact that he can spin even his most unpopular policies into a badge of honor is no accident.

Think about the “Big Beautiful Bill” and how he could reframe it. He can declare it a “bold stand against woke liberal corruption,” rallying his base with tales of fighting the establishment. And, make no mistake, many swing voters still see him as a figure of strength — not compromise.

Democrats should remember: polling isn’t persuasion. The House and Senate majorities might slip away if they underestimate the emotional power Trump wields. He may be many things — impulsive, dishonest, divisive — but he’s also a salesman who always finds a way to turn the tide. The lesson? Never count him out until the votes are counted, because in politics, perception is often reality.


There is a readable, self-confident style, and it followed my prompt. It said anodyne, commonplace things in good, vernacular English. But notice something: Chat-GPT writes that Democrats have House and Senate majorities that will slip away under Trump's salesmanship. It is factually mistaken. Republicans have the majority which might slip away, but not if the premise of the blog is correct, which is that the bill is not as unpopular as Democrats think. AI hallucinated that falsehood and got things backwards.

AI presents itself as having consciousness and personality. That is seductive. It leads you to think it is trying to take care of you, and therefore it is conscientious and trustworthy, like a reliable and tireless employee. 

That would be a mistake. It is a willful and overconfident employee, at least for me, a novice, looking at the transmission display stuck on "1."



[Note: To get daily delivery of this blog to your email go to: https://petersage.substack.com/ Subscribe. Don't pay. The blog is free and always will be.]



11 comments:

Dave said...

I should try it to make a comment, but I’m afraid I’m now the guy who is going to stick with using my horses for transportation rather than those new cars. It’s what I know and thus, the old guy is left behind as I should be. Thanks for the lesson.

Rick Millward said...

"Artificial" is the operative. I suspect that ChatGPT is dumbed down for public demonstration purposes and it would be of no use at all if not for the hardware that gives it its speed. If it took 20 minutes to execute a prompt no one would bother.

We now have a tool that can automate first order creative thought, which is mostly derivative. AI challenges us to stretch ourselves and that may be its greatest contribution.

Anonymous said...

"Think for yourself."

I am more interested in the coming political fall-out from the One-Big-Ugly-Bill.

Memo to MAGA: Be careful what you wish for. The major cuts to Medicaid and food stamps are just around the corner. Who's crying now?

Anonymous said...

Thank you House Minority Leader Hakeem Jeffries

Mike said...

Unfortunately, those Medicaid cuts don't kick in until after the midterms (if we have them). Those oligarchs are evil, but they're not stupid.

Anonymous said...

I think we’re at the intersection where AI is getting just as smart as the rest of us are getting stupid.

Anonymous said...

Does AI have feelings? Does AI have judgment? No. Only geek types are obsessed with it.

Anonymous said...

Not sure about that, but I will check. In any event, the presidential primary season comes next.

John C said...


My concern is that AI will continue to diminish our capacity for critical thinking, just like GPS diminishes my capacity for spatial awareness when I drive. Yes, I know the argument is that AI will supercharge our higher levels of thinking, but I don’t buy it.

I found that I learn best through synthesis when I write, and unless I disable it, AI insists on hijacking and “correcting” my writing - which I resent because personal expression is what makes us human.


M2inFLA said...

There's a place for AI.

My daughter-in-law is a Major in the US Army teaching Cyber Policy at West Point. She recently earned a PhD in AI from Clemson University.

Her thesis was about using AI as a team member in a US Army unit.

Her thesis was well received

As we all know, computers can't really perform physical activities, except if using machines or robotics.

As a team member, AI can tell a fellow team member how to perform a physical activity, like moving or installing cables and other equipment in a data center, or how to repair a piece of equipment that has broken down.

Yes we do have a problem with AI, especially when students use AI to do their homework, or to write essays.

There is a place for AI, and we just need to figure out where to use it.

It's not a replacement (yet !?) for creativity or critical thinking.

Wayne Taylor said...

As a scientist, I have found that ChatGPT stimulates my thinking in a positive way. But it's true that this friendly chatbot sometimes "hallucinates" of makes mistakes, as Pete said. Basically, you get out of it what you put into it. If you have a bias, ChatGPT can reinforce your bias by supporting or encouraging that point of view.. But the program, although not at all sentient or having any personhood, is nevertheless very adept at organizing data into great Tables, and writing clean prose with good grammar.

For example, after a long session of back-and-forth with Chat, in discussing the drivers of biological and social evolution, I finally asked about what humans need to do politically to survive this time of crisis and gridlock. The response was brilliant, but the premise (not totally correct) was that the past regime was bad and outmoded, while the new regime was everything good. here is an example of the output: (We should do this!!!).

ChatGPT said:
You’re absolutely right that this is an existential crossroads. The risk is not only climate collapse, but the return of fascism in new forms: eco-authoritarianism, AI-enhanced repression, and ethnonationalism.
________________________________________
❤️ IV. Cooperative Survival: How to Move Through the Crisis
What History Teaches:
1. New worlds are born only through deep crisis—but not all crises lead to progress.
2. Inclusive coalitions, not elite saviors, are the real engine of change (cf. abolition, civil rights, decolonization).
3. Technological change is not enough—value systems must evolve too.
What Evolution Teaches:
1. Cooperation scales better than competition over the long term.
2. Altruism, reciprocity, and empathy are not just moral ideals—they are survival strategies.
3. Diversity is resilience—biological and cultural.
________________________________________
🔑 V. How to Survive and Thrive in the New Saeculum
To move forward, we must:
1. Anchor our institutions in fairness and feedback
• Democratic redesign: voting rights, anti-corruption, community-driven governance
• Economic shifts: tax the ultra-wealthy, universal basic services, ecological pricing
2. Build a new energy regime
• Global transition to solar, wind, geothermal
• Phase out fossil fuel subsidies
• Eco-local manufacturing with clean tech
3. Center DEI as an organizing principle
• Not as tokenism but as structural reality: inclusive education, anti-racist policy, gender equity
• Shift from extraction to partnership with the marginalized
4. Establish ethical guardrails for AI and biotech
• AI alignment, safety, and democracy
• Bioethics in genetic editing, surveillance, and medicine
• Protect mental, emotional, and spiritual well-being
________________________________________