In 2015, the president of a small Catholic university in Maryland, Mount St. Mary's, wanted to improve what schools call the "retention rate," meaning the share of incoming students who stay. He had freshmen fill out a survey to find the ones who were struggling. Sweet, right? Find them so you can help them. Not quite. The plan, according to what professors said in the meeting and what was later published, was to nudge the most fragile students out in the first few weeks, before the date when they officially count as "enrolled." That way, the statistic would look gorgeous. When professors objected, the president told them they thought of the students as cuddly bunnies, and that you just had to drown the bunnies. His words. He resigned in February 2016.

Now, that president was not using artificial intelligence. But think about what happens when you hand somebody like that a tool that says "I predict who will fail." He does not have to say it in a faculty meeting, where somebody can leak it to the student paper. A screen does it, quietly, with a number, and nobody knows why. And the screen never resigns.

Today we are going to talk about that character. The machine that hands down a sentence without explaining, without an appeal, and without blushing. Five ways predictive AI fails. Five. With real cases, real dates, and a bit of good news at the end.

A number with no explanation is a sentence with no trial.

Every good episode needs a villain with a name and a face. Vague villains do not work. If I tell you "technology is bad," you shrug and go back to watching videos of cats falling off couches. So today's villain has a name: Sake Oil AI, the Digital Medicine Show.

Who is it? It is any person, company, or headline that uses the label "artificial intelligence" as a giant umbrella to cover things that have nothing to do with each other. Under that one umbrella you find wonderful things, like the autocorrect on your phone, and things that do not work and never will, like the software that claims to guess whether you will be a good employee by watching you talk for thirty seconds. Same umbrella. Same sales pitch. Very different reality.

In this first episode I am not going to scare you or sell you the apocalypse. I am going to give you something more useful: a dictionary. By the end, when somebody says "an AI did this," you will be able to squint like a building inspector and ask, "Which one?" That question, honestly, is worth gold.

And since this is a show for people who say "tech just isn't my thing," here is the good news. You will not need a single equation. You will need exactly what you need when you buy a used pickup truck: healthy distrust, good questions, and a firm decision not to fall in love with the salesman.

In nineteen twenty five, a man sold the Eiffel Tower. Twice. No laptop, no phone, not even a Wi-Fi password taped to a router.

And here is the part that should keep you up tonight: his descendants have your phone number.

Last year, Americans filed more than one million complaints with the FBI about crime on the internet. A million. First time in history. Nearly twenty-one billion dollars gone. And the most reported crime of all was not some genius cracking a government vault. It was phishing. Somebody pretending to be somebody else. That is it. That is the whole crime wave. A bunch of people doing bad improv.

Think about that. Nobody picked a lock. Nobody cut a wire. There were no green characters raining down a screen like in the Matrix movies. We opened the door, offered them a cold drink, asked if they wanted to see the basement, and then helped them carry out the TV, the silverware and the jewelry.

The thief of the future does not need a crowbar. He just needs you to be polite.

I am Dr. Qubit, and this is Behind the digital curtain, the only show where technology gets explained without making you feel like you skipped a class everyone else attended. And since it is October, my birthday month, which happens to be Cybersecurity Awareness Month in the United States, today we are talking about the oldest trick in the book with the most expensive name in the world: social engineering.

Today we're breaking down a documentary called "The AI Doc: Or How I Became an Apocaloptimist." It premiered at the Sundance Film Festival on January 27, 2026, hit U.S. theaters on March 27, and landed on Netflix in the U.S. on September 15. Runtime: an hour and forty-four minutes. I'm gonna give it to you shorter, funnier, and with a musical number. You're welcome.

Who made it? It's directed by Daniel Roher and Charlie Tyrell. Roher is the guy who made "Navalny," the documentary about the Russian opposition leader, which won the Oscar. And it's produced by Daniel Kwan and Jonathan Wang, the team behind "Everything Everywhere All at Once," the hot dog fingers movie that won Best Picture. So this isn't some dude in a basement with a ring light yelling about lizard people. These are serious filmmakers. With Oscars. And it made about 1.3 million dollars at the box office. A film about the most powerful technology in human history grossed what a Marvel movie makes during the Thursday preview trailers. That tells you everything about how much attention we're paying.

Today we are discussing artificial intelligence without using a single slide titled “Disruptive Synergy.” Every time someone says “disruptive synergy,” a printer somewhere decides it is offline.

Our guide is chapter 4 of The Coming Wave, by Mustafa Suleyman and Michael Bhaskar. The chapter is called “The Technology of Intelligence.” Its central claim is simple and unsettling. For centuries we built tools to manipulate matter. Then we learned to manipulate information. Now we are building systems that manipulate something even more powerful: the ability to learn, predict, plan, and act.

The old hammer hit where you swung it. The new hammer studies the wall, prices the nails, suggests an open floor plan, and launches a home-renovation channel. It can still mistake a load-bearing beam for a toaster, but it gets more capable every year.

The useful question is not whether the machine has feelings or secretly wants a long weekend. The useful question is: What can it do? What information can it reach? What tools and permissions does it have? Who checks the result?

We will travel from an Atari game to systems that can speak, see, code, and string tasks together. We will learn why AlphaGo’s famous Move 37 mattered, why a language model is more than autocomplete after three espressos, why bigger is not automatically wiser, and how to use AI without handing it the keys, the credit card, and your mother’s maiden name.

This episode is coming out in the week that we commemorate 9-11.

Today, 25 years ago I was in Manhattan. A day I will remember forever. Please remember those people that lost their lives today. Today I dedicated this post to all those friends that we lost that day. The only silver lining from that moment was the beautiful union between Americans, facing the tragedy and sadly, that union is nowhere to be found today.

Right now, while you are pretending to work at your desk or sitting in traffic on the 405 questioning your life choices, an artificial intelligence has already figured out the shape of two hundred million proteins. All of them. Every protein known to science. A job that used to take a PhD student their entire career, done in an afternoon of computing.

And do not think this is some faraway lab experiment that has nothing to do with your morning commute. This exact same protein technology is already being used to design new medications faster, to understand rare diseases, and even to dream up new batteries for the electric car you might actually be able to afford someday. So I am not here tonight to tell you this is pure apocalypse. I am here to tell you it is both things at once, the cure and the problem, sitting in the same box, with no instruction manual included.

Here is the part that is going to sting a little. That exact same technology, the one that just handed medical research fifty free years, is a close cousin of the algorithm deciding whether your credit card interest rate goes up, whether the bank trusts you, and whether your next job application gets read by an actual human or gets tossed by a keyword filter that never even opens the PDF.

Can you believe this? The same technology that might cure your uncle's cancer is the one deciding if you are "approved" to rent an apartment in NYC. And nobody, and I mean nobody, sent you a survey asking if that was okay with you.

Last night I asked an AI to help me sleep. It designed a breathing routine, dimmed the lights, played gentle rain, and reminded me that eight hours of sleep improves memory. Everything went beautifully until it sent me seventeen notifications asking whether I was asleep yet.

That is when I understood the future: a machine that can solve problems we never had... until we bought the machine.

Welcome to Behind the digital curtain. I'm Dr. Sergio Sánchez, Dr. Qubit. Today we're talking about artificial intelligence without incense, without a tinfoil hat, and without saying "this changes everything" every thirty seconds. Whenever somebody says AI will change everything, ask: everything what? My job? My kid's homework? The election? Sunday's sermon? Or just my cousin's profile picture, where he now looks like an astronaut even though he gets carsick in an Uber?

We will cover what AI is, what it is not, how it helps, what goes wrong when we overuse it, and what it could mean socially, economically, educationally, culturally, politically, and religiously. The short answer: it is neither angel nor demon. It is a powerful tool, built by people, trained on human data, and deployed inside human institutions. That means it can amplify our intelligence... and our bad habits.

A real doctor. Licensed, white coat, eight years of school. Once told a patient who got bleach in his eye, and I am quoting this directly, "sounds like you will be fine, take 2 aspirins and call me tomorrow." Then handed him the number for Poison Control, hangs up like he is hanging up on his mother in law, fast, awkward, done talking.

A ROBOT, asked the exact same question, wrote back practically a Hallmark card with a medical license. Apologized. Explained why he probably was not going blind. Gave him seven sentences of real advice. And wished him well, with all the sincerity of a machine that does not actually have feelings.

Guess which one patients preferred. (long pause) Yep. The robot. Welcome, cloud dwellers, to the era where your chatbot has better bedside manner than the guy who spent eight years in medical school and fifteen minutes with you.

This is Behind the digital screen, the podcast where a old medical dr, that would be me, Dr. Qubit, explains technology to normal people without using words that only make sense in a Silicon Valley board meeting or a Zoom call full of people pretending they understand them. Today we are talking about a study, published in one of the most serious medical journals on the planet, that is going to change how you look at your doctor, and at your phone.