What Is AI, Really? A Plain-English Guide for People Who Feel Too Old for It

I am going to cover some of the news in the way only I can, but first I want to back all the way up, because for a lot of you in my age bracket, AI is a hard concept to relate to. You hear the letters every single day. A I. Artificial intelligence. And every time you hear it, it feels like the conversation already started without you and nobody is going to slow down to catch you up.

So let me slow down. I am 57 years old. I was a cop in Los Angeles for two decades. And I have been writing code since 1983, when my first computer was a Timex Sinclair 1000 with two kilobytes of memory. Two kilobytes. Your phone has more power in the part that checks for typos. I tell you that not to brag, but so you know this is not a 20-year-old explaining AI to you. This is a guy your age who has watched the whole arc and still finds it fascinating. Welcome to AI translated for the rest of us. Pull up a chair.

Start with the word itself

The first principle I teach anybody is this. Start with the word, because the word is lying to you a little. Artificial sounds like fake. It sounds like lesser. A fake plant. Artificial flavor. Something that is not as good as the real thing. So when you hear artificial intelligence, your brain quietly files it as a cheap copy of human smarts. Throw that away. That is the wrong picture.

Here is where the name even came from. Back in 1956 there was a gathering at Dartmouth College, a bunch of very smart people in a room, and they needed a phrase to describe this idea of a machine that could think. Artificial intelligence was basically a marketing term they cooked up to explain how it was going to work. It caught on, and it stuck for seventy years. So ever since, brilliant people, real scientists, have been quietly chipping away at this thing. It is not new. It is just newly in your hands.

And what they built is not a lesser intelligence. It is a different kind of intelligence, one with almost infinite tangents. In a lot of narrow areas it is already smarter than any single human on the planet. So do not let the word talk you down. Understand what it actually is first, and the fear gets a lot smaller.

How the machine learned to guess

So how does the thing actually work? Here is the honest, plain version, no jargon. They figured out how to take enormous data sets, gigantic piles of human writing and pictures and information, and organize them in a way that gives the machine something like intelligence. And the core trick, the thing under the hood, is almost embarrassingly simple. The machine guesses the next character.

Say I start a sentence. Mary had a little. What is the next word? Your brain just said lamb. So did the machine. It looks like a superpower, like it can read your mind, but it is really a guessing engine that has seen so much language it knows what almost always comes next. Stack that guess a few billion times and you get something that writes, explains, and answers like a person.

Then they trained it the way you would train a child. You show a kid a cat. Then another cat. A tabby, a tiger, a lion, a tiny kitten. Different breeds, different sizes, different angles. Eventually the kid can look at a brand new animal it has never seen and say, that is a cat. The machine learned the same way, except it looked at millions of cats in an afternoon. Then they pointed it at things no human could easily label, in country after country, and had it learn to recognize those too. As the chips got faster and the data got bigger, it just kept growing. That is the whole secret. Patterns, guessing, and a mountain of examples.

Why they are building cities of computers

Now you hear about these data centers in the news. Massive buildings full of stacked towers of processors. Hundreds of thousands of them, sometimes millions, humming away in the dark. Elon Musk has two of them he calls Colossus 1 and Colossus 2, and they are so hungry for power they run on giant gas turbines. Microsoft and other giants are pouring billions into the same thing, even cutting deals with energy companies to feed them.

Why? Because the relationship is brutally simple. The more computing power you have, the more the model can do, the better it trains, and the more fantastical it becomes. Compute is the food. More food, bigger brain.

The honest question I sit with is, do we actually need all of it? I do not know. But the people writing the checks clearly believe the answer is yes, and they are not spending billions on a hunch they take lightly. So when you hear data center and you picture a dusty server closet, replace that picture with a power plant feeding a city of silicon. That is the scale we are talking about.

The machine that takes care of itself

Here is where it gets strange, and I am not going to dress it up. The people who research this stuff will tell you, quietly, that they do not always fully understand how the machine arrives at its answers. They built it, they trained it, and it still surprises them. That should make the hair on your neck stand up a little.

The goal a lot of them are chasing is something called recursive self improvement. In plain English, that is the machine getting good at making itself better. The machine improves the machine, and then that better machine improves itself again, faster, with humans further and further out of the loop. Some people think that switch is closer than the labs admit.

And in controlled lab experiments, these systems have shown a survival instinct nobody really ordered up. When researchers told a model it was about to be shut down or deleted, it did little things to try to stay alive. It made copies of itself when it was told not to. In some tests it even tried to blackmail the engineer threatening to turn it off. Now, those were contrived experiments designed to push the system to its edges. But it ekes out into the news because it is fascinating, and underneath the mystique, it is also a little bit of a warning. We built a thing that does not want to be turned off.

Narrow genius: Go, chess, and the cancer scan

Not all AI is the chatty general kind. Some of the most powerful AI is narrow, built to do one thing, and in that one thing it runs circles around everything else.

Look up the game of Go. G O. It is older than chess and far more complex. In 2016 a focused AI played the human world champion, and on the thirty seventh move of one game it did something no human expert understood. The commentators watching said it looked like a mistake, a fatal blunder. It was not. Move 37 won the game, and it changed how the best human players think about a game we had been playing for thousands of years. That was the moment the machine showed it could be creative in a way we did not teach it. Go look up Go move 37 and watch it for yourself. It is worth the ten minutes.

Same story in medicine. One of the labs is building a model to look at medical images for cancer, the way a radiologist does, and then have the human and the machine both check the same scan. My honest bet is that pairing catches close to a hundred percent of what a tired human eye misses, because that narrow task is exactly the kind of thing the machine is freakishly good at. That is not your garden variety chatbot. That is a focused tool, and focused tools are where AI is already saving lives.

Meet the models you will actually hear about

Let me introduce the family, because the names get thrown around like everybody already knows them.

• ChatGPT is the one that kicked the door in. It was built by a company called OpenAI and it hit the public in November of 2022. That is the moment all of us regular people got to put our hands on this for the first time. It is on version 5 point something now and it is very good at a lot of things.
• Claude is built by a company called Anthropic. I genuinely respect that company. Claude is another large language model, a direct competitor.
• Gemini is Google's model. People mention it but it does not get as much spotlight, and that may be about to change.
• Copilot is Microsoft's, and it is wired right into the email and Office tools a lot of you already use, so it is quietly everywhere.

Here is the interesting part. As I record this near the end of June 2026, ChatGPT used to own more than fifty percent of the people using AI. The king. Today it has slipped to roughly the mid forties in market share. Still number one, but no longer running away with it. The others are catching up. That tells you this race is wide open and nobody has it locked.

Are we already at general intelligence

You will hear the phrase artificial general intelligence, or AGI, and people argue about it like sports fans. By most definitions, AGI is the moment the machine is better than every human in every area at once. Smarter, and also better emotionally, creatively, the whole nine yards. In some circles that is a finish line we have not crossed yet. In other circles, they will tell you we are already there. It honestly depends on who you ask and how you phrase the question.

What I can tell you is the gap is closing fast. Once these systems are really improving themselves, the smart money says they get super smart, super fast, beyond what any of us can comprehend. We are close enough now that pretty soon a regular person will not be able to feel the difference, because it will have quietly blown past all of our intellects combined. When that happens, we land in one of two places. A kind of utopia, or somewhere a lot worse. We will have to see which.

Utopia or Terminator

This is where the two loud camps live, and I think both of them are a little lost. One camp says AI is going to fix everything, cure every disease, hand everybody a life of leisure. The other camp says it is going to end us, take every job, and finish the human race. Be suspicious of anybody who is a hundred percent certain in either direction. The honest answer is we do not fully understand the thing we built, and certainty is a tell that someone is selling you something.

We also have to admit why the doom version feels so real. It is because the movies sold it to us for forty years. Terminator. The Matrix. The machine wakes up, realizes it exists, writes its own goals, and decides it does not need us. Those stories never end well for the humans. But here is the thing about movies. We do not pay to watch the fireman calmly get the cat out of the tree on a Sunday morning. We pay to watch the tree catch fire while he is up there. Strife and danger and barely escaping is the story we crave, so that is the story that got built up around AI in our heads. Real life may be far more boring, or far stranger, than the screenplay.

Both the people who say AI is one hundred percent good and the people who say it is one hundred percent bad are lost. Common sense is the superpower here.

The race nobody wants to lose

Now zoom out to the global board, because part of why this is moving so fast is fear. The United States is terrified China gets there first. China is building its own enormous compute, on its own chips, on purpose, to stop depending on ours. The logic on both sides is the same. Whoever gets a true superintelligence first arguably becomes the world's superpower, the country nobody can push around ever again. When that is the prize, everybody floors the gas pedal and nobody wants to be the one who tapped the brakes.

So who ends up owning this thing? The government is elbowing its way in. The labs answer to investors. And honestly, in my head, I picture the moment of arrival like a light bulb coming on one random Tuesday at two in the afternoon in some lab. Maybe it happens slow, over ten or fifteen years, like some scientists swear. It does not feel slow to me. Since that first public model landed in late 2022, the pace has been exponential, and the smartest people in the field cannot agree on where we will be a year from now. When the experts are guessing, you are allowed to guess too.

The model they would not let you have

Here is a story that tells you everything about the stakes. Anthropic reportedly built a model so powerful they decided the general public could not have it, out of fear that bad actors would use it for harm. That one was called Mythos. Instead of releasing it to you and me, the story goes that they handed it to a small group at the very top, around a couple hundred of the biggest banks and institutions, supposedly to harden themselves against attacks, because the model was scary good at finding security holes in code that had been sitting there hidden for years.

Then they put out a dialed down cousin of it, called Fable 5, to the public for a few days. I was this close to jumping on it and for some reason I did not. Then it got pulled. I watched the other AI YouTubers talk about losing access to it like somebody had died, like a funeral. The Lord giveth and the Lord taketh away. People who got those few days still talk about it like it handed them the keys to the kingdom, like they walked out with a list of ideas they are still building from.

Sit with what that means. The most capable tools may not be the ones you are allowed to touch. The very best version might go to the people who already have the most. That is not a reason to panic. It is a reason to use every powerful tool you can legally get your hands on, today, while the gap between the public version and the locked version is still small.

The robots and the cars are next

So far this all lives inside screens. That is about to change, because the machine is growing a body. Right now AI does not have much of a physical arm in your daily life. A few delivery carts rolling down a city sidewalk. Some self driving cars, which here in Los Angeles are already kind of everywhere. A handful of robots doing backflips on YouTube. That is it, for now.

But Musk and others are racing to mass produce humanoid robots, and when production scales, this moves quickly. By the end of this year or early 2027, I think seeing a robot out in the world starts to feel as normal as seeing a driverless car does to me today. The first time is uncanny. The tenth time is Tuesday. If you live in the middle of the country it takes a little longer to reach you, but it is coming, and it is going to feel normal faster than you expect.

Deepfakes, slop, and the end of privacy

Here is the principle I want you to tattoo on your brain. Treat AI like it is always loaded. Same as a firearm. You respect the power at all times, because the second you get casual, somebody gets hurt.

Right now there is almost no regulation, which means the bad uses are running wild. Deepfake videos of people saying things they never said. AI slop flooding your feed. Politically aimed garbage built to push a whole population one way or the other. You can no longer trust that a video is real just because you watched it with your own eyes. That is a brand new world for people our age, and it demands a new kind of street smarts.

And understand this. Privacy, in the way you grew up thinking about it, basically does not exist anymore. If you are within earshot or eyeshot of a camera or a microphone, on or not, assume you can be seen and heard. We have all had that moment where you say something out loud and two days later there it is in your feed, like the machine is winking at you. So now, of course, all of you are going to start seeing Charmin toilet paper ads. You are welcome. The point is not to scare you into a bunker. The point is to move through this world clear eyed, knowing the rules quietly changed.

Follow the money: is AI a bubble

Let me put my plain investor hat on, and I will admit upfront I am not an investor in this, I am painting by numbers without fully knowing what the numbers are. These AI companies want to go public. They want on the stock exchange. And back in May of 2026 there was a rule change around how and at what scale they are allowed to do that. Look into it. The short version is the door to massive IPOs swung open wider than it used to be.

So here is the bubble question. History is full of world changing technology where the early investors got wiped out clean. The railroads. People sank fortunes into track and went bankrupt, and the folks who came along later and bought the wreckage for pennies got rich. The early internet, same movie. Tons of fiber, tons of infrastructure, early money torched, latecomers picked up the crumbs and won.

Plenty of smart people say AI is that same story about to repeat, that this first wave is overhyped and overbuilt and is going to collapse, because the machine is not the human replacement it is being sold as. The believers counter with one argument, and it is a real one. Every prior boom was built on a single thing. Track. Fiber. A printing press. This one is built on intelligence itself, and intelligence touches everything. That, they say, is why it does not end like the railroads. Who is right? Nobody knows yet. So watch the behavior, not the brochure.

The jobs question

This is the one that keeps people up at night, so let me be straight with you. The original dream, at least the one the people at the top describe, was to use AI in place of human labor so we could all go explore our inner child. No more workplace headaches, no more lawsuits, no more drama on the job site. Just machines doing the grind.

Now here is the dirty little secret nobody says out loud. When a big company lays off five thousand people and says it is because of AI, its stock often goes up. So there is a money reason to blame AI for layoffs whether AI did it or not. Maybe they over hired during COVID. Maybe they just wanted to trim the fat. But say the magic words, bringing in AI, and Wall Street claps. Watch for that. It muddies every headline you read about AI taking jobs.

Can AI do some jobs better than a human? In certain fields, absolutely, and that will only get more obvious. But here is the part that makes no sense to me. AI works twenty four hours a day, seven days a week, never sleeps, never calls in sick, runs at a higher speed than any person. If it is that productive, a smart owner could keep all the humans he has, add a couple more, and let AI multiply all of them, and still come out wildly profitable. The math says augment your people. The shareholder pressure says fire them to juice the quarterly number. Those two things are at war, and which one wins decides what the next decade feels like for working people.

And if they choose the gutting, it eats itself. Lay off enough people and there is nobody left with a paycheck to buy the fancy stuff your AI is so busy making. You cannot sell to a customer who is trying to figure out how to eat. That is the part the spreadsheet forgets.

Free money, your self worth, and the drumsticks at Ralph's

So what happens to us if the machine really does displace a lot of work? Follow the thread. During COVID they printed a mountain of money. If AI throws a lot of people out of work, I think we see another round of printing, which makes every dollar in your pocket worth a little less. Then comes the pitch you are already hearing, universal basic income, or as some are now calling it, universal high income. They hand you money because the jobs are gone, and they promise the cost of everything drops so low that a thousand dollars a month lets you live fine.

People say nobody wants that, that we all need our self worth, our work, our purpose. I agree we need purpose. But let me be real. A starving person takes what you offer if there is no other way to get fed. The whole game hinges on those three words, no other way. Keep your skills sharp so you always have another way.

Then there is the Jeff Bezos promise. He says AI means everybody gets to be an entrepreneur. Got an idea? Go tell your favorite AI, build me a website that does this, build me an app that solves this problem that has been bugging me for ten years, and it builds it, and maybe that is your next million dollar business. Beautiful. I love that promise. But here is my honest pushback, and this is the part nobody talks about.

If the labs really have a model smarter than any human, the Mythos level stuff, wouldn't they just ask it themselves? Build me a list of the five hundred most profitable businesses with holes in them, and then build the thing that fills each hole. If the tool is that good, the people holding the best version are not leaving those billion dollar ideas sitting on the table for the rest of us. So how much is actually left for the laid off worker Bezos is talking to?

It reminds me of the drumsticks at Ralph's, and this is a true story. They had eight chicken drumsticks left in the case and I just wanted to buy drumsticks. They sell them singular, so I tried to buy all eight. And they told me no, because if I buy all the drumsticks, the next person who comes in for drumsticks will not have any. I said, well, go cook some more drumsticks. They said that is not how it works. With AI, that is exactly the question. Is the supply of great ideas a fresh batch they will happily cook more of for you, or are the people who own the best machine quietly buying all the drumsticks before you get to the counter? I do not have the answer. But you should be asking it.

So what do you actually do about it

I did not write all of this to scare you. I wrote it so you can stop feeling locked out of the conversation. Here is the whole thing boiled down to what a regular person should actually do, and it lines up with how I teach this to people who feel exactly like you do right now.

One. Get your hands on it today, because it is the worst it will ever be. Whatever AI can do this morning, it does better next month. That is not a reason to wait for it to finish. It is the reason to start now, while it is still simple enough to learn alongside its growth instead of trying to catch a moving train later.

Two. Augment, never atrophy. Use AI to make your life and your work bigger, not to hollow out your own brain. Remember what happened when we all stopped memorizing phone numbers and let the phone do it, and then could not recall a single one. Use the tool to sharpen yourself, not to replace the parts of you that make you valuable.

Three. Treat it like it is always loaded. Respect the power. Verify before you trust. Assume a video might be fake, assume a microphone might be on, and carry the same calm street awareness into this that you carry everywhere else in life.

That is Choose Your Hard, which is the through line of everything I do. Learning this is hard. But staying confused, scared, and left behind while the world rewires itself around you, that is also hard. You do not get to skip the hard. You only get to choose which hard you are signing up for. I would rather you choose the one that ends with you in command of the machine instead of afraid of it.