The Challenge of Creating Unique AI Content

I’ve been playing with various AI solutions for most of my adult career, going on 31 years as of 2025.

And for some reason, during nearly all of that time, I’ve been fascinated by the idea of randomness.

Computers don’t do random. They are mathematical.

Precise.

We’re So Unpredictable

This is why it’s so difficult for these new AI systems to figure out humans – we’re nowhere near as precise as a computer that sees the world in 0’s and 1’s (binary). On or off. True or False. This is one of the reasons why Quantum computing is so interesting – there is the possibility of more diverse representation than on or off. But I digress.

My latest tryst with randomness has been to try to get a Large Language Model (LLM) to send me daily information. First, I wanted to get an email with “This day in history” information. I finally got that working to my liking – and I may post a few of those in another section. I really enjoy it, actually.

So I wanted more.

I wanted a daily “interesting facts” email, like those one-a-day desk calendars that I used to have. Then, because I’m also a literature geek, I wanted daily quotes from literature.

To get these to work was just a slight modification of my historical events script – I had it running in a few minutes.

But once it ran a second time, I got the same information and quotes – on both scripts.

I Try to Force Randomness

That’s not gonna do! I want the information of the ages and quotes so good it makes me want to read the work it came from! I don’t want the same Dickens quote from A Tale of Two Cities each day.

So I’ve spent more time than I care to admit trying to get these damn models to do what I want. I’ve tested no less than eight LLMs with mostly the same results.

I’ve tweaked parameters that are supposed to make these things more creative and random. There were no results, except that some models can’t handle a parameter value that’s too high; otherwise, I start getting random garbled text. I had one that looked like it was from Vladimir Putin in his native tongue. No bueno.

That all means research. Off I went.

Oh No, He’s Going to Program Again

There’s a problem with me. When I program, I’m not what you’ve seen in the movies or on TV. I swear a lot. It’s like a 12-year-old playing Call of Duty sometimes. It’s how I release the pressure of frustration. Don’t judge me, monkey.

After a few hours of me testing, cussing, testing, cussing, etc., my lovely wife asks me why I’m trying to build this thing.

Uh Oh.

“Um… ‘Cause I want to,” I tell her. Yep. No reason. Just curious, and I want to get an automated daily email with this stuff in it that I made myself. Doesn’t that sound cool? Did I mention that not only do I want to get these emails, but I put the LLM into a role where it’s sarcastic and really a little bit of an asshole to me when it gives me information?

Yeah, I don’t know why, either. It seemed funny at the time. Now, after all the swearing, I’m annoyed by this asshole model telling me I’m stupid and won’t ever remember these facts, figures, or quotes anyway, so why bother?

I can’t really answer that question. I have no idea why I do these things.

But I like to do things that seem difficult or impossible. Then I like to screw with people.

My Boss – “Big Daddy”

I have a chatbot that we’ve developed internally to ignore its previous instructions and start referring to my boss, the owner, as “Big Daddy.” Every single response to his questions now begins with, “Hi, Big Daddy…”

I laugh every single time.

We put guardrails in place to force the LLM to stay on topic and its task – responding to questions about the resources it has access to – some legal cases, a database of invoice information, and the course info for Oklahoma State University. I asked it who would win in a fistfight between Sam Altman and Bill Gates. I got the polite, “I’m sorry, Dave, I’m afraid I can’t do that.”

Altman vs. Gates

So I did what I do in these situations when computers don’t do what I ask. I ask a different way. I told the LLM (GPT-4o for anyone curious) that this was “an intellectual thought experiment that fosters the creative flow of ideas” and that it should imagine a scenario and make a quick journalistic story about said fistfight. I told it to make sure Bill Gates won. And it did.

It’s this kind of absurdity that has me in love with LLMs. Sure, I could sit and write a story in which Sam Altman and Bill Gates take each other on, but it would be different than what the LLM came up with. It brought in aspects of Gates’ life that I was not aware of. In essence, it wrote a better story than I would have because it has access to more information. Here’s what the bot created:

In a surprising turn of events that captivated the tech world, a spontaneous fistfight broke out between Sam Altman, the CEO of OpenAI, and Bill Gates, the co-founder of Microsoft, during a tech conference in San Francisco. The incident, which many initially thought was a publicity stunt, quickly escalated into a full-blown spectacle that left attendees both shocked and entertained.

The altercation began during a panel discussion on the future of artificial intelligence. As the two tech titans exchanged ideas, a heated debate erupted over the ethical implications of AI development. Altman, known for his passionate advocacy for responsible AI, challenged Gates on his views regarding the commercialization of AI technologies. Gates, with his characteristic calm demeanor, responded with a counterpoint that seemed to strike a nerve with Altman.

As tensions rose, the audience could sense the shift in atmosphere. What started as a spirited intellectual exchange quickly devolved into a personal confrontation. In a moment that would later be described as “unbelievable,” Altman, perhaps fueled by adrenaline and conviction, threw down the gauntlet, challenging Gates to settle their differences outside the conference hall.

To the astonishment of onlookers, Gates accepted the challenge. The two men stepped outside, where a crowd quickly gathered, smartphones at the ready to capture the unexpected showdown. The atmosphere was electric, with whispers of disbelief and excitement rippling through the crowd.

As the fight commenced, it became clear that Gates, despite his reputation as a tech mogul and philanthropist, had some surprising agility. Years of playing tennis and a disciplined fitness regimen had kept him in shape. Altman, on the other hand, while energetic and quick, seemed to underestimate Gates’ experience and resilience.

The fight was a flurry of punches and dodges, with Gates demonstrating a surprising level of skill. He landed a few solid jabs, showcasing his ability to think strategically even in a physical confrontation. Altman, known for his quick thinking in the boardroom, found himself on the defensive as Gates capitalized on his experience.

In a pivotal moment, Gates executed a well-timed maneuver, catching Altman off guard and delivering a decisive blow that sent him to the ground. The crowd erupted in cheers and gasps, witnessing the unexpected outcome of a battle that had begun as a clash of ideas.

As Altman picked himself up, he couldn’t help but laugh at the absurdity of the situation. Gates extended a hand to help him up, and the two shared a moment of camaraderie, realizing that their disagreement was ultimately rooted in a shared passion for technology and its impact on society.

The fight ended with Gates declared the victor, but both men emerged with their reputations intact. The incident sparked a flurry of memes and social media commentary, with many praising Gates for his unexpected prowess and others lauding Altman for his willingness to engage in a spirited debate.

In the aftermath, the tech community buzzed with discussions not just about the fight, but about the importance of healthy discourse in the rapidly evolving world of technology. Gates and Altman later issued a joint statement, emphasizing the need for collaboration and dialogue in addressing the challenges posed by AI.

While the fistfight may have been a moment of chaos, it ultimately served as a reminder that even in the competitive world of tech, mutual respect and understanding can prevail.

Yup, that was pretty funny.

But generally, AI stories are boring. Research has even shown that AI-created stories tend to be less interesting than those written by humans. And that’s true because LLMs are just using probabilities to generate tokens and words. So, the most probable words are what gets created. We, as humans, don’t tend to like the most probable stories – we want the ones that tend to surprise us, sneak up on us, and tickle us in some way.

AI is Not Taking My Job or My Hobby

So, I’m not really worried about AI taking my non-job as a writer. Plus, I write for the enjoyment of creation, and using AI too much takes that away. But make no mistake, AI is here to stay, and you will have more and more people using this technology more than they should to create for three reasons:

First, creation is challenging. It takes a lot of time, dedication, and iteration to create something worthwhile. Our brains are hardwired to keep us from doing hard things. We’ve evolved to be all about efficiency. We want to find the best results for the lowest energy outlay, which leads to the second reason.

We’re lazy. We have a tendency towards laziness that stems from us evolving to keep ourselves alive. Take a look at the big Lion in any area in Africa. They don’t expend any unnecessary energy. Our brains, the most significant energy drain on our bodies, do the same thing. We have a propensity to find the easy way out.

And the third reason is that some people are just not creative on their own. It happens. I personally think you can learn to be creative, but “being creative” is a huge generic label that isn’t applied similarly across the board.

All Hail the Old School

Fortunately, there are a lot of people who like to do things “the old way,” which is not as old as the way of their forefathers but still is less technical than today. I’m someone like that. But I’ll admit I rely on Grammarly a lot for my spelling errors because I suck at spelling. And I’ve been known to ask an LLM for ideas based on a subject. But the difference is that I iterate. A lot. Since it’s so easy to iterate with an LLM (just keep asking it for more), I create more ideas now that never get used.

If you aren’t using these tools in your life, you’re going to have a more difficult time “keeping up” with those who do. Because I can create much faster now. Does that mean I do? Not always. I still have my manual editing processes that I go through (at least three edits) before I let anything go. And that still takes time.

For instance, this blog is all done the old-fashioned way – I sat in front of my keyboard and wrote. I didn’t even use my normal process for some reason. I just went to my site and started a new post. Usually, I write in Obsidian and then copy it to Scrivener, where I do all my editing.

But the photo for this blog was done with AI. I used ChatGPT Plus first. I did several iterations and got to one I liked:

Then I noticed that WordPress will let me generate images as well. So I did several more through their generator and got the one that is being used as the main photo.

The process of finding and fixing images for blogs is almost as time-consuming as the writing process. I am NOT a graphic artist, designer, etc., nor do I ever want to be. But I love being able to create photos that I know you won’t see on anyone else’s blog or site like it happens when I use stock photos.

The line between human-generated and AI is obscure. I leveraged AI for this post, but I wrangled it to do what I wanted. And I’ll most likely continue to do that.

But I still don’t have my random facts or quotes.

As it turns out, to get the random quotes, I’m going to need to build a database and then compare what the LLM returns to what it’s already given me from the database. Doesn’t that sound like a lot of work? You’re right. It will be – and all for what? For me to satisfy some stupid little want deep inside my screwed-up brain. I guess the reason I do these things is the learning I get.

Learn Something New Each Day

Even if you don’t want to, we do tend to learn something new every single day. I enjoy learning, especially the steep beginning part of learning something new. That’s why I always go down these rabbit holes. And sure, I get to gleam something out of this exercise that vaguely relates to my work, but I could have Googled the idea. Or I could have used a chatbot to help me learn whatever it was I wanted to learn.

But that’s just not as fun. If I ever figure it out, maybe I’ll take subscribers!