Washington Already Solved AI. Then the Money Showed Up. | Monologue Monday

🎥 WATCH COMMON SENSE ON RUMBLE
https://rumble.com/c/CommonSenseWithChadLaw

Washington is crawling with psychics.

Politicians, television personalities, and technology executives can tell you exactly what artificial intelligence will do to your job, your children, and the country twenty years from now.

What they cannot tell you is what the law says today—because Congress still hasn’t produced one comprehensive AI framework.

While Washington predicts the future, rival factions of the AI industry are pouring roughly a quarter-billion dollars into writing the rules. A resolved fight is worth nothing. An unresolved fight keeps the lobbyists, political campaigns, and fundraising machinery running.

But America already has a blueprint.

In the 1990s, a Republican Congress and a Democratic president created an Internet Truce. They didn’t pretend to know what the technology would become. They built a foundation that allowed Americans to find out.

In this episode, Chad explains:

🔮 Why prophecy is easier than legislating
💰 How Big Tech can use regulation to protect itself
🤝 The forgotten Clinton–Gingrich Internet Truce
🇺🇸 What Trump’s AI framework gets right—and leaves unfinished
🇪🇺 Why Europe got the rules while America built the companies
🚗 What electric-vehicle subsidies reveal about government certainty
📜 The seven-part Common Sense AI Truce Congress should pass now

The central question isn’t whether Washington can guarantee the right AI future.

It’s whether Washington can admit that it doesn’t know—and create room for Americans to build it anyway.

💬 JOIN AFTER HOURS Q&A

Call or text 252-CHAD-LAW with your comments, critiques, and episode ideas. Your message could be featured on the show—and every text enters you in the drawing for signed books.

📚 Get the free Common Sense Conservative Manifesto, read Chad’s editorials, and find everything Common Sense at ChadParkerLaw.com.

👍 Follow the show, leave a review, and share this episode with someone tired of being told that fear counts as governing.

COMMON SENSE WEEKLY

Monologue Monday
Wacky Wednesday
Freedom Friday
Satire Saturday
Sequel Sunday

Transcript

SPEAKER_00 0:00

America, we are crawling with psychics. Nobody in Washington can tell you what a gallon of milk costs. Every one of them can tell you exactly what artificial intelligence is going to do to your grandchildren in the year 2049. Apparently we didn't elect a Congress, we elected a psychic hotline. And the readings are free, which is the only thing they've ever given away. Turn on anything this week. You'll get the surveillance state your grandkids live under, and the exact month your job disappears, delivered by people who could not tell you what the law says about any of it. Why? Because there isn't one. Here's the number of comprehensive AI framework bills Congress has managed to move out of committee. Zero. Now, that's not not enough. Not too weak. No. Zero. Zero. They held hearings. They made the concerned faces. They went on the podcasts. Six actual members of Congress even produced a real bipartisan draft. Three Republicans, three Democrats, which in this town is basically a Christmas miracle. And leadership in both parties waved it off because it wasn't robust enough. Which is Washington speak for, I would rather keep the issue than solve it. So here's where we are. A government that cannot produce a framework for artificial intelligence, but cannot stop producing predictions about it. Apparently prophecy is easier than legislating. Prophecy doesn't have to get out of committee. In America, Chad Law here, the last gay conservative, and this is common sense. The show that reads the fine print so you don't have to. And if you've got something to add while we're doing this or you think I've got it dead wrong, text me, 252 Chad Law. I read them at the end of the show. So let's take this thing apart. The people warning you the loudest about what AI might do to you have not agreed on what AI should be allowed to do. That is not governing. That is a ghost story told by people who did not do the reading and who are counting on you to not ask whether they did. And I want to tell you why they do this, because once you see it, you'll see it everywhere. Nobody in Washington is allowed to say the three most useful words in the English language. I don't know. You cannot say it in a hearing, you can't say it on a panel, you can't say it at a press conference because the next question is, then why are you here? And there's no good answer to that one. So a gap opens up. Nobody actually knows what this technology becomes in twenty years. Not the senators, not the anchors, not the CEOs, and not me. And I've been reading about this until two in the morning like a lunatic. But politics cannot leave a gap empty. So it fills that gap with the loudest emotional thing available. Robots, surveillance, your job, your children. Fear is the only substance Washington can manufacture in unlimited quantities without knowing anything at all. And most of these people are not lying. They believe every word. The fear is a substitute. It's what they say instead of saying, I don't know. But that only explains the noise, it does not explain the silence. Because a gap that stays open for three straight years isn't confusion. Confusion resolves. Somebody eventually reads the thing. A gap that stays open for three years is a gap somebody is being paid to leave open. And I've got a receipt on that. It's got a date on it, and the date is gonna make you sick. Now, here's the part where I sound crazy. We already know what to do. Not because we can predict artificial intelligence, because we've faced this exact governing choice before. The last time a technology arrived that promised to reorder commerce, communication, culture, privacy, and power, a Democratic president and a Republican Congress did something almost unimaginable today. They agreed on the rules of the road. And then they had the restraint to let Americans build the actual road. It worked. It produced the greatest explosion of innovation and prosperity in modern American history. And the blueprint is sitting in an archive right now, under an inch of dust. We do not need Washington to predict the future, contrary to their opinion. We need Washington to remember how it governed the future once before. So tonight, what Washington is doing instead of governing and who's paying for it? How two men who could not stand each other won the last one, and what it costs a country when it decides it already knows the answer. So let's talk about Washington has been doing instead. I'll be fair about it. I said Congress hasn't produced a framework. I did not say Congress has been idle. Congress has been very busy. Three years of hearings, subcommittees, task forces, working groups, a bipartisan house commission, open letters to CEOs, site visits, a caucus, there's always a caucus. You should understand something about a caucus. A caucus is what you form when you would like the appearance of a committee without the consequences of a committee. A committee has to vote. A caucus just has to meet. We've got the White House framework in March, a real document, a serious document, and I'll give it its due later tonight. But parts of it are right. But here's what a framework is. A framework is merely a piece of paper that tells Congress what it ought to do. Congress read it. Congress did not do it. Now here's the part that got me. This is the number I could not stop thinking about all week. While Washington was doing all of that, the 50 state legislators were legislating. In 2025, lawmakers in all 50 states introduced twelve hundred and eight artificial intelligence bills. 145 of them became law. By March of this year, 45 states had already filed 1,561 more. 3,000 bills in two years, every state in the Union. Wyoming has a position on a large language model. Wyoming got there first. And in that same window, the number of comprehensive AI framework bills the United States Congress moved out of committee is the number I gave you at the top of the show. Still zero. Now, someone's going to text in to me about this, so let's go get ahead of it. Congress did pass one AI-related law in that stretch. It's called the Take It Down Act. It's about nonconsensual intimate images, including deep fakes, and it's a good and necessary law. I'm glad it exists. It's also about four inches wide. It tells you what happens when somebody makes a fake nude picture of you. It tells you nothing about who is responsible when a model does something nobody intended, or which government gets to make the rules, or whether a company in Ohio can build something without asking permission from California. One law about the worst thing. Zero laws about the thing itself. Now I know somebody's already yelling at their television, Chad, the president has done plenty. He has. And I'll score it honestly later. In December, he signed an executive order that stood up a litigation task for us at the Justice Department to go after state AI laws in court, told commerce to go find the offending ones and put certain federal broadband dollars on the table as leverage. That is a real strategy. It's aggressive, it's coherent, and it's fast. And it has the lifespan of a house plant. Because that's what an executive order is. An executive order is a policy with a shelf life of one election. The next guy walks in, signs a piece of paper, and it's gone before lunch. Here's the comparison that should end this argument forever. When they did this the last time, they didn't do it by lawsuit, they did it by law. The Internet Tax Freedom Act passed the House without opposition and cleared the Senate 96 to two. You can't get 96 senators to agree on a lunch order. Then Bill Clinton signed it and it held through him, through Bush, through Obama, Trump, Biden, Trump again, and Congress made it permanent in 2016. That is not a preference, that's a foundation. What we have now is a president renting a policy in a Congress that won't buy one. Which brings me to the question I couldn't answer all week, and it bothered me. Why isn't anyone asking about this? You watch an hour of cable news about artificial intelligence, and you will get robots, jobs, china, and a very serious man saying we're at an inflection point. You will not get one question about who is liable when the thing screws up. And I went looking for the conspiracy because that's the fun answer, there isn't one. It's worse. It's structural. The framework question is unbookable. Think about what a panel segment needs. It needs two people who disagree, it needs a villain, it needs a picture. Now try to book this. Should liability attach to the model developer, the company that deploys it, or the person who used it? There is no video, nobody's crying, there's no bad guy. There are four possible bad guys, and it depends on the contract. You can't put a Chiron on liability allocation. I understand that, you understand that. But meanwhile, there's a clip of a senator asking a CEO whether the computer has feelings, gets four million views before dinner. And there's a second thing. Nobody's job is this. When I went looking for where this actually stands, I found it in congressional trade outlets and law firm client memos. Lawyers writing to their clients. That's where the reporting on the future of American technology policy is living right now. Not one minute of it was on television. Because to ask a real question, you'd have to read a 269-page discussion draft, and every newsroom in America laid off the person who used to do that. That person never once produced a segment that rated. Here's the way to think about it. It's the difference between covering the kitchen fire and covering the lease renewal. The fire is television. Flamed sirens, a guy with a hose. The lease determines whether the restaurant exists in five years. Nobody covers the lease. Now, there's one more group filling the sentence, and these people I want you to watch carefully because they sound like the good guys. But first, remember the psychics? The loudest one in America is not on cable news. He's the one who's running the AI companies. Last year, the CEO of Enthropic told Axios that this technology could wipe out half of all entry-level white-collar jobs inside five years and push unemployment to somewhere between 10 and 20%. Finance, law, consulting tech. He has since doubled down. Meanwhile, the CEO of Nvidia says that productivity gains create more jobs, not fewer. So these are two men in the same industry and the same technology, same week, with completely opposite readings. And I'll be straight with you because it would be cheap not to be. The anthropic guy is predicting something that is bad for his own business. That's not a marketing stunt, and I'm not going to pretend that it is. He might be right. But sincere or not, it is still a reading, and you cannot write a law on a reading. You can't put somewhere between 10 and 20% in one to five years, probably into a statute and expect a judge to enforce it. So back to the good guys. The largest artificial companies on earth keep going to Washington and asking to be regulated. And everybody hears humility. Look at that. They're worried too. They even want guardrails on themselves. Let me tell you what I hear. Before marketing, before any of this, I was a licensed cosmetologist. I stood behind a chair. And I am not against licensing. I'll go first on that. If you're putting a chemical on a stranger's scalp that can burn them or blind them if it gets into their eye, you should have to prove that you know what you're doing. That's the floor, and I want the floor. Here's the part that isn't the floor. In Oregon, that license takes twenty one hundred hours of school. In New York, it's one thousand. They use the same chemicals, the same scissors, the same human head. Oregon hair is not twice as dangerous as New York hair. It's on the other side of the line on a map. That's it. I always figured the real safety material, sanitation and chemistry, and infection control and don't blind anyone was maybe four hundred hours. Turns out I was even being generous to my own industry. The Institute for Justice went through the actual curricula across eleven states, averaging sixteen hundred and fifty hours. Safety and sanitation is about ten percent of the training. Ten percent. So call it 160 hours. So what are the other 1,500? Their tuition. And nobody sat in a room and drew that up as a scheme. That's the part people get wrong. It happened because everybody who benefits from the rule shows up to the hearing about the rule. The schools testify. The board is staffed by people who came up through the schools, the trade association brings a very nice binder. And the person who is not in that room is the 22-year-old who can't afford 18 months of tuition and who is doing hair out of her kitchen instead. Illegally, by the way. So when the biggest AI companies in the world stand up and ask for sensible safety rules, I don't hear a villain. I hear the beauty schools. I hear people who probably believe every single word, what they're saying, and who have also noticed something. Compliance cost is flat. A licensing regime, an audit requirement, a registration threshold, those cost about the same whether you have four employees or forty thousand. To a company worth $300 billion, that's a rounding error. To the company that would have beaten them, it's the entire runway. The incumbent never asks for the rule that hurts him. He asks for the rule that hurts the guy behind him. And this is not a theory because you can read the filings. There is a super PAC called Leading the Future, launched in 2025, north of $100 million, and Dreessen Horowitz, the president of Open AI and his wife at $25 million, a Palantir co-founder, a Silicon Valley investor or two, their goal is one national rule that preempts the states. There's a second pack seeded with about 20 million by Anthropic. And I'll note for the record that Anthropic makes the AI I use to research my show, which I've never hidden and I'm not going to start tonight. Their group wants safety requirements and opposes broad preemption. There's a third pack run by a White House advisor that has pledged $100 million. $50 million of it has already been spent on this year's midterms. Now I want you to hold on to something. Last year, the United States Senate did vote on whether to freeze state AI laws. It failed 99 to 1. 99 to 1, ladies and gentlemen, that is not a close question. That is the Senate agreeing about something, which happens roughly as often as, I don't know, Haley's comet, and a fight that was settled by a vote of 99 to 1 is now being rerun with a quarter of a billion dollars in campaign money. So understand what's in front of you. This is not an industry lobbying against regulation. It is two rival industries, one buying preemption, one buying safety rules, spending against each other at a scale that rivals the political parties. And neither one of them is buying anything for you. Whichever side wins, the rule book gets written by whoever could afford to be in the room. And the four people in a garage who might have built the better version of this were never invited because they don't exist yet. Companies that don't exist yet do not have lobbyists. So that's the theater, and it's not gridlock. Gridlock is when two sides want different laws. This is what happens when nobody can say I don't know. So everybody performs instead, and the only people in the room with a straight answer are the ones with a financial interest in it. Which would be a hopeless place to end. Except we've been here before, folks. And last time we won. Let me show you how. Here's what nobody tells you about winning. When policy fails, you get a hearing, a documentary, and somebody's name on a scandal. But when policy works, you get nothing. There's no anniversary for the tax that never got levied. Nobody's making a limited series about the lawsuit nobody filed. A policy victory doesn't look like a victory. It looks like a completely normal Tuesday. And normal Tuesdays don't trend. That's why we forgot. Not because it was complicated, because it worked so well it became invisible. And thirty years later, a country that already won this fight can be sold the exact same fear a second time in nicer packaging by people who were in middle school for the first one. So let me remind you what we did. August fourth, nineteen ninety five, the House of Representatives votes on an amendment written by a Republican from California named Chris Cox and a Democrat from Oregon named Ron Wyden. four hundred and twenty to four. Not one member of Congress stood up and spoke against it. That amendment became Section 230, the rule that says if a stranger says something on your platform, that's the stranger's problem, not yours. Now here's the date I promised you. Five days after that vote, five, a company called Netscape went public, opened at $28 a share, closed the day worth over $2 billion. That was the starting gun. That was the morning the internet stopped being a science project and became a business. So run the clock. The United States Congress wrote the rule book for the commercial internet five days before the commercial internet had a stock price. There was no internet lobby in 1995. There couldn't be. There was nothing to lobby with. No PAC, no trade association, no war chest. Nobody in that industry had a dollar in Washington because nobody in that industry had a dollar, period. And that is why Chris Cox could say what he said. Nobody came at him looking for an advantage over a competitor, and he admitted years later it was the only time in his career that ever happened. That's not a man describing his own integrity. That's a man describing an empty hallway. Nobody bought the internet's freedom. It was free because nobody had thought to charge for it yet. Now, everybody thinks that this came out of the Clinton White House. It did not. Clinton's framework for the Internet came out of uh in July of 1997. Congress, the Republican Congress, beat the president to it by two years. Two branches, two parties, no coordination, same answer. But nobody was following anybody. So let's put the credit where it goes, because everybody's version of this story is wrong. Gingrich's Republican House wrote the law and passed it twice, Section 230 in 95, the tax moratorium in 98, both out of a Republican Congress, both with Democrat co-authors on the bill. Clinton's White House supplied the doctrine, and it took the rest of the world. Cox and Wyden built the bridge, one Republican and one Democrat, and they did it twice on purpose, and Clinton signed it. Not one of those men gets to claim the internet by himself. The Republicans can't have it, the Democrats can't have it, and anyone who tells you otherwise is selling you a t-shirt. That's not an agenda, that's what a truce looks like. Compromise, actual governance. And that's the name for it. The Internet truce. Somebody should have put it on a plaque. And then in 97, the White House put out five principles. And why they were brilliant is the opposite of what people assume. The genius was not what they wrote down. It was what they refused to write down. They did not decide what the internet was for. They did not decide who was responsible for what a stranger typed. They did not appoint one single human being anywhere in the United States government to approve it. Five principles that amounted to this. We don't know what this is yet, so we are going to protect it from ourselves until we figure it out. There it is. The three words in a federal policy document on purpose from a president and a Congress who spent that entire decade trying to end each other. And one of those attempts involved an impeachment. They didn't fill the gap. They built around the gap. You want to know what that bought? This morning, you deposited a check by taking a photograph of it. Somebody in Nebraska sold a set of used golf clubs to somebody in Norway before lunch. Nobody permitted either of those. Nobody licensed them, and if you had described them out loud in 1997, somebody would have taken your temperature. That wasn't luck. That was a decision made on purpose by people who knew they had no idea what they were making room for and who did it anyways. They built a country sized empty lot and refused to tell anybody what to build on it. Now not everyone agreed, and this was my favorite fact in the entire episode, so stay with me. The same bill that carried section two hundred thirty also carried something called. The Communications Decency Act. Senator James Exxon from Nebraska. His bill made indecent material online a crime with jail time and fines. So Congress passed the right answer and the wrong answer in the same law on the same day, which is the single most Congress thing that has ever happened, by the way. Sixteen months later, the Supreme Court threw the Exxon half out. Two justices dissented only in part. The other half is still on the books today. It built the thing you're looking at right now. America ran a controlled experiment on this exact question with one bill, same day, and we have the results. And by the way, Jim Exxon wasn't a monster. He was a grandfather who'd seen what a modem in 1995 and wanted to take it away from his kids. That's a decent instinct, and I get it. But being scared is not a policy. Scared is a feeling. And when you legislate a feeling, you get something the Supreme Court can throw out before the technology finishes booting up. Then there were the psychics. Of course, there were psychics then too. 1995, Newsweek runs a piece headlined, The Internet? Bah. The author calls online shopping baloney. Telecommuting baloney. He writes, and I promise I'm not making this up, that no online database will ever replace your daily newspaper, and no computer network will change the way government works. He published that in a magazine, which you now read on a computer network about a government that changed. But here's the thing that gets me. That man, Clifford Stoll, came back years later and said it flat out. I was wrong, called it his howler. Said now, when he's sure he knows what's happening, he reminds himself he might be wrong. The single wrongest man in the history of the internet found a way to say, I just don't know. He just had to do it afterwards. Not one person currently on your television about artificial intelligence has managed it beforehand. So when somebody asks which side of history you want to be on, that's the choice. Not the smart people versus the dumb people. The people who admitted they couldn't see the future and the people who charged for the reading. Now here's where it gets funny, and then it gets a little sad. The funny part is that a lot of these people are still there. Not their portraits, them. There are members of the current United States Congress who were in that building when this passed. Some of them have now served longer than the Internet has existed, which means we've got people voting on artificial intelligence who were already senior when the answering machine was aspirational. Here's the sad part. One of the two men who wrote it is a sitting United States Senator. Section 230 and the Internet Tax Freedom Act twice, both times with a Republican on purpose. Right now, today, Ron Wyden from Oregon, in Congress since 1981, 45 years, which is not a compliment and not an insult. It's a number that should probably alarm all of us. And he's not in a museum. He's not on a stamp. He's in the Dirksen building, and his phone number is published on the internet he helped protect. He's also my senator, and I disagree with Ron Wyden about 99 out of a hundred things. I'll give you the list some other night, but he's sitting on the answer. He wrote the answer. He wrote it twice. But boy, do they love milk in this lobbying money. And after three years of hearings, open letters, panels, predictions, and a quarter of a billion dollars in campaign money, apparently nobody in Washington has walked down the hall and knocked on his door. That's the amnesia. The instructions aren't lost. The author is alive, employed, and 300 feet away, and everyone would rather be a psychic. So here's my entire proposal for artificial intelligence, and I cannot believe I'm the guy saying it out loud. Copy it! Not adapt it, not modernize it, not convene a bipartisan commission to study a modernized adaptation of it and report back in 18 months with findings named after somebody's dead colleague. Copy it. Find the word internet, replace it with artificial intelligence, print it, vote, go home. All right, don't literally run find and replace. I can hear the lawyers reaching for their phones. AI is not a bulletin board. A model writes the sentence. Section 230 does not port over, and I'm not asking for it to. What I'm saying is copy the posture, copy the restraint, copy the one national market and the tax neutrality, and then write the narrow rules this thing actually needs, which is about 40 pages of work that nobody wants to do because there's no camera involved. And then then we can argue about everything else, about the robots, argue about the jobs, the surveillance, whether the thing is conscience, whether it writes a better country song than Merle Haggart. I've got opinions on all of it, and I'll happily give you 11 hours. Argue for 30 years. I don't care. Because you cannot argue about how to govern a technology before you've agreed on the first rule of governing it. That's not a debate. It's a hundred people in an empty parking lot screaming about where to park. Yeah, somebody's already got an objection ready, and it's a good one. So I'm gonna make it for them. The objection is this. You can't just photocopy 1996 because Section 230 protected a bulletin board from what a stranger typed, and a model doesn't host the sentence. It writes it. That's correct. That's a real difference, and I'm not gonna pretend it isn't. Also not my point. And that argument, who's responsible when the machine writes the sentence, is narrow, technical, boring, and exactly the fight Congress should be having and isn't. Nobody's having it because it doesn't book on cable. There's no Chiron. We already covered this up top. All right. I told you I'd give the president's framework its due, and I'm a man of my word, so let's score it, honestly. Which means you're gonna hear me praise him and hit him inside the same two minutes, and if that scrambles you, that's a you problem. What he gets right, and he gets a lot right. No new federal AI agency, correct? And it's the single most important thing in the document. Use the regulators we already have inside their own sectors instead of pretending artificial intelligence is one industry. Also correct. Industry-led standards, testing environments, sandboxes, correct? And Utah's already proving it works. And this one nobody expected from a Republican White House. The recognition that if you make liability broad enough, you don't punish the giants, you protect them. The giants can afford the lawyers, the startup can't. Correct. And it took Washington 30 years to figure that out, by the way. Now what he gets wrong, and it's less a wrong than an unfinished. That document is a philosophy, it's not an operating system. It does not say who is responsible when something goes wrong, developer, deployer, the company that packages it, the person who typed the prompt. That's four different roles, and the framework treats them like one guy. Its national security review thresholds are vague enough that they could quietly become a permission slip, and a permission slip is the European model wearing a cowboy hat. Open weight models, unresolved, copyright and provenance, unresolved. And the whole thing is an executive order, which means it lives as exactly as long as the president does. So here's my grade. Trump obviously understands what could kill American AI. He has not yet built what will allow it to live. And that's not a dig, that's a job description. The building part was never the president's job in the first place. It's Congress's. It was Congress's in 1995 too, and Congress did it. Which brings me to the thing that's really been bothering me all week. America, the choice you're being offered and the choice you actually have are not the same thing. Build it or stop it. Cameras or no cameras. Where the data sleeps at night. Every one of those is a prediction wearing a policy costume. That's a reading. Because all of them are arguments about what artificial intelligence is going to become. And nobody gets to choose that. Not you, not me, not a senator, not a CEO, not the guy with the newsletter. You cannot legislate an outcome you can't observe yet. Period. There's exactly one choice on the table, and it isn't about outcomes, it's about method. Do we demand certainty before we let anyone build? Or do we let people build and find out which guarantees were ever necessary? And that is not a thought experiment because it has already been run on this exact technology at the same time on two continents. You see, in the late nineties, Europe and the United States got the internet on the same day. One of them decided in advance what it was for and who was responsible for it. The other one decided to find out. Thirty years later, one of them has the companies, the other one has the rules. Now, I'm not going to stand here and tell you Europe has no technology industry because that's lazy and it's false and somebody will clip it. Europe has brilliant engineers, world class universities, and plenty of money. In 2024, the European Commission asked Mario Draghi to write a report on why Europe was falling behind. Draghi ran the European Central Bank. He was the Prime Minister of Italy. This is not a guy from the Heritage Foundation. He's a very leftist. Here's what he found. There is no company in the European Union founded from scratch in the last fifty years worth more than a hundred billion euros. Not one. In fifty years. And every American company worth over a trillion, all six of them, was created within that same window. Of the fifty largest technology companies on Earth, thirty-six are American, Europe has four. And this is the detail that stopped me cold. Europe's three biggest research and development spenders have been car companies for 20 years running. America's top three are the ones you're arguing with on your phone. We had that same profile once, autos and pharma in the early 2000s. Ours changed, theirs didn't. What did Europe get instead? They got the cookies banner. That's not a joke I'm making. That is Europe's most widely experienced contribution to the internet. A button on every website on earth that you click to make a box go away. Europe regulates uncertainty. America regulates demonstrated harm. One of those builds compliance departments. The other one builds trillion on trillion dollar industries. And look, I will put the American internet settlement up against any technology policy any government on this planet has ever written. Bring me a better one doesn't exist. But I'll wait. Now don't get comfortable because we did our own version of this. And we did it recently. And I want to be honest about it. America has a second way to lose. We don't forbid the technology. We decide we already know which one wins, and then we buy the customers. Electric cars. I'm not against electric cars. If you want one, buy one. It's a free country and they're quick. I own one. But watch the mechanics. Two federal tax credits expired September 30th of last year. The month before they expired, battery electrics hit a record 12% of new car sales. Fourth quarter, right before, all eight time record, four hundred and thirty-seven thousand units. Then the credit went away. First quarter of this year, two hundred and sixteen thousand, cut in half in one quarter. That was not demand. That was a deadline. And here's the part that should keep somebody in Washington up at night. The hybrid, which never qualified for a nickel of any of it, hit a record 16% share. The product nobody paid you to buy outgrew the product they paid you to buy. That's not a story about cars. That's a story about a government that was certain it knew what Americans wanted, spent your money proving it, and got corrected by a Toyota. So there are two ways to lose this AI deal. Europe demanded guarantees before construction and ended up regulating other people's inventions. We decided we already knew the answer and paid for it until the check bounced. Different politics, but the same disease. Both of them started with somebody in a Capitol building decided they can see the future. Which brings us back to the only real question on the table, and it turns out somebody already wrote the answer down. So if the answer is sitting right there, why hasn't anyone picked it up? Three reasons, and none of them are the ones you'd guess. One, nothing is forcing them. The internet tax bill happened because 30,000 local taxing authorities were lining up to tax your modem, and businesses screamed loud enough to create a deadline. Deadlines produce laws. Artificial intelligence has no deadline. The cost of doing nothing lands on companies that don't exist yet, and companies that don't exist yet don't have lobbyists. Two, this is the one that actually matters. An unresolved fight pays, a resolved one doesn't. There's a name for this. Economists call it rent extraction, and it comes out of the work of a law professor named Fred McCensey or McKenzie back in the nineties. The plain version is simpler than the name. We all understand that politicians can get paid to pass something. What McKenzie documented is that they can also get paid to not pass something. You float the bill, the people it would hurt come running. The bill quietly dies in committee and everybody's happy. State legislators have their own word for it. In California they call them juice bills, in Illinois, Fetcher bills, everywhere else cash cows. The industry named this practice itself because the industry does it on purpose. And I want to be careful, because I told you earlier there's no conspiracy here, and I mean it. See, nobody has to sit in a room and plan this, nobody has to even say a word out loud. You just have to notice what's true. Right now, a resolved AI fight is worth zero dollars and an unresolved one is worth a quarter of a billion and climbing. Two industries buying opposite outcomes in an election year. Nobody has to be corrupt for that to work. The incentive doesn't for them. That's worse than a conspiracy because you can't indict an incentive. And that is the entire difference between nineteen ninety five and tonight, not the technology, not the people, not some vanished golden age of statesmanship. The money. In nineteen ninety five there was nothing to milk. Today there's a herd. That's not a moral victory back then, and it's not a moral failure now. It's timing and we spent it. three, preemption is the poison pill. Republicans want one national rule. Democrats want the states to keep their authority. Neither will trade. And the Senate already settled it ninety-nine to one, which is now being relitigated with a quarter of a billion dollars in campaign money in an election year when a solution is just a win the other guy gets to claim in October. An issue you can fundraise on, a law you can't. So here's the deal I'd put on the table. I call it the AI truce. Seven pieces, I'll be quick, and I want you to notice something as I go. Not one of them requires anyone to predict anything. One, one national market, federal rules for how these things get built. States keep everything they've always had over how they get used, fraud, kids, consumer schools, COP zoning. Fifty different rule books for building a model is not federalism, it's a tariff. Two, no new AI agency. We have regulators. The FDA exists, the FTC exists, banking regulators exist and have opinions about everything. We do not need a Department of Artificial Intelligence with a nice lobby and a 40-year lease. Three, conduct over capability. We punish what people do, not what a tool could theoretically enable. Fraud is already illegal. Discrimination is already illegal. Running unpermitted turbines is already illegal. You don't need an AI law to enforce the Clean Air Act. You need somebody to enforce the Clean Air Act. Four, say who's responsible. Developer, deployer, the company, the packages it, the person who used it. Four roles, four levels of responsibility written down. This is the boring one. This is also the one that matters the most, which is why nobody's doing it. Five, protect the garage. Research startup sandboxes, open weight development. And this isn't theory. Utah built a regulatory sandbox for legal services in 2020 to test whether you could loosen the rules without hurting consumers. Fourteen complaints in the first stretch. About one per 400,000 services. It also turned up real problems, and they fixed them, which is the entire point. A sandbox finds harm, a ban finds nothing forever. Six, infrastructure without corporate welfare. And I mean this at both ends. No permitting obstruction, but also no abatements, no sweetheart deals, no handing a trillion dollar company a tax holiday. Bring your own power, pay your own freight. My grandmother should not be subsidizing anybody's data center in her electric bill. And last seven, tax neutrality, which brings me to the one piece of actual legislation I want. The AI Innovation Tax Freedom Act. Same idea as 1998. Same structure, same reasoning. No special taxes invented to punish or milk this technology because it's new and unpopular. No tax on tokens, on training, on interference, on API calls, on a subscription just because there's AI inside it. Ordinary taxes still apply, income payroll, property sales, all of it. Same as your dry cleaner. And somebody will lie about this by Thursday, so hear it from me first. There is no tax immunity for trillion dollar companies. It is tax neutrality for a technology the entire economy is about to be using. Look, if you want to tax Google, tax Google. What you don't get to do is invent a brand new category of tax that lands hardest on the 26-year-old building something in a spare bedroom. That's seven pieces. Not one of them requires a single person in Washington to know what happens in 2049. That's the entire trick and the genius of the late 90s internet framework. Now, before I let you go, I owe you something. I've spent an hour telling you that the people on your television can't say three words. So I'm gonna go first. Is a truce even still possible now that there's a quarter of a billion dollars in the room and two industries buying opposite outcomes in an election in the middle of it? I don't know. I don't. I think it's possible. I can't prove it. I could be completely wrong about all of it. And if I am, some guy in 2055 is gonna put this episode on television and laugh at me the way I laugh at the man who said online chopping was baloney. Fine, put me on the list. At least I'll have said it out loud while it still mattered. And I'll tell you the one qualification I've got for saying it, nobody is paying me either way. Because that's the whole argument. Not that I know what artificial intelligence becomes, not that Clinton knew or Gingrich knew or Cox and Wyden knew. They didn't know either. And they refused to pretend. And then they built a country where the finding out could happen. Which brings me to Reagan. Not the government is the problem line everybody uses. A different day. January 28th, 1986. The challenger comes apart 73 seconds off the pad on live television in front of a nation of school children who are watching because their teacher was on board. Reagan tears up the State of the Union and goes on air that evening. And what he says to those kids is not that it was safe, or that it was worth it, or that somebody will be held accountable. He tells them this. He tells them that this is what exploration is, that taking a chance is part of it. And then he says the future doesn't belong to the faint hearted, it belongs to the brave. He said that on the worst day the program ever had. He did not know what came next. But he still said go anyways. That is not recklessness. Reagan didn't say the future belongs to the careless. He said it belongs to the brave. And brave means you can see the risk clearly and you move anyways, because you understand that the alternative is a country that never leaves the ground and calls it safety. We did not win the internet because one party controlled the future. We won because for one extraordinary moment both parties agreed not to. It's time for an AI truce. And America. That was common sense. Stick around for about ten seconds while I take all the text that came in as part of our after hours QA. Alright, let's take these questions. Marcus 4417. Be honest, did an AI write tonight's show? No, but one helped me research it, and I'd be a hypocrite to dodge that after the hour we just had. Here's my rule. It reads faster than I do. It finds the vote counts and the filings, and it argues with me, which is the useful part. Then I decide what's true and say the words. Nothing goes on air I didn't put there. Late night hosts have twelve writers. I've got a very fast research assistant and a ranch doing the same job. Priya. Hey Priya. Priya says you spent an hour defending section 230, and half your audience wants it repealed. Explain that. Yeah. And I heard it in my own head while I was writing it. So let me split it. 230 says a platform isn't the author of what you post. That's just true. And if you kill it, every website with a comment section either turns off comments or turns into a swamp. Those are the only two outcomes. The actual complaint you're referring to isn't with that. It's with platforms deciding which Legal speech gets seen. That's a real fight, and I'll take that fight all day. But it's a different fight, and I'm angry at Facebook is not a reason to blow up the rule that made the internet possible. Fix the thing you're mad about. Don't burn the foundation because you don't like the tenant. Tobias nine one five three. You mocked Exxon for wanting to protect kids. Some of us have kids. Great, Tobias. I didn't mock him. And you're not wrong to be watching for that. Exxon looked at what a modem could reach in 1995 and wanted it away from children. That instinct is correct. I share it. What I said is that his method failed. He wrote a law so broad the Supreme Court threw it out inside sixteen months, and at the end of it, kids were exactly as exposed as they'd been before. You got a feeling into the statute books, and it accomplished nothing. You want a rule that protects a nine year old? Write the narrow one that survives court after the foundation is built. That's not me going soft on it. That's me wanting it to work. Great question. All right, next one. Renee 6820. Have you actually called Wyden's office? No. Guy's half dead. I think he's on a breathing machine or something. And I sat with that for about a day because it's the obvious thing to throw back at me, and it should be. Honest answer. I'm a guy with the show. He's got a hundred million things in front of him, and I figured a phone call for me goes into a log somewhere, which is exactly the excuse everybody in Washington is using. So I don't love how that sounds as it's coming out of my mouth and I'm thinking about it. But ask me again in a month. That's what I do here. Raw honesty. Sorry. All right, Grant ending in 3374. You said you don't know what AI becomes, then why should I take your seven-point plan seriously? Nice try. It's a fair one, so I'll take it, but you're not going to win this one. My plan doesn't require me to know. That's the design. Not one of those seven pieces depends on a prediction being right. Compare that to what's on the table right now. Every serious proposal in Washington asks you to legislate a forecast. This capability is dangerous. That job disappears. This timeline holds. Those all fall apart the second the forecast is wrong. A rule about who's responsible when something goes wrong doesn't care what the technology turns into. It works in the good version and the bad one. I'm not asking you to trust my crystal ball. I'm telling you I don't have one and I wrote a plan for people who don't. Thank you. Great pushback. Keep it coming. Try next time. Hollis 5109. China. You didn't say the word once. Yeah. I'm glad you caught it, and I did it on purpose. Because China has become the button people press when they've run out of argument. Both sides do it. Want a subsidy? China. Want a ban? China. Want to skip a permit? China. It's not a fake concern, it's a real one, but it's being used as a substitute for thinking, which is the same disease as the physics wearing a flag. And here's the thing nobody wants to hear. If the race is real, the answer is the same one I gave you. The country that lets people build the fastest wins the race. Certainty first is not a national security strategy. It's a way of losing politely like Europe did. All right, great question. Next one, Danielle 7742. What if the anthropic guy is right and half of entry-level jobs are gone in five years? Does your truth survive that? Honestly? I guess that's the question that keeps me up. And the answer is that the truth survives it, but it doesn't solve it. Those are different things, and I'm not going to blur them for you. The truth is about who writes the rules and how. It's not an employment policy. If we lose half of entry-level white collar work in five years, this country needs a real conversation about what a twenty two-year-old does for a living. And none of my seven points answer that. What I'd say is this. Every version of that problem gets worse if the technology gets built somewhere else and we're renting it. There's no scenario where America loses the industry and keeps the jobs. But no, I'm not going to sit here and tell you the tax bill fixes your kid's career. It doesn't. Alright, next one, Abram 2695. What's the first thing that should actually pass, Chad? The boring one. Number four. Write down who's responsible. Is it the company that builds the model, the company that deploys it, the outfit that packages it into a product, and the person who used it? Four roles, four levels of liability on paper. It's not a headline, nobody's going to fundraise off of it, doesn't have a villain in it. Every argument we're having right now, the copyright fight and the safety fight and the open source fight is an argument about that. And nobody will say so. Because liability allocation is a phrase that generally clears a room. Pass that one, and the rest gets easier. Skip it, and we'll be having this exact conversation in 2032 with better graphics. All right, guys, that's the show. Go text somebody the thing about Nebraska and the golf clubs, and I'll see you next time. God bless you, President Reagan. And as always, may God save America.

SPEAKER_01 49:15

It's marvelous that you should care for me. Yes, you've made my life so glamorous. You can't blame me for feeling amorous. It's wonderful. It's so marvelous that you should care for me. It's wonderful, it's marble dance, story as it's glamour.