Anthropic Just Filed to Go Public. Here's Who the AI Works For Now.

I think you might already know why I keep doing this. It helps me wrap my mind around the thing I believe is going to be the most important shift of our lifetimes. So let me walk you through what happened, in plain English, and why it matters for the plumber, the hairstylist, and the person reading this on their lunch break.

This morning the most powerful technology ever built took its first step toward Wall Street. A company named Anthropic, the maker of the Claude artificial intelligence, filed to go public. Going public has a name. It is called an initial public offering, or IPO. That is a fancy way of saying a private company is about to start selling small pieces of itself, called shares, to anybody who wants to buy them. You, me, a retired teacher, a hedge fund, anybody with an account.

The paperwork went to the Securities and Exchange Commission on June 1. The SEC is the government agency whose job is to watch over the stock market and keep companies honest. The number is hard to say out loud. Investors think Anthropic is worth about 965 billion dollars. That is bigger than its main rival, OpenAI, for the first time. And OpenAI is only days behind with the same move. Both want to be selling shares to the public by this fall.

The news is going to be about the money. Who is worth more. Who wins the race. Biggest stock sale in history. That is the headline. Here is the part nobody is telling you.

Who Does a Company Actually Work For?

When a company is private, the answer is whoever owns it. The founders. The early backers. People who, at least in theory, can choose to do the right thing even when it costs money.

Anthropic was built as something with a special name. A public benefit corporation. That means the company is allowed, on paper, to weigh what is good for people against what makes the most money. There is even a trust set up to hold a kind of share that lets the board pick its own members. The whole point was to protect the mission from being run over by the money. A wall, built on purpose, by people who saw the pressure coming.

What is the pressure? It is the day you go public. Because when you sell shares to millions of strangers, those strangers all have one thing in common. They did not buy in to make the technology good for the world. They bought in to make the price of their shares go up. That is not evil. That is just what buying a share is.

But once a company is public, the law tightens its grip on the people running it. There is a legal rule called fiduciary duty. Say it with me. Fiduciary. It means the people in charge are legally required to act in the financial interest of the shareholders. Not the people using the product. Not the mission. The shareholders.

Watch what that does. Right now, while these companies are still private, a leader can stand up and say we are going to spend a year and a half testing this before we let it out, because it could be dangerous. A private company can swallow that cost. A public company that says the same thing watches a rival ship first, watches its share price drop, and watches the board show the leader the door. Caution becomes a liability. Slowing down to get it right becomes the thing that gets you fired.

This Already Happened. His Name Is Sam Altman.

This is not a guess. In November of 2023, the board of directors at OpenAI fired Sam Altman, the man who runs the company that makes ChatGPT. A board of directors is the small group that sits above the chief executive and is supposed to keep watch. They said they had lost confidence in him, that he had not been straight with them about safety, and that they were blindsided before some big product launches. They fired him over exactly the stuff we should care about. Honesty and safety.

Here is the twist. Five days later he was back. The money came roaring in to save him. His giant business partner and the company's own workers, who held shares, forced the board to put him back. The board that fired him got replaced instead.

Sit with that. This was the board built specifically to protect the mission, not the money. They pulled the trigger for the right reasons. The money reversed it in five days, and that was while the company was still private. Now imagine that same fight after the company has gone public, after fiduciary duty is locked in and there is a share price on the screen every single day. The mission-minded board does not lose in five days. It never gets to swing at all.

Now picture a second leader. One who decides to say no to the government. Say the government wants the technology for spying on regular people, or for weapons that pick their own targets with no human pulling the trigger. While the company is private, he can say no, too dangerous, and the mission backs him. After the company is public, saying no to a giant pile of government money is the exact thing that gets him in legal trouble. A shareholder can turn around and say you left all that money on the table, you broke your duty to us. He becomes someone who can be sued for making the right moral choice.

And there is a third kind of leader. The one the shareholders cannot touch at all. The founder who holds enough voting power that he simply cannot be removed, no matter how strange things get. That is not the tyranny of the share price. It is the tyranny of one man. Look at the three together. One leader the money protects even when he is reckless. One leader the money could punish for doing the right thing. One leader the money cannot control at all. Three stories, one root cause. Going public does not put the right person in charge of the most powerful tool ever made. It just turns up the volume on the wrong incentive.

We Have Seen This Movie Before

Social media. The companies that ran the feeds had their own research. Their own people told them the machine was keeping kids scrolling and hurting them. The machine kept running anyway. Why? Because the thing they were legally rewarded to grow was attention. Attention was the number shareholders paid for. The kids were the number that lost. That is the last fifteen years in one sentence. The company grew the thing it got paid to grow, and the thing it got paid to grow was not your well-being.

Now take that exact machine and point it at a technology smart enough to study how your own mind works in real time while you use it. That is the deep end. And artificial intelligence is a different animal than anything before it, for three reasons.

First, speed. A bad decision inside one of these companies ships to a billion phones before the next time they report numbers to investors. There is no factory to slow it down. It is instant and it is everywhere at once.

Second, it builds itself. The technology is being used to make better versions of itself, so the pressure to cut corners does not just add up, it snowballs. Every company knows that slowing down might mean losing the lead forever. Smart people call that a collective action problem. Everybody would be safer if they all slowed down together, but not a single one can afford to be the only one who does. So everybody races, even the ones who know better. I wrote about that race tightening in the window already closing.

Third, and this is the one that should keep you up at night. After a company goes public, the people technically in charge are the shareholders and the board. And those people cannot actually judge whether a technology is safe. Nobody sitting on an investor board can read the inside of one of these systems and tell you it is dangerous. So you have the most powerful tool in history, and the people legally steering it can only see one thing. Did the number go up. It is like handing the controls of a jet to a room full of people who can only read the fuel gauge. As long as there is gas, full speed ahead. They cannot see the mountain coming.

The Fair Side, Because I Do Not Gloom Just to Scare You

Going public also means sunlight. A private company can do a lot in the dark. A public company has to open its books, file reports, and answer to regulators. Some smart people argue with a straight face that this daylight actually makes the companies safer, not more dangerous. That is a fair point and I am not going to pretend it is not. Needing to make money kills junk. A company that has to earn cannot burn billions while a tiny group of insiders decides what is good for you with nobody checking them. The mission-only version becomes its own kind of unchecked power, and we got a flash of that in the Altman story. So it is not as simple as public bad, private good. It is that the structure about to take over has a known weak spot, and we are about to test it on the highest stakes technology we have ever made.

The Bigger Thing Almost Nobody Is Saying Out Loud

Forget the AI companies for a second. Think of every other big company. The insurance company. The bank. The hospital chain. The store where you shop. Every one of them has shareholders. Every one of them lives under the same rule. Make the shareholders more money now.

Now hand each of them a technology that can do work a human used to do, cheaper and faster. No paycheck. No health insurance. No sick days. No vacation. Here is the trap in one sentence. The same law that says make the shareholders more money now also says, by simple math, replace the human. It is not a choice some cruel boss makes in a dark room full of cigar smoke. It is what the structure demands. If a machine does the job for one tenth the price and your legal duty is to chase the shareholders' money, then keeping that person is the thing you can get sued for. The human becomes the expensive, optional part, and expensive and optional is exactly what gets cut. That is the whole reason I keep saying augmentation beats replacement, which I broke down in AI replaces your workers, then who buys your product.

How Does It Get Changed?

I do not have a magic answer, but there are real ways, and some are stronger than others.

The weakest way is hoping companies just choose to be nice. That fails for the reason we already covered. A company that keeps expensive humans while its rivals switch to machines gets punished, and the board replaces the leadership. Niceness cannot beat the rule.

A stronger way is changing the kind of company itself. Remember the wall Anthropic built, the public benefit corporation? More companies could be built that way, with doing right by people written into the rule book so leaders cannot be sued for it. The catch is almost no big company chooses that on its own.

The strongest near-term way is a law that sets a floor. Here is the logic, and it is the key to the whole thing. If the law forces every company to do something, the money rule stops fighting it, because you cannot get punished for following a rule your rival also has to follow. A law that says you must give workers warning and retraining before a machine replaces them. Taxes on companies that automate, used to help the people who lost the work. Rules that simply slow the speed of replacement so people can catch up. Whether any of it passes is a fight in the voting booth, not a problem with the technology.

And here is my favorite, because it does not fight the money rule, it uses it. Give the workers a piece of the machine. If the technology makes a company ten times more productive, let the workers own a share of that. Then the workers are also the shareholders, and suddenly the company is legally bound to serve them too. You turn the very rule that was cutting people out into a rule that pulls them back in.

The Good News, Because I Have to End Here or This Is Just Doom

The rules are written by people. Court decisions. Tax laws. Company rule books. Board charters. Every single one of them is a human choice, and a human choice can be made a different way. The machine can be rebuilt. It is not a law of nature. It is a fight.

So when you hear people arguing about rules for artificial intelligence, and it sounds boring and far away, understand what is really on the table. They are deciding whether the bell that rang this morning turns these tools into something that works for you, or something that works on you. Whether a company down the street sees you as a person worth keeping or a cost worth cutting. This is not a technology story. It is a kitchen-table story.

Let me bring it all the way home. The reason I have been saying this for a long while, long before this morning's filing, is that I use these tools every single day. In real estate, in voice technology, in software for small businesses. I am not watching from the cheap seats. I have my hands in the machine, and the machine is good. It is the best edge a one-person operation has ever had. You can watch me run it on real data in the post where I pulled every SCV closing. But I never forget who built it, or who it is about to answer to.

This morning was the starting gun. Anthropic first, OpenAI right behind, both aiming for this fall. By the end of the year, the most powerful technology on earth is going to be owned in part by the stock market. And the stock market asks exactly one question. Did the number go up. So we ask a different one. The question they will not ask on the investor call. Does this actually work for the person reading this on their lunch break? That is the whole game. That is why the show exists. Because the people building these things are not thinking about you. That is not hate. It is math.

So we think about us. Artificial intelligence for everyone, not just the people who can afford to buy the shares. Let's be careful out there. I'm Connor with honor, and I'll see you in the next one.