Is the Government Finally Stepping In? (Federal AI Regulation)

Trump just outlined his plan for AI regulation and it actually makes a lot of sense. I'm excited. Let me tell you

of sense. I'm excited. Let me tell you about it. I'll break it all down. Now,

about it. I'll break it all down. Now,

first, let me read his tweet. There must

be only one rule book. If we are going to continue to lead in AI, we are beating all countries at this point in the race, but that won't last long if we're going to have 50 states. Many of

them bad actors involved in rules and the approval process. That is the crux of what Trump and his administration are proposing, that AI regulation needs to be done at the federal level. And if

you've watched this channel at all, I agree. We do not need a patchwork of 50

agree. We do not need a patchwork of 50 different states coming up with their own regulation, often extremely partisan regulation trying to compete with each

other on that regulation. And thus,

startups get hurt. And I'll explain all of this in a moment. Let's continue.

There can be no doubt about this. AI

will be destroyed in its infancy. I

don't know if I should be yelling when I'm reading these all caps. I will be doing a one rule executive order this week. You can't expect a company to get

week. You can't expect a company to get 50 approvals every time they want to do something. That will never work. Now,

something. That will never work. Now,

that was the tweet. And David Saxs, the AI and cryptozar, the guy who works for Trump and advises him on these AI and crypto policy, basically followed up and

gave his legal and logical argument for why AI needs to be regulated at the federal level. So let's read this and

federal level. So let's read this and I'm going to explain what he means.

First, this is not an AI amnesty or AI moratorum. It is an attempt to settle a

moratorum. It is an attempt to settle a question of jurisdiction. What he's

saying is that they are not saying, "Hey, AI doesn't need regulation." They

are saying that the federal government should be regulating AI, not each individual state. And I already know

individual state. And I already know California has passed their own AI regulation. I know they've tried to pass

regulation. I know they've tried to pass even more. I know other states are

even more. I know other states are thinking about or have already passed some AI regulation. So, we're already heading in that direction. And this is their attempt to say, "No, we're going

to handle it." Now, here is his legal argument for why regulation, AI regulation needs to be done at the federal level. When an AI model is

federal level. When an AI model is developed in state A, let's say California, because usually it is trained in state B, let's say Texas, because there are major data centers in

Texas, inferenced in state C, that might also be Texas, it might be another state. That's where they're actually

state. That's where they're actually running inference on the GPUs and serving you your answers and delivered over the internet which is everywhere through national telecommunications

infrastructure that is clearly interstate commerce. So that is legal

interstate commerce. So that is legal argument number one and it would be difficult for me to imagine any counterarguments to that argument. It is

true artificial intelligence is already multi-state and it is being delivered through the internet which is in every single state and international and exactly the type of economic activity

that the framers of the constitution intended to reserve for the federal government to regulate. Now what if we didn't have federal level AI regulation?

What if we left it up to each individual state? Well, we would have a complex

state? Well, we would have a complex patchwork of AI regulation, legal regulation that really makes it almost

impossible for companies that are not incredibly wellunded, the top companies in the world, Google, Meta, OpenAI, Anthropic, companies that are not those

major companies, startups, the lifeblood of the economy of the United States to actually compete. It would make it

actually compete. It would make it impossible. Imagine as a startup, as a

impossible. Imagine as a startup, as a fivep person startup trying to build the next incredible AI model, the next incredible AI application, that you had

to build multiple versions of the model just to satisfy regulation between California and Florida and Texas and New York and all of these. It costs extra

money to train. It costs lawyer investment for them to go lawyer things and all of this makes it quite prohibitive to actually start a new

business. This is not good. It will not

business. This is not good. It will not allow us to compete with China. Think

about how China operates. There's

effectively one man who controls all of it. He just says the word and the entire

it. He just says the word and the entire economy shifts and changes into exactly what he says. Now, we definitely don't want that here, but if we went the

opposite route and allowed each state to choose their AI regulation, it would not make sense. And one of those AI startups

make sense. And one of those AI startups that is growing really quickly that I want to tell you about is the sponsor of today's video, Lindy. AI agents are no longer a developer's bestkept secret.

They are going mainstream. Wire just

published an article exploring companies that are almost entirely run by AI agents, examining what happens when not only the individual contributors, but the executives are essentially just AI.

And the company leading this transformation is Lindy. I've covered a lot of agent platforms on this channel, but what sets Lindy apart is just how easy they make it to build AI agents.

You don't need to be a developer. You

don't even need to understand complex workflows. You just tell Lindy what you

workflows. You just tell Lindy what you want to build. Three examples. Create a

social media spam monitor. Create a QA engineer agent for my app. And help me manage my personal CRM. From sales to support to operations, Lindy's agents

work 24/7 for you. And they're dead simple to set up so you can focus on growing your business, not on the grunt work. So check out Lindy. Thanks to

work. So check out Lindy. Thanks to

Lindy for sponsoring this video. I'll

drop all the links down in the description below. Now, back to the

description below. Now, back to the video. Now he finishes this argument

video. Now he finishes this argument with over 1,200 bills have been introduced in state legislatores and over a 100 measures have already passed.

So we're again as I mentioned earlier well on our way to having this crazy complex patchwork of AI regulation between states. Now let me give a very

between states. Now let me give a very specific example. the automobile

specific example. the automobile industry. California has very strict

industry. California has very strict laws on pollution from cars that was very different from other states. And

there's a reason for it. I grew up in Los Angeles and I remember in the '9s you couldn't see even a couple miles out because there was so much pollution.

There was so much smog. It was really bad back then. And then California passed laws that increased the threshold of what was required to get a smog check

passed and to have these vehicles be more efficient and capture more of that pollution. And then one day the skies

pollution. And then one day the skies cleared up. It was really very sudden

cleared up. It was really very sudden and it did work. But you're probably saying, Matt, aren't you making the argument against federal regulation?

Aren't you making the argument for states rights? And that is the

states rights? And that is the counterargument is states rights. And in

that example with automobiles, it actually made more sense. But the flip side of the California example is that automobile companies had to make two versions of their cars. They had to make

one for California and one for the rest of the states. That caused them to have additional cost, additional logistics.

And so with cars specifically, a car in California is emitting pollution in California. So it was a very California

California. So it was a very California specific problem. Obviously, cars are in

specific problem. Obviously, cars are in every state, but every state got to choose how much they cared or optimized for some other metric based on their pollution threshold levels. But in

California, we really cared. And so,

since everything about that car is based in California, California seems like it should have the right to regulate. But

here's the thing with AI. It is by nature, by definition, not only interstate, but international. And so,

as I mentioned before, AI is created here. It's trained somewhere else. It's

here. It's trained somewhere else. It's

inferenced somewhere else. It's

delivered throughout the entire world.

So, just by that nature, it seems like it negates the right of a state to regulate it since it's everywhere already. And so, what ended up happening

already. And so, what ended up happening as an automaker, you had a car that was built for California, then you had a car that was built for the other states.

Eventually, the automakers figured out, well, our cars that we sell to California, we actually sell the majority of cars to California because the population is so large. And thus,

maybe we'll just have a California car and we'll sell that California car to the rest of the country. California

asserted its opinion on what that threshold should be. And so, for cars, California basically decided for the rest of the country what their emission standards should be. Now, here's a

little difference between the car example and AI. With cars, there is a very clear metric to measure risk, and that is the emissions. Some states don't

care as much about higher emissions.

They care more about the cost to produce it because when you have additional regulation, you have higher cost.

California cared a lot and they didn't mind that it costs more. But with

artificial intelligence, the risk involved is actually much more nebulous.

It's not clear exactly what the risks are. We know some of them, but we

are. We know some of them, but we certainly don't know all of them. And of

course, we don't know what we don't know. And also, the profile of the risk

know. And also, the profile of the risk is quite different. With automobiles,

the risk is very specific to the location that the car is in. If a car is emitting pollution in California, it is a risk to California. with AI. It is not that way. If somebody's using AI in

that way. If somebody's using AI in California or it gets created in California, it can be a risk to the entire world. But again, bringing it

entire world. But again, bringing it back to having this patchwork of 50 different states determining what the regulation should be. And of course, we have blue states and red states and

purple states, and they're competing on regulation. What's important to them? Is

regulation. What's important to them? Is

it censorship? Is it bias? Is it

discrimination? Obviously, you probably know which states care about which issues. So, now back to David Sax's

issues. So, now back to David Sax's post. So, he gives a few examples. Of

post. So, he gives a few examples. Of

course, keep in mind David Sax is a Republican. He leans quite right. He

Republican. He leans quite right. He

tends to be pretty consistent about his views, but it is very right-leaning. So

he says, for example, states like Colorado, California, and Illinois have made AI developers liable if their models contribute to algorithmic discrimination, which is defined as

having a disperate impact on a protected group. Colorado's list of protected

group. Colorado's list of protected groups even includes English language proficiency. So presumably, it's against

proficiency. So presumably, it's against the law for an AI model to criticize illegal aliens. So obviously what he's

illegal aliens. So obviously what he's saying in a very partisan way is that blue states care about those issues whereas red states might not. And if all of these states are fighting to get

their issues front and center in AI regulation, that's going to be a big problem. He gives the example of this is

problem. He gives the example of this is how we ended up with black George Washington. I don't know if that's

Washington. I don't know if that's actually how we ended up with black George Washington. And he's referencing

George Washington. And he's referencing Google about 18 months ago. They had an AI model that said, "What does George Washington look like?" and it produced a black George Washington. Then he goes on

to say only a federal framework can achieve this goal. And of course here he makes a dig on Europe at best will end up with 50 different AI models for 50

different states. A regulatory morass

different states. A regulatory morass worse than Europe. And what is he referencing? Well, Europe is basically

referencing? Well, Europe is basically in stagnation because they have regulated everything. It is very

regulated everything. It is very difficult to get anything done in Europe right now. Obviously, there's a lot of

right now. Obviously, there's a lot of arguments for the way they're doing things as good, but that is my perception is it is very difficult to get stuff done in Europe right now. And

I think specifically one thing that I'm reminded of daily is the fact that I have to accept cookies on every website that I go to. That is GDPR. That came

from Europe. That came from overregulation.

It is absurd. It does nothing. And every

day my web browsing experience is negatively affected by the fact that I have this popup and I have to click it.

And it doesn't sound like a big deal, but imagine how many different times you've had to do that. And then multiply that by the billions of internet users every single day. Imagine how much time

is wasted clicking on that just because Europe wanted to regulate the whole world. And most importantly, if that

world. And most importantly, if that happens, China will race ahead. And also

as I said before, startups can't compete. This is regulatory capture and

compete. This is regulatory capture and that can happen both at the state level where multiple states are competing to get their regulation passed and at the federal level. If there's too much

federal level. If there's too much regulation at the federal level, we will also have regulatory capture in which the major companies, the major tech companies like OpenAI and Google are able to basically just capture the

market because startups can't compete.

They don't have the resources, both human capital and capital capital. But

what about the four C's? He says, let me address those concerns. Child safety, a very serious issue that of course we need to keep a close eye on. Preeemption

would not apply to generally applicable state laws. So, state laws requiring

state laws. So, state laws requiring online platforms to protect children from online predators or sexually explicit material would remain in effect. Okay, good. Communities. AI

effect. Okay, good. Communities. AI

preeemption would not apply to local infrastructure. That's a separate issue.

infrastructure. That's a separate issue.

In short, preeemption would not force communities to host data centers they don't want. He's specifically talking

don't want. He's specifically talking about there's been a lot of push back lately, a lot of nimism towards data centers. Of course, a lot of politicians

centers. Of course, a lot of politicians on the right have been saying, "Hey, anytime you set up a data center in my state, the electricity costs skyrocket for my constituents, and I don't want that." Next, something that's obviously

that." Next, something that's obviously very important to me, creators.

Copyright law is already federal. So

there is no need for preeemption here.

Questions about how copyright law should be applied to AI are already playing out in the courts. That's where this issue will be decided. Next, censorship. As

mentioned, the biggest threat of censorship is coming from certain blue states. Of course, remember he's very

states. Of course, remember he's very right-leaning. Red states can't stop

right-leaning. Red states can't stop this. Only President Trump's leadership

this. Only President Trump's leadership at the federal level can. Now, put

politics aside. I'm not sharing my opinion on this statement, but it is obviously extremely partisan. Censorship

is obviously not good. But he's saying only one side is doing the censorship, which is unlikely to be true. And then

he finishes with there is actually a fifth C, and that's competitiveness. If

we want America to win the AI race, a confusing patchwork of regulation will not work. So Trump is looking to pass an

not work. So Trump is looking to pass an executive order this week making it so that AI regulation is done at the federal level. So, what do you think

federal level. So, what do you think about this? Let me know in the comments.

about this? Let me know in the comments.

If you enjoyed this video, please consider giving a like and subscribe.