🧠 Are We Letting AI Do Our Thinking?

In today’s email:

🧠 Is AI making us smarter or just really good at outsourcing our own brains? Here's how we might be turning into cognitive couch potatoes.

🤥 New AI models are hallucinating more—and we may just have to live with our confident, fact-challenged chatbot friends.

🎓 College in the AI age is getting weird—think AI-graded essays and digital cheating—but does higher education still matter?

Curious? Keep scrolling!

🥔How AI Is Turning Us Into Cognitive Couch Potatoes

AI is kinda making us dumb-dumb. Not in an “oops, I forgot my keys” way, but in a “why bother thinking when Skynet can summarize it?” way. Rich Haridy (tech bro turned techno-skeptic) wrote a love letter to the good ol’ days when you had to actually read stuff, not just vibe-check a ChatGPT summary.

Some key takeaways:

• Using AI for everything = cognitive couch potato

• Creativity needs time, boredom, and a little wandering

• LLMs (like ChatGPT) don’t "think" — they remix existing info like a DJ on autopilot

• Offloading too much? You risk forgetting how to actually think

Also, there’s a Black Mirror episode in which your brain implant charges you monthly. Yes, Netflix said capitalism, but make it neurological.

Even Microsoft’s CEO is AI-pilled. He skips podcasts and lets an AI give him SparkNotes during his morning commute.

The moral of the story is to use AI, but don’t outsource your entire frontal lobe. Otherwise, we’ll all be sending AI-generated emails to people who have AI summarizing them back into bullet points.

🤦‍♂️ AI Keeps Hallucinating and We Might Be Stuck With It

New AI models from OpenAI, Google, and pals are... getting worse at telling the truth. AI “hallucinations” (aka confidently made-up nonsense) are actually increasing in newer models that were supposed to be smarter.

These AI bots are like overly confident friends who think they know stuff, but are often just winging it. Even OpenAI’s newer models, like o3 and o4-mini, hallucinate significantly more than their older sibling, o1. So much for aging like fine wine.

🔍 A quick cheat sheet:

• Hallucination = when AI spouts false info, or gives an answer that technically sounds right but isn’t helpful.

• o1 hallucinated 16% of the time.

• o3? 33%.

• o4-mini? A wild 48%.

• “Reasoning models” didn’t help—they might actually be worse.

Experts say hallucinations aren’t just bugs; they’re baked into how AI works. So we might be stuck with glitchy, overly enthusiastic robo-know-it-alls for a while.

Bottom line? Always double-check your chatbot’s "facts." Or better yet, don’t ask your AI to summarize your legal case. Unless you like imaginary court drama.

🎓 Should We Still Bother Going to College?

College life has never been easier... or weirder. Professors are grading AI-written essays with AI. Students are cheating with bots. And parents are paying thousands for what’s basically a high-tech staring contest between humans and machines.

But wait—does college even matter anymore?

It turns out, yes, it still does. While AI can summarize Plato in two seconds flat, it can’t give you wisdom or turn you into a functioning adult who can argue politely at dinner.

Here’s the tea:

• AI in class? Yes, but use it with guidance, not instead of thinking.

• Cheating’s too easy? Time to bring back the ol’ pencil-and-paper tests.

• Professors stressed? Yes, but they need to adapt, not just grade bots with bots.

The point of college isn’t just job skills. It’s about learning to think critically, forming opinions, and maybe even reading Aristotle without sobbing. Schools need clearer AI policies, real consequences for misuse, and creative ways to make students actually learn.

Other cool AI stuff that is trending right now 🔥🔥

⚖️ Trump's latest “big, beautiful” AI bill could bulldoze state laws and trigger a full-blown constitutional cage match. - Read more

👀 AI just figured out how to eavesdrop on what your eyes see and ears hear—without any human help—and it’s about to make robots way smarter (and maybe sneakier) in the real world. - Read more

💃 Tired of juggling logins? OpenAI might soon let you waltz into your favorite apps just by being a ChatGPT user. - Read more

💼 AI might not have taken your job yet—but if you're a new grad eyeing tech, it may have already swiped your seat before you even got to the interview. - Read more

😏 Why show up to your own earnings call when your AI clone can do it for you? — Tech CEOs are literally outsourcing themselves now, and it's as weird as it sounds. - Read more