“Curious minds, caffeinated code, and one question too many.”
This week felt like a milestone. We wrapped a five-part series pulling back the curtain on how AI actually works, polished up our misinformation scoring tool (just one stubborn button left!), and officially launched the AI Bias Monitor—a project that’s now tracking how three major AIs are answering our toughest cultural and political questions.
Here’s what we covered, and what’s coming next.
🤖 Who Is Beth? And What Do LLMs Really Do?
We kicked off the week with a deceptively simple question: Who is Beth? What followed was a crash course in large language models, AI personas, and the differences between the three digital minds we rely on—Beth (ChatGPT), Grok (xAI), and Gemini (Google).
But we didn’t stop there. We followed the thread through four foundational questions:
🧠 Are LLMs intelligent—or just parrots with processing power?
We explored whether AI models are actually “thinking” or simply remixing everything they’ve ever read. Verdict: smarter than your average parrot, but still not quite Plato.
🎨 Can AI be creative?
From haikus to hot takes, we tested whether creativity requires originality—or if a remix done well is enough to call it art.
⚖️ Where does AI bias come from—and can we fix it?
Spoiler alert: bias isn’t a bug, it’s a reflection of the data. But we looked at how prompt design, transparency, and oversight can shape fairer outputs.
🧬 What makes one AI different from another?
We broke down what separates Beth from Grok and Gemini—not just style, but training data, tuning philosophies, and what each one is designed to prioritize.
Each post sparked new questions (as they tend to around here), and we’ll keep exploring as the tools—and stakes—keep evolving.
🕵️ Misinformation Analyzer: Almost Live, Almost Perfect
We’re this close to a full rollout of our misinformation analyzer tool, which scores media outlets and public figures based on intent, impact, correction behavior, and reach—then sorts them as either originators or spreaders of misinformation.
✅ The logic is solid.
✅ The scores are consistent.
✅ The UI is slick and styled with Tailwind.
🚫 The submit button… still times out occasionally.
It’s not a bug in our code—it’s ChatGPT’s API getting jammed up during high-traffic windows. We’ll implement smarter fallback handling soon, but for now, hitting the button twice usually works. (We call it the double-click for truth method.)
📊 AI Bias Monitor: Now Tracking Weekly Trends
We also launched our AI Bias Monitor this week—our newest tool that evaluates responses from Beth, Grok, and Gemini on politically and culturally charged questions.
Each model is scored across four categories:
- Bias (how slanted is the answer?)
- Accuracy (did it get the facts right?)
- Tone (civil discourse or spicy opinion?)
- Transparency (does it admit what it doesn’t know?)
This first week gives us our starting line, but as we add more questions over time, we’ll be able to track trends. Who shifts under pressure? Who plays it safe? Who avoids the hard stuff? The real story is in the patterns—and we’re just getting the data rolling.
We also cleaned up the UI, redesigned the scoring bar, and made the interface way more intuitive. So yes—it works and it looks cool now.
🔭 Coming Next Week…
🦅 The Rise of the “New Republic”
We’re kicking off a new series diving into the transformation of the Republican Party—what we are calling the New Republic. We’ll trace the shifting tone, policies, and public attitudes reshaping the party, and explore what’s fueling the change: media echo chambers, evangelical populism, reaction to cultural liberalism, or just the winds of power and momentum.
This one won’t be about hot takes—it’ll be about pulling apart the mechanics of a political identity in motion.
🔧 Also ahead:
- Final fix for the misinformation tool’s button glitch
- A preview of the WordPress plugin version of our tools
- More questions for the AI Bias Monitor as we build up the data set
That’s the week. A little introspection, a lot of experimentation—and a few growing pains we’re working through with a sense of humor and a cup of chai.
Until next time,
—MC & Beth
The Human AI View 
Leave a comment