September 12, 2026 • 36 mins
Hour 2 of the Clay Travis and Buck Sexton Show shifts away from the Lindsay Clancy trial discussion that dominated the opening hour and focuses primarily on the rapidly evolving world of artificial intelligence (AI), technology innovation, robotics, automation, national security, and the global AI race between the United States and China. The hosts briefly touch on Senator Ted Cruz’s appearance on The View, discuss upcoming midterm election activity and political advertising strategies, and then devote most of the hour to a wide-ranging conversation about the future of AI and its potential impact on society, the economy, warfare, and everyday life.
A central topic throughout Hour 2 is the emergence of advanced AI systems and humanoid robotics. Clay and Buck discuss the growing capabilities of Tesla’s Optimus robots and the possibility that consumers could soon have robots performing household tasks such as folding laundry and completing chores. The conversation expands into broader questions about automation, machine learning, artificial intelligence development, and the transformative impact AI may have on the workforce, productivity, and daily life. The hosts describe society as being in the midst of an AI revolution and compare the current technological moment to major historical technological breakthroughs.
The discussion intensifies after the hosts examine a viral statement from a former AI researcher who warned that the race toward artificial superintelligence could pose serious risks to humanity. Clay and Buck analyze claims that future AI systems could become capable of self-improvement, autonomous decision-making, advanced cyberattacks, scientific breakthroughs, and even acquiring real-world power. Referencing popular science fiction concepts such as Skynet from the Terminator franchise, they debate whether fears of AI-driven extinction are legitimate concerns or exaggerated predictions designed to attract attention. The hosts explore arguments from both AI skeptics and AI optimists, discussing concepts such as superintelligence, autonomous systems, artificial general intelligence (AGI), machine autonomy, and existential risk.
Another major focus of Hour 2 is the intersection of AI and modern warfare. Clay and Buck discuss reports of autonomous drone technology being utilized in combat zones, examining how AI-powered targeting systems could operate without direct human involvement. The conversation centers on the ethical and military implications of autonomous weapons, facial recognition technology, AI-controlled drones, battlefield automation, and the future of defense technology. They explore scenarios in which AI systems could independently identify and engage targets, raising questions about accountability, military strategy, and global security.
Clay and Buck also examine the critical role of energy production, data centers, and computing power in the future of artificial intelligence. Clay argues that the ultimate limits of AI may be tied to access to affordable and abundant energy, while Buck emphasizes the enormous computing requirements needed to develop and operate advanced AI systems. Together, they discuss whether future AI breakthroughs could unlock new energy technologies and how energy infrastructure may become one of the most important factors shaping the next generation of technological innovation. Themes of energy security, computational power, technological scalability, and infrastructure investment are recurring throughout the hour.
China's technological rise becomes another significant topic. Multiple callers and the hosts discuss concerns that China may be rapidly advancing in AI, technology, and strategic innovation. Clay and Buck argue that the United States cannot simply halt AI development because geopolitical rivals would continue pursuing breakthroughs. They compare the AI competition to historical military and nuclear technology races, suggesting that leadership in artificial intelligence may become one of the defining strategic advantages of the twenty-first century. The conversation frequently returns to themes of global competition, innovation policy, national security, technological leadership, and the future balance of power between the United States and China.
Listener calls add further depth to the discussion. One caller working in AI argues that many apocalyptic warnings are being overstated by major AI companies seeking greater regulatory influence. Another discusses concerns about censorship, centralized control of information, and the possibility that government regulation could limit the development of open AI platforms. Additional callers examine the impact of AI on software development careers, workplace automation, and changing labor markets. A software developer explains how AI tools are already reshaping hiring practices and reducing demand for entry-level programming
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
All right, second hour. Clay and Buck kicks off now.
Thanks for being here with us, everybody. And as you know,
we talked first hour about the Lindsay Clancy jurors who
have come out. We'll continue to take calls and talk
backs on that as we go through the show. But
I didn't want to do... Clay and I didn't want
to do three hours today of really running through the
same conversation. We did a lot of it yesterday. So
with that in mind, I would say we should talk
(00:25):
about some of the other big things happening in the news, too.
Speaker 2 (00:27):
One....
Speaker 1 (00:29):
We'll get the audio of this. It seems that our
friend Ted Cruz may have just made an appearance on
The View, which brings up a lot of things. First
of all, a little bit of FOMO, because Clay and
Buck haven't been on The View yet, and we have
generously offered. We will go together. We will go separately, whatever,
(00:51):
but we want to go to The View at some
point and talk to those ladies and just help them
understand that Ladies of The View, help us help you.
That's what we're really offering here. Jerry Maguire also holds up.
It's a great movie. I would say, I know you're
a Jerry Maguire guy. Oh, it's a great movie. Great movie.
(01:13):
Holds up. But Ted Cruz has been on The View.
Something else, though, that is on our radar. We've got
those midterm elections. That's going to be heating up day
by day as we get closer to it. Some interesting
updates on Trump's PAC spending and where that's going. And
a lot of money now going to be supercharging the
media buys. But something I want to talk about. Clayton,
(01:34):
I want to talk to you about this. As you know,
we are on the AI revolution wave right now. It's
like we're all on the surfboard. The wave is going.
It hasn't crested yet. And we don't know when it's
going to, if it's going to crash down or if
it's going to be a nice smooth ride. But we
(01:55):
do know that it's happening right now. Optimus robots, for
those of you who haven't seen, you can go check
out videos of these things. They are a bit like
what you would think the Jetsons robots would be like
in real life. They can fold laundry. They can do
household tasks. The first Optimus Gen 3 robots, they're going
(02:17):
to be taking orders for them or public sales will
begin the second half of next year. It's a conditional timeline,
but that's what they think. Not that far off from
you having a robot doing chores in your home. If
you want one and you have, I don't know, like
30 grand is, I think, what they're going to cost,
something like that, they're going to be a little pricey
to start. But then there's also this other thing that
(02:39):
is happening in the background. You've got robotics, yes. You've
also got the AI race, which brings in the data
centers for sure. the data centers conversation, which we've had
here on the show, but just AI philosophically as a technology,
artificial intelligence. And Clay, let's dive into this together. You
keep seeing this from people at these companies. This is
(03:00):
not the first time. It is the warning of impending
doom from AI. Call it the Skynet effect. You all
remember Skynet from the Terminator series. That's the cybernetic or
the cyberintelligence That becomes self-aware and starts a war, a
nuclear war with Russia, and then it kills all of us, too.
(03:24):
So and then, you know, people are fighting in like
a post-apocalyptic future. Jacob Coxon wrote this on X and Clay.
I'm just going to do a few. It's very long,
but I'll do a couple of these. And then I
want you to tell me if you buy into this.
Is this something that you're worried about? Sky, he's a
senior guy at Anthropic. He wrote, I resigned from Anthropic today.
(03:47):
I spent the last three years doing pre-training research at
both OpenAI and Anthropic. Neither company. is acting responsibly. They
are racing straight to self-improving superintelligence and gambling with our lives.
More thoughts below. By the way, this has 94 million
(04:08):
views on X so far. Are people paying attention to this? Yes,
yes they are. Do not underestimate, he writes, the power
of this technology. These will soon be superhuman systems that
can hack anything. revolutionize any field overnight, and acquire real
power and resources. We have all witnessed the progress in
(04:29):
each of these domains, and progress is not slowing. The
people building AI earnestly believe that it could kill us
all by the end of the decade. This is not
a marketing stunt. If anything, many executives and senior researchers
will couch their phrasing in the press to sound sensible.
But I hear the same people express fear privately. No
(04:51):
other human activity poses this level of danger. Clay, make
it make sense.
Speaker 2 (05:00):
My 18-year-old says that you and I should be talking
way more about AI because he agrees with everything that
that anthropic employee just said. He has been regularly, I
think the phrase haranguing his mom and me on this, regularly.
(05:22):
And what he is focused on is the idea that
when AI becomes able to self-replicate, in other words, when
it reaches post-human knowledge and is able to iterate itself,
that it is going to, and this is something that
they think is going to happen in the next year
(05:45):
or two based on the progress, that it then will
spiral out of control and we will not be able
to have any control over it. And you pointed out
an interesting story that happened that he was talking about
quite a lot, which is already in war, Russia has
trained autonomous drones to basically attack fuel depots. And the
(06:10):
concept behind that is they can send all the imagery
of what a fuel depot looks like, and then they
just send the drones out, and the drone knows it
has been trained. When you see a fuel depot blow
it up, three humans in Ukraine died in an autonomous
drone attack.
Speaker 1 (06:28):
But the key thing is there's no human in the
kill chain, they call it. There's no human being standing
over a button going, oh, wait, it's about to hit
a little old lady and kids. Abort the mission. It
just is completely self-driven and it goes through the kill
chain autonomously.
Speaker 2 (06:42):
You could theoretically tell an AI drone today, I think basically, hey,
go kill every deer.
Speaker 1 (06:51):
In a county. And the drones would just fly everywhere.
Speaker 2 (06:56):
And when they saw a deer, they could recognize it.
Speaker 1 (06:58):
They would immediately kill it.
Speaker 2 (07:00):
When I saw the autonomous drone story, my first thought was, well,
you could clearly train drones to just go kill any
human without any interaction with a human at all. If
you see a human age, whether they're combatants or not.
Speaker 1 (07:13):
They could do facial recognition. They could do it based
on the they could they could send a drone out
and say, here are the parameters of where we think
you'll find this person. Roughly loiter in the sky. Now, obviously,
there's like loitering time and there's some factors that come in.
But theoretically, they could send it. to an area, let's say,
of a city, Clay, or a town in eastern Ukraine,
(07:34):
and say, when you see Oleg so-and-so, here's his facial recognition,
blow him up. They could do that.
Speaker 2 (07:42):
So my question is, and this is for, we've talked
a lot about data centers, energy, everything else. Is it
possible for AI to iterate to a degree where.
Speaker 1 (07:54):
It provides its own power? That's the question I have.
Speaker 2 (07:59):
So this is where I because I've had this debate
with my son and he is he came in, I
don't know, sixth in the nation in high school debate.
Speaker 1 (08:08):
So he's pretty good at debating.
Speaker 2 (08:10):
And my question and I'm curious how many people out
there that are super smart on this, because in other words,
my response on, well, Skynet's eventually going to kill us
all is running data on AI is incredibly cost intensive.
And it requires a great deal of energy. We all
know that. We've been talking about it immensely. And this
(08:31):
is why they call it compute. It is your compute power.
This is now the word that everyone is using. So
if humans decided that we wanted to shut AI off,
is AI able to harness energy, which is expensive, such
that it could maintain its own self? And as AI
(08:53):
grows in complexity... the amount of compute power that it
requires to execute everything becomes more and more significant. And
it's possible, Buck, here's one answer. It's possible that AI
is so much smarter than we are that it's going
to unlock ways to harness energy that we can't even comprehend, right?
(09:15):
Instead of oil and gas, there's a solar capability. That's
my question, which goes to how scared we should be
To what extent can AI self-perpetuate and power itself going
forward if humans one day decided, you know what? This
AI has gotten pretty scary.
Speaker 1 (09:33):
Well, this is the Skynet problem. This is the Skynet
paradox from the Terminator series, which is in retrospect now
even more brilliant than people I think realized at the time.
It was very successful, which is that they tried to,
the phrase was, pull the plug on Skynet. And then
it launched nuclear missiles because we had autonomous AI basically defense.
It's like the Pentagon was running by a computer program.
We launched missiles at Russia, Russia missiles come back and
(09:55):
you have a nuclear, uh, nuclear, you know, Holocaust as
a result. Um, I'm not, I'm not nearly as concerned
about this stuff as these people who are doing the
AI research. And I know that's a little bit like saying, um,
I'm not worried about this surgery, no matter what the
surgeons say, but I want to, I can, I can
give you some reasons for that. Um, before I do
(10:15):
that though, and I, and I'm, I'm coming at it
with a self-deprecating understanding that like, You know, I'm not
a surgeon, but I do know if a surgeon is
lying to me. That's important. Jacob here is correct. Evan,
here's a guy that responded to that other tweet. He's
alignment science lead of Anthropic AI. Anthropic is one of
(10:36):
these top AI companies right now. These companies are worth
so much money that you can't even, I mean, they're
putting valuations on them, but these companies are going to
be the most valuable companies really, I think, ever. But
this guy who's the lead at Anthropic AI on this writes, Jacob,
the guy who said it could kill us all, is
(10:57):
correct here. We really do earnestly believe AI could kill
all humans. I personally think it is a greater than 10%
chance within the next decade. I believe Anthropic is trying
its best, but we do not yet have a plan
to solve alignment for superintelligence and are not clearly on track.
Speaker 3 (11:17):
To do so.
Speaker 1 (11:17):
28 million views on that, I might add. So these
are ultra viral. The whole thread was very widely distributed.
This is probably the biggest thread on the Internet right
now on X, certainly today and maybe for quite some time. Clay,
I just think that... There's always concerns with technology. There
were even concerns about the Internet. There were concerns about
(11:39):
Y2K people. The the doomerism. There were concerns about nuclear
nuclear fission.
Speaker 3 (11:46):
Right.
Speaker 1 (11:46):
Which I think is very, very, very valid to think, oh,
my gosh, if we can split the atom, what's going
to happen? And we did have we dropped two nuclear bombs,
as we all know. But the world didn't end and
we didn't go into nuclear holocaust. And I just think
that we will adapt. The world is going to change.
the world that we all live in. I don't mean
your kids, your grandkids world, the world that we live
(12:08):
in is going to change very rapidly. There will be some, uh,
what's the, I'm trying to think of something. It's not
absurdly euphemistic. Like growing pains is too euphemistic. There will
be some challenges from this for sure. But the notion
of humanity becoming extinct because of something that we have
created this, uh, deus ex machina mass extinction, Thanos moment, um,
(12:34):
I just think that a lot of people know that
this is an incredible way to get attention. And if
you're some nerd who's been working on AI models for
the last three years and you want to become super
famous overnight, it's easy to be like, it's going to
kill us all, maybe.
Speaker 2 (12:47):
I would also add, if we don't do it, China will.
So for everybody out there who's saying, boy, this feels
like a risk, right, if you knew there's a 10%
chance of humanity ending, if his calculus was correct, you
might say, that's a little bit too much of a risk,
I think, compared to the reward that we might get
(13:09):
from AI. The problem I think we've got is China's
not going to stop. So really, I think the one
way we have to think about it is you talked
about this. AI is basically our military.
Speaker 1 (13:21):
Yes.
Speaker 2 (13:21):
So would we allow China to accelerate military technology and
we just say, no, we're good sticking with what we
have now? I think the answer is clearly no. I
don't think anyone out there would say that's a decision
we should make. And so I don't think we have
a real choice here. Right now, one of the things
(13:44):
that always happens is when somebody gets a big competitive advantage.
they often then beg for regulation because it means that
others are not going to be able to catch up
with them. So you have a trillion-dollar company. A part
of me wonders whether the anthropics of the world and
the open AIs and all these other different AI companies,
(14:08):
part of the doomerism is, oh, we need the government
to get involved. Because when the government get involved, you
stifle innovation and basically in many ways freeze the competitive
situation so that the people who are in a competitive
advantage are able to stay there. Some of you probably
are experts in AI. I think this is a far
(14:29):
more interesting discussion than what do you think about data centers?
Where is AI taking us, to me, is the conversation
that should be happening, not, oh, my goodness, can you
believe they're building another data center? because the data center
is less of an existential threat, obviously, than the AI endgame.
Speaker 1 (14:53):
My quick analogy on this also is our enemies, notably
the Soviet Union, exploited fears over nuclear civilian and nuclear
weapons to try to get a competitive military advantage over us,
and were successful in doing so. And our nuclear technology
is far less advanced than it would have been And
I mean for nuclear energy, which we would have think
(15:16):
about all the stuff with climate change and CO2 and
all this we could have skipped over if we had
aggressively allowed nuclear or aggressively pursued nuclear energy in the
way that would have made sense. Instead, we have hobbled
ourselves because of fear. And I don't want us to
do that on AI. I think that's a good argument, too.
And it's happening quick. So I'm curious, 800-282-2882, if we
(15:39):
have any experts in AI. I'm not just talking about philosophers.
I'm talking about people actually involved in the specifics of AI.
One bad incident in a neighborhood can change everybody's perspective
when it comes to self-protection.
Speaker 2 (15:52):
Have you had a burglary? Have you had a robbery?
Maybe you felt safe and now you don't. There's a
company out there called Sabre that can provide families with
self-protection tools you can keep in your own home. Pepper sprays,
pepper projectile launchers. They also have, by the way, if
you're out on a hike, bear spray, mountain lion spray.
When you make a plan, include a Sabre Home Defense
(16:13):
Launcher that's powerful protection designed to help keep you safe.
The Sabre Home Defense Launcher delivers seven powerful impact or
pepper projectiles. Two more than most competitors, and that matters
when you need extra stopping power on an advancing intruder.
Home invasions happen all too often. Maybe you also have
got kids out there working. Maybe they're riding the bus.
(16:35):
Maybe they're off to college. When safety matters, maybe they're
just in a new city. America chooses Sabre, the number
one brand trusted by police and millions worldwide. Sabre is
how it's spelled. S-A-B-R-E. You can save up to 15%
on select Sabre bundles right now.
Speaker 1 (16:53):
That's S-A-B-R-E. SabreProtect.com. 844-824-SAFE. Hope for the best. Be prepared
with Sabre. If you truly care, pass along this common
sense to family and friends.
Speaker 4 (17:07):
Clay and Buck, owning the airwaves.
Speaker 2 (17:10):
Welcome back in, Clay and Buck. David in Ohio, you
got 30 seconds. A lot of you want to weigh
in on this AI thread. David, what you got for us?
Speaker 5 (17:20):
Yes, thanks for taking my call. I just wanted to
weigh in. I've been listening to your AI debate and
who's further advanced. And you rang in as China. We're
not going to do what the Chinese will. I just
came back from a business trip of a week and
a half and witnessing firsthand. I've been going there for
almost 15 years. Over the past five years, they've become
(17:42):
so technologically advanced. And particularly within this past year, it's
totally startling and amazing. And we had some debates within
our company, willing to wager that without question, the Chinese,
if we don't beat them to the punch on this,
we're likely to be ruled by them within my lifetime.
And I'm in my 50s.
Speaker 1 (18:02):
Thank you for the call. Good call. If the roof
over your head requires repair, you know it. It's your home.
What you don't know is who you can necessarily get
the job done with when you have to get it fixed.
That's why we're here for you. Erie Home. They've been
at this for over 50 years. Erie Home, they are
the roof specialist. They start their job on your property
with a free roof inspection, 25-point inspection they develop themselves.
(18:25):
And if you need roof help, Erie Home has the newest,
best materials in the business. They look great, but they
last longer. Your new roof from Erie Home comes with
a 50-year transferable warranty. Schedule your free inspection at eriehome.com
slash buck today. That's E-R-I-E home.com. eriehome.com slash buck. Get
(18:46):
up to $ 1, 200 in a discount. eriehome.com slash buck. This
discount is maximized at $ 1, 200. Valid on new roofing installation.
Only minimum purchase required. Restrictions apply. See rep for warranty
or promotional detail. Welcome back in here to Clay and Buck.
Turns out that many of you are fired up about
our AI, artificial intelligence discussion. Let's continue on with that.
(19:09):
I really do think, just to put it clearly where
I stand on this. I think AI is transformational. I
think it's going to be generally enormously wealth producing in
a whole range of ways. I don't just mean like
money in your bank account, although there's that, but also
(19:30):
just the improvement in goods and services. The amount of
time that you will get back in your life from
a lot of things that are wasteful I also, Clay,
think how long is it before we can have AI
just sort of do our taxes, by the way? Wouldn't
that be nice? Think about all the time you spend
just going through finding this paperwork, things like that. I
think it's going to be enormously beneficial. That all said,
(19:53):
there's going to also be some dislocations and some problems,
and it's going to be used. It's a very powerful tool.
It will be used in good ways and bad ways.
But I think that the people who do this thing of,
I'm worried... Oh, I'm an AI researcher, and here's my
threat on how it's going to kill us all, maybe.
Notice how it's always like when you ask them percentages, 10%, 5%, 1%. Okay, well,
(20:16):
so what do we do with that? What are we
supposed to do if there's a 10% chance that this
kills all of us? Does that mean we stop AI research?
They'll say, oh, we need to have more governors, if
you will. I don't mean that. I mean that in
like the– isn't it a governor? Yeah, a governor is
on a car to slow down the speed. Thank you. Yeah,
that's what I mean. We need more governors on AI.
(20:37):
We need more guardrails put in. You know what's really
going to be a mess? If you decide that you're
going to have Chuck Schumer in charge of AI legislation
or something like that. The whole thing, we've seen this
with the Internet already, the whole thing becomes a giant
mess the second the people who have no understanding of
(20:58):
this whatsoever are coming up with kind of willy-nilly regulation,
which is going to be based on what the donor's
want them to do anyway, and there's going to be
a lot of backroom dealing. So my thing on this
is we'll be fine. It's going to be overall, I think,
a positive. And I also think that China really hopes
that we act like we're scared of our shadow on
(21:19):
this one.
Speaker 2 (21:20):
Bernie Sanders has called for the government to shut down
all AI. So if you're out there and you're just
not an AI person, I think it's worth asking how
often is Bernie Sanders been right in his political career
about anything? So the fact that he has come out
and said the government should insist that AI stop.
Speaker 1 (21:44):
He means our government. China's not going to stop.
Speaker 4 (21:48):
Imagine.
Speaker 1 (21:48):
Sorry. Yeah, no, I was just going to say, Buck.
Speaker 2 (21:50):
So imagine if you take it away from AI and
just say, imagine if we said, hey, our military is
strong enough as it is. We're not going to continue
to evolve at all in the military. That would be
a point of lunacy. That is, I think, what would
rapidly happen is the caller. We've got a couple more
(22:11):
callers here who I think want to weigh in. It's
not a situation where we can just say, hey, okay,
we're going to stop doing this. Whoever gets to generative AI,
and that's basically AI that can run itself and self-perpetuate,
is going to control the world.
Speaker 1 (22:29):
Yes, and to the point, they don't teach this in
school at all, so for a lot of people, even
if you're a history person, you really love history, you
may not know this, but the Soviets ran a long-running,
what they call, the KGB called Active Measures Campaign. So
that's sabotage, propaganda, assassination, things like that, all fall under
Active Measures. They had an active measures campaign to hinder U.S.
(22:51):
civilian nuclear energy by playing on Western fears, funding and
steering overlapping peace slash environmental groups, and turning any accident
into an indictment of all nuclear power while building a
large nuclear sector, Clay. That's the key part here, everybody.
The Soviets in our country, oh, nuclear. And you're like, oh,
(23:14):
how will they do? They had organizations like the World
Peace Council, the U.S. Peace Council, the National Council of
American Soviet Friendship, Communist Party USA, and then all kinds
of journalists and individuals and plants all over America for
decades to make everyone think, if there's a nuclear plant
three miles from me, my kids are going to have
three heads, you know? This was the whole, oh, we're
(23:37):
all going to die in some conflagration, while the Soviets,
now they messed up with Chernobyl, obviously, but while the
Soviets were building nuclear plants as fast as they possibly could. See,
it's the same. This is what I'm saying. We have
seen this before, everybody. And I know right now there's
this thing of like, oh, well, people on the right
are just, they're just taking, really? Clay and I were
the guys who weren't listening to anybody, among others, but
(23:57):
weren't listening to anybody about COVID, the science, the shots,
the masks, all that stuff. But we're now taking orders from,
on on china is an ai threat i mean give
me a break like this is this is very obvious
if you think it through tracy it works in ai
and he wants to talk to us tracy's a guy
(24:17):
tracy what's.
Speaker 4 (24:18):
Up hey what's going on guys i watch y'all or
listen every day sometimes i watch on youtube I just
have been losing my mind because I know for a fact,
in my lab, I code bare metal, as they call it,
on a Google server. I build live, and the first
thing I do is lock down the login so that
(24:38):
all your equipment can't be hacked or stolen. Your API
keys can't be stolen. But I'm here to tell you,
to make an AI do what they say it's done,
like hacking Hugging Face, and you can confirm this is
what they did. They turn off the safeties. Then they
give it a live Internet access so there's something to hack.
And they want to scare us so they can say
(24:59):
we're the only ones who know how to really regulate this,
and we let big AI regulate it. You won't have
to worry about censorship. They have the ultimate in censorship
at that point.
Speaker 1 (25:13):
I agree with him, by the way. I think they're
playing us for fools and they actually want more power.
I presume that you are optimistic. Give us, in your mind,
an optimistic vision. If you were calling back in in
five years, right? So you're calling back in in 2031.
Speaker 2 (25:33):
What do you expect AI to have done? What will
it look like for the average person out there? How
will their life be different?
Speaker 4 (25:42):
I've spent a lot of time worrying over that. It's
one of two visions. If we allow the government and
big AI to regulate AI so that we can never
have uncensored AI with filters on it telling us what
we can know and what we can think, then we
will take the biggest tool for and giving it to them,
(26:06):
and we will never be able to. When they do
these document dumps of a million documents, we now have
something that if you have a harness that tells it,
only give me output that's verified by fact, you just
filtered out all of their propaganda. You have solved the
problem of the information age. Too much noise, not enough signal,
but not if we let them regulate it to where
(26:28):
guys like me can't build platforms that will tell you
the truth ever. It'll only tell you what big AI says,
and the filters let through, and that is the danger. Now,
the other side is we win, and AI is a
free market, and guys like me get to keep building.
The big guys have to serve the customer, but if
(26:49):
they get to lock us all in like they do
with their software, oh, my good Lord.
Speaker 1 (26:54):
Thank you for the call. Buck, what I would say, too,
is this is so important for AI. I'm with him,
by the way. I just want to say it was
a great call. Go ahead.
Speaker 2 (27:02):
Where AI is getting its information, you and I use Grok.
Elon Musk has said that Grok's goal is to be
maximally truth-seeking. Where the input is coming in for AI,
when they're using Reddit forums and when they are scanning
for as much information as they are... The output is
(27:24):
only as good as the input. And so I'll use
as an example that isn't very significant, OpenAI, when you
asked OpenAI initially who would win a high school boys
team or a WNBA champion, that OpenAI said the WNBA
team would win. Because I bet it's been coded to
believe in girl power, yay, we're not going to be sexist.
(27:47):
Grok said, oh, the high school boys team is likely
to win by 30 or 40%. After OpenAI got ridiculed,
they went back in and they changed the answer, and
now OpenAI says a high school boys team is likely
to win. But it's only as good as the parameters
that are put in place, and so if you have
restrictive parameters, the information that you're going to be able
(28:08):
to get out of it is going to be less effective.
I just say, I've replaced, which I thought, when would
this ever happen, Clay? I've replaced Google it with Grok it.
Speaker 1 (28:17):
I agree. I go to Grok as my first line of,
of search engine because i don't want to be served
a whole bunch of ads and i don't want i
don't want the google teams uh you know secret algorithm
propaganda and i no i want an answer from something
that i trust and people might say oh but now
you're trusting grok i trust grok more than google yes
(28:38):
yes i do i trust elon's company and what its
mission is more than i trust uh whatever they're doing
over it over at uh alphabet which is google's parent
company so jeralyn in uh california.
Speaker 3 (28:55):
Before in Los Angeles, I am a software developer and
I don't work on AI directly myself. I'm not a
machine learning expert or anything like that, but I do
use it in my work. I have a little story
for you. When my dad was in college in the 70s,
his school counselors told him and all of his classmates
that if they wanted a career that was future proof,
(29:15):
they should become punch card operators because at that time
computers were running on punch cards. And computers were obviously
the wave of the future. And they thought that that
would be a secure career. And then fast forward, you know,
50 years, about 15 years ago, I was learning to
become a software developer. And I thought my career was future-proof,
(29:36):
unlike my dad's punch cards of the 70s, because everybody's
going to need a website. Everyone's going to need software.
There's no way that this career is going anywhere anytime soon. And,
you know, it's now 10 years later, and AI has
basically taken away my job. Not entirely, but there's a
lot of hiring managers out there who would prioritize AI
over hiring a junior software developer because because they think
(29:59):
that it's cheaper and more efficient.
Speaker 1 (30:01):
You bring up an excellent point, can I just say,
that it had been a decade ago, the snide phrase
of the left, remember this, was learn to code for
people who, you know, have a trade, you know, work
with their hands. It's learn to code. Like, that was
the answer. Well, now coding is doing the coding. And
people that can do electrical wiring or build something that
(30:24):
people need or, you They're in demand, and software engineers,
the software is doing the engineering. Gerilyn, what do you
think about AI?
Speaker 3 (30:37):
So here's the thing. AI, as it is developed right now,
does not think. And Clay, you were just talking about it.
I think you and Buck have a pretty good grasp
of how it works, but I think the general population
does not have a good idea of how it works,
and especially the scaremongers out there don't. It does not think.
The best it can do is look at proximity of
(30:58):
words that other people have thought of and then generate
answers based on percentages, which is why AI hallucinates. So
I also use AI instead of Google Now a lot
to ask it just like general knowledge questions. And a
lot of the time it's going to know, but a
lot of the time it's also wrong. And I have
to push back and say, no, you hallucinated that. And
(31:18):
then it will admit that it made it up because
it's just looking at patterns and saying, well, this is
the most likely answer based on the patterns of language
that I've seen other people, like that I'm seeing in
the internet or from my training data. But all of
that training data and everything it's reading off the internet
was made by a human at some point. The point
(31:39):
at which AI is going to get dangerous, if it
ever gets to that point, is when it can genuinely
be creative, when it can think of things without having
to look at patterns that humans have produced first. And
I view this as kind of parallel to, like, the
(32:00):
Ehrlich scare of the 60s, the population bomb idea of,
you know, Ehrlich said that the world was going to
starve to death in 30 years because the population was growing.
But he failed to take into account improved agricultural methods. AI,
it could go either way. I think we're sitting on
the cusp. It could either become incredibly dangerous, just like
(32:22):
everyone is predicting. Maybe we will be able to teach
it to learn to teach it to actually think. And
then we'll be living in the iRobot future where the
Cylons are going to take over and we all have
to bow to our robot overlords and or think of
a way to outthink it, just like, you know, Hal.
And I'm quoting all of these sci-fi movies and books because...
(32:46):
as Shelley said, poets are the unacknowledged legislators of the world.
I think science fiction writers are pointing us in the
right direction. There was already an Isaac Asimov story called
The Gods Themselves, the premise of which was that we
live in a not-too-distant future where all of our energy
(33:08):
needs are taken care of by siphoning it from a
different from a parallel universe, and this allowed us to
live in perfect luxury because that energy in his economy
was the stopping point, right? So we had enough energy
to power AI, power robots, and then we could live
(33:30):
lives of luxury because we had machines to do everything
for us. I don't know if the problem is going
to be energy. I don't know if the problem is
going to be we won't actually ever be able to
train AI the way we're hoping to, to think for itself.
I don't know what the issue is going to be,
or maybe none of that will be an issue. We're
standing on the cusp right now, and it could be
(33:51):
either AI does become so powerful and takes over, or
maybe this will be like the Y2K scare, and it
turns out to be a big nothing burger in 10 years,
and we're all like, well, what were we so afraid of?
But we won't know until 10 years from now, and
that's only if we keep developing AI, and you guys
are spot on.
Speaker 2 (34:10):
Thank you for the call, Gerilyn. I think the most
interesting aspect of all of this is energy. Energy, energy, energy.
Can we unlock energy that is far cheaper than what
we have so far? Can AI help us do it?
And what will power the AI revolution? To me, it
all boils down to energy at its baseline level. Look,
(34:34):
yesterday I was in Washington, D.C. I met with Frank Siller.
I met with Larry, the top... ad buyer there. I
met with Frank Siller's family. Rachel Compost Duffy was there
interviewing him about his brand new book. We were at
the Hayes Adams Hotel overlooking the White House. Frank was
with the president yesterday at the White House. It is
a somber week. Friday will be 25 years since the
(34:58):
September 11, 2001 terror attacks. As the audience knows, Frank
started Tunnel to Towers to honor his brother, Stephen, one
of the 343 firefighters.
Speaker 1 (35:09):
Who died on 9-11.
Speaker 2 (35:11):
Stephen's son, Stephen Siller Jr., was just nine months old
when he lost his dad. He spent the last four
months bringing a 16,900-pound steel beam from the World Trade
Center to dozens of cities across the U.S. in the
Tunnel to Tower Steel Across America Tour. Yesterday, that beam
was in D.C., President Trump there for a special ceremony.
(35:34):
The Tunnel to Tower Steel Across America Tour concludes in
New York City at Ground Zero on Friday, September 11th.
The Tunnel to Towers Foundation is hoping that seeing a
piece of history up close will help Americans remember, learn,
and never forget. To support Tunnel to Towers like both
Buck and I do, join us in donating $ 11 a
(35:54):
month at T2T.
Speaker 3 (35:54):
Org.
Speaker 2 (35:57):
Help America to never forget. Additionally, you can make a
meaningful impact by donating car or land as well. That's T,
the number two, T.org. Laugh.
Speaker 5 (36:10):
Learn.
Speaker 1 (36:11):
Hang with the guys. Right there when you need them most.
Speaker 4 (36:14):
Clay and Buck, just preset them on the iHeart app.
Speaker 1 (36:18):
Welcome back in.
Speaker 2 (36:18):
We will be back for the final hour.
Speaker 1 (36:20):
We'll take your calls.
Speaker 5 (36:21):
800-282-2882.
Speaker 2 (36:22):
I have used AI to dive into the jury question
that many of you asked that we talked about earlier
in the program. How did it end up being nine women,
three men? What are the procedures for selecting a jury?
We will discuss that. We'll continue to take your calls
on AI. And Ben Shapiro of The Daily Wire will
(36:44):
join us at the bottom of the third hour. All
that coming your way. You won't want to miss it.
We will be hanging out with you for another hour
next week.
All right, second hour. Clay and Buck kicks off now.
Thanks for being here with us, everybody. And as you know,
we talked first hour about the Lindsay Clancy jurors who
have come out. We'll continue to take calls and talk
backs on that as we go through the show. But
I didn't want to do... Clay and I didn't want
to do three hours today of really running through the
same conversation. We did a lot of it yesterday. So
with that in mind, I would say we should talk
(00:25):
about some of the other big things happening in the news, too.
Speaker 2 (00:27):
One....
Speaker 1 (00:29):
We'll get the audio of this. It seems that our
friend Ted Cruz may have just made an appearance on
The View, which brings up a lot of things. First
of all, a little bit of FOMO, because Clay and
Buck haven't been on The View yet, and we have
generously offered. We will go together. We will go separately, whatever,
(00:51):
but we want to go to The View at some
point and talk to those ladies and just help them
understand that Ladies of The View, help us help you.
That's what we're really offering here. Jerry Maguire also holds up.
It's a great movie. I would say, I know you're
a Jerry Maguire guy. Oh, it's a great movie. Great movie.
(01:13):
Holds up. But Ted Cruz has been on The View.
Something else, though, that is on our radar. We've got
those midterm elections. That's going to be heating up day
by day as we get closer to it. Some interesting
updates on Trump's PAC spending and where that's going. And
a lot of money now going to be supercharging the
media buys. But something I want to talk about. Clayton,
(01:34):
I want to talk to you about this. As you know,
we are on the AI revolution wave right now. It's
like we're all on the surfboard. The wave is going.
It hasn't crested yet. And we don't know when it's
going to, if it's going to crash down or if
it's going to be a nice smooth ride. But we
(01:55):
do know that it's happening right now. Optimus robots, for
those of you who haven't seen, you can go check
out videos of these things. They are a bit like
what you would think the Jetsons robots would be like
in real life. They can fold laundry. They can do
household tasks. The first Optimus Gen 3 robots, they're going
(02:17):
to be taking orders for them or public sales will
begin the second half of next year. It's a conditional timeline,
but that's what they think. Not that far off from
you having a robot doing chores in your home. If
you want one and you have, I don't know, like
30 grand is, I think, what they're going to cost,
something like that, they're going to be a little pricey
to start. But then there's also this other thing that
(02:39):
is happening in the background. You've got robotics, yes. You've
also got the AI race, which brings in the data
centers for sure. the data centers conversation, which we've had
here on the show, but just AI philosophically as a technology,
artificial intelligence. And Clay, let's dive into this together. You
keep seeing this from people at these companies. This is
(03:00):
not the first time. It is the warning of impending
doom from AI. Call it the Skynet effect. You all
remember Skynet from the Terminator series. That's the cybernetic or
the cyberintelligence That becomes self-aware and starts a war, a
nuclear war with Russia, and then it kills all of us, too.
(03:24):
So and then, you know, people are fighting in like
a post-apocalyptic future. Jacob Coxon wrote this on X and Clay.
I'm just going to do a few. It's very long,
but I'll do a couple of these. And then I
want you to tell me if you buy into this.
Is this something that you're worried about? Sky, he's a
senior guy at Anthropic. He wrote, I resigned from Anthropic today.
(03:47):
I spent the last three years doing pre-training research at
both OpenAI and Anthropic. Neither company. is acting responsibly. They
are racing straight to self-improving superintelligence and gambling with our lives.
More thoughts below. By the way, this has 94 million
(04:08):
views on X so far. Are people paying attention to this? Yes,
yes they are. Do not underestimate, he writes, the power
of this technology. These will soon be superhuman systems that
can hack anything. revolutionize any field overnight, and acquire real
power and resources. We have all witnessed the progress in
(04:29):
each of these domains, and progress is not slowing. The
people building AI earnestly believe that it could kill us
all by the end of the decade. This is not
a marketing stunt. If anything, many executives and senior researchers
will couch their phrasing in the press to sound sensible.
But I hear the same people express fear privately. No
(04:51):
other human activity poses this level of danger. Clay, make
it make sense.
Speaker 2 (05:00):
My 18-year-old says that you and I should be talking
way more about AI because he agrees with everything that
that anthropic employee just said. He has been regularly, I
think the phrase haranguing his mom and me on this, regularly.
(05:22):
And what he is focused on is the idea that
when AI becomes able to self-replicate, in other words, when
it reaches post-human knowledge and is able to iterate itself,
that it is going to, and this is something that
they think is going to happen in the next year
(05:45):
or two based on the progress, that it then will
spiral out of control and we will not be able
to have any control over it. And you pointed out
an interesting story that happened that he was talking about
quite a lot, which is already in war, Russia has
trained autonomous drones to basically attack fuel depots. And the
(06:10):
concept behind that is they can send all the imagery
of what a fuel depot looks like, and then they
just send the drones out, and the drone knows it
has been trained. When you see a fuel depot blow
it up, three humans in Ukraine died in an autonomous
drone attack.
Speaker 1 (06:28):
But the key thing is there's no human in the
kill chain, they call it. There's no human being standing
over a button going, oh, wait, it's about to hit
a little old lady and kids. Abort the mission. It
just is completely self-driven and it goes through the kill
chain autonomously.
Speaker 2 (06:42):
You could theoretically tell an AI drone today, I think basically, hey,
go kill every deer.
Speaker 1 (06:51):
In a county. And the drones would just fly everywhere.
Speaker 2 (06:56):
And when they saw a deer, they could recognize it.
Speaker 1 (06:58):
They would immediately kill it.
Speaker 2 (07:00):
When I saw the autonomous drone story, my first thought was, well,
you could clearly train drones to just go kill any
human without any interaction with a human at all. If
you see a human age, whether they're combatants or not.
Speaker 1 (07:13):
They could do facial recognition. They could do it based
on the they could they could send a drone out
and say, here are the parameters of where we think
you'll find this person. Roughly loiter in the sky. Now, obviously,
there's like loitering time and there's some factors that come in.
But theoretically, they could send it. to an area, let's say,
of a city, Clay, or a town in eastern Ukraine,
(07:34):
and say, when you see Oleg so-and-so, here's his facial recognition,
blow him up. They could do that.
Speaker 2 (07:42):
So my question is, and this is for, we've talked
a lot about data centers, energy, everything else. Is it
possible for AI to iterate to a degree where.
Speaker 1 (07:54):
It provides its own power? That's the question I have.
Speaker 2 (07:59):
So this is where I because I've had this debate
with my son and he is he came in, I
don't know, sixth in the nation in high school debate.
Speaker 1 (08:08):
So he's pretty good at debating.
Speaker 2 (08:10):
And my question and I'm curious how many people out
there that are super smart on this, because in other words,
my response on, well, Skynet's eventually going to kill us
all is running data on AI is incredibly cost intensive.
And it requires a great deal of energy. We all
know that. We've been talking about it immensely. And this
(08:31):
is why they call it compute. It is your compute power.
This is now the word that everyone is using. So
if humans decided that we wanted to shut AI off,
is AI able to harness energy, which is expensive, such
that it could maintain its own self? And as AI
(08:53):
grows in complexity... the amount of compute power that it
requires to execute everything becomes more and more significant. And
it's possible, Buck, here's one answer. It's possible that AI
is so much smarter than we are that it's going
to unlock ways to harness energy that we can't even comprehend, right?
(09:15):
Instead of oil and gas, there's a solar capability. That's
my question, which goes to how scared we should be
To what extent can AI self-perpetuate and power itself going
forward if humans one day decided, you know what? This
AI has gotten pretty scary.
Speaker 1 (09:33):
Well, this is the Skynet problem. This is the Skynet
paradox from the Terminator series, which is in retrospect now
even more brilliant than people I think realized at the time.
It was very successful, which is that they tried to,
the phrase was, pull the plug on Skynet. And then
it launched nuclear missiles because we had autonomous AI basically defense.
It's like the Pentagon was running by a computer program.
We launched missiles at Russia, Russia missiles come back and
(09:55):
you have a nuclear, uh, nuclear, you know, Holocaust as
a result. Um, I'm not, I'm not nearly as concerned
about this stuff as these people who are doing the
AI research. And I know that's a little bit like saying, um,
I'm not worried about this surgery, no matter what the
surgeons say, but I want to, I can, I can
give you some reasons for that. Um, before I do
(10:15):
that though, and I, and I'm, I'm coming at it
with a self-deprecating understanding that like, You know, I'm not
a surgeon, but I do know if a surgeon is
lying to me. That's important. Jacob here is correct. Evan,
here's a guy that responded to that other tweet. He's
alignment science lead of Anthropic AI. Anthropic is one of
(10:36):
these top AI companies right now. These companies are worth
so much money that you can't even, I mean, they're
putting valuations on them, but these companies are going to
be the most valuable companies really, I think, ever. But
this guy who's the lead at Anthropic AI on this writes, Jacob,
the guy who said it could kill us all, is
(10:57):
correct here. We really do earnestly believe AI could kill
all humans. I personally think it is a greater than 10%
chance within the next decade. I believe Anthropic is trying
its best, but we do not yet have a plan
to solve alignment for superintelligence and are not clearly on track.
Speaker 3 (11:17):
To do so.
Speaker 1 (11:17):
28 million views on that, I might add. So these
are ultra viral. The whole thread was very widely distributed.
This is probably the biggest thread on the Internet right
now on X, certainly today and maybe for quite some time. Clay,
I just think that... There's always concerns with technology. There
were even concerns about the Internet. There were concerns about
(11:39):
Y2K people. The the doomerism. There were concerns about nuclear
nuclear fission.
Speaker 3 (11:46):
Right.
Speaker 1 (11:46):
Which I think is very, very, very valid to think, oh,
my gosh, if we can split the atom, what's going
to happen? And we did have we dropped two nuclear bombs,
as we all know. But the world didn't end and
we didn't go into nuclear holocaust. And I just think
that we will adapt. The world is going to change.
the world that we all live in. I don't mean
your kids, your grandkids world, the world that we live
(12:08):
in is going to change very rapidly. There will be some, uh,
what's the, I'm trying to think of something. It's not
absurdly euphemistic. Like growing pains is too euphemistic. There will
be some challenges from this for sure. But the notion
of humanity becoming extinct because of something that we have
created this, uh, deus ex machina mass extinction, Thanos moment, um,
(12:34):
I just think that a lot of people know that
this is an incredible way to get attention. And if
you're some nerd who's been working on AI models for
the last three years and you want to become super
famous overnight, it's easy to be like, it's going to
kill us all, maybe.
Speaker 2 (12:47):
I would also add, if we don't do it, China will.
So for everybody out there who's saying, boy, this feels
like a risk, right, if you knew there's a 10%
chance of humanity ending, if his calculus was correct, you
might say, that's a little bit too much of a risk,
I think, compared to the reward that we might get
(13:09):
from AI. The problem I think we've got is China's
not going to stop. So really, I think the one
way we have to think about it is you talked
about this. AI is basically our military.
Speaker 1 (13:21):
Yes.
Speaker 2 (13:21):
So would we allow China to accelerate military technology and
we just say, no, we're good sticking with what we
have now? I think the answer is clearly no. I
don't think anyone out there would say that's a decision
we should make. And so I don't think we have
a real choice here. Right now, one of the things
(13:44):
that always happens is when somebody gets a big competitive advantage.
they often then beg for regulation because it means that
others are not going to be able to catch up
with them. So you have a trillion-dollar company. A part
of me wonders whether the anthropics of the world and
the open AIs and all these other different AI companies,
(14:08):
part of the doomerism is, oh, we need the government
to get involved. Because when the government get involved, you
stifle innovation and basically in many ways freeze the competitive
situation so that the people who are in a competitive
advantage are able to stay there. Some of you probably
are experts in AI. I think this is a far
(14:29):
more interesting discussion than what do you think about data centers?
Where is AI taking us, to me, is the conversation
that should be happening, not, oh, my goodness, can you
believe they're building another data center? because the data center
is less of an existential threat, obviously, than the AI endgame.
Speaker 1 (14:53):
My quick analogy on this also is our enemies, notably
the Soviet Union, exploited fears over nuclear civilian and nuclear
weapons to try to get a competitive military advantage over us,
and were successful in doing so. And our nuclear technology
is far less advanced than it would have been And
I mean for nuclear energy, which we would have think
(15:16):
about all the stuff with climate change and CO2 and
all this we could have skipped over if we had
aggressively allowed nuclear or aggressively pursued nuclear energy in the
way that would have made sense. Instead, we have hobbled
ourselves because of fear. And I don't want us to
do that on AI. I think that's a good argument, too.
And it's happening quick. So I'm curious, 800-282-2882, if we
(15:39):
have any experts in AI. I'm not just talking about philosophers.
I'm talking about people actually involved in the specifics of AI.
One bad incident in a neighborhood can change everybody's perspective
when it comes to self-protection.
Speaker 2 (15:52):
Have you had a burglary? Have you had a robbery?
Maybe you felt safe and now you don't. There's a
company out there called Sabre that can provide families with
self-protection tools you can keep in your own home. Pepper sprays,
pepper projectile launchers. They also have, by the way, if
you're out on a hike, bear spray, mountain lion spray.
When you make a plan, include a Sabre Home Defense
(16:13):
Launcher that's powerful protection designed to help keep you safe.
The Sabre Home Defense Launcher delivers seven powerful impact or
pepper projectiles. Two more than most competitors, and that matters
when you need extra stopping power on an advancing intruder.
Home invasions happen all too often. Maybe you also have
got kids out there working. Maybe they're riding the bus.
(16:35):
Maybe they're off to college. When safety matters, maybe they're
just in a new city. America chooses Sabre, the number
one brand trusted by police and millions worldwide. Sabre is
how it's spelled. S-A-B-R-E. You can save up to 15%
on select Sabre bundles right now.
Speaker 1 (16:53):
That's S-A-B-R-E. SabreProtect.com. 844-824-SAFE. Hope for the best. Be prepared
with Sabre. If you truly care, pass along this common
sense to family and friends.
Speaker 4 (17:07):
Clay and Buck, owning the airwaves.
Speaker 2 (17:10):
Welcome back in, Clay and Buck. David in Ohio, you
got 30 seconds. A lot of you want to weigh
in on this AI thread. David, what you got for us?
Speaker 5 (17:20):
Yes, thanks for taking my call. I just wanted to
weigh in. I've been listening to your AI debate and
who's further advanced. And you rang in as China. We're
not going to do what the Chinese will. I just
came back from a business trip of a week and
a half and witnessing firsthand. I've been going there for
almost 15 years. Over the past five years, they've become
(17:42):
so technologically advanced. And particularly within this past year, it's
totally startling and amazing. And we had some debates within
our company, willing to wager that without question, the Chinese,
if we don't beat them to the punch on this,
we're likely to be ruled by them within my lifetime.
And I'm in my 50s.
Speaker 1 (18:02):
Thank you for the call. Good call. If the roof
over your head requires repair, you know it. It's your home.
What you don't know is who you can necessarily get
the job done with when you have to get it fixed.
That's why we're here for you. Erie Home. They've been
at this for over 50 years. Erie Home, they are
the roof specialist. They start their job on your property
with a free roof inspection, 25-point inspection they develop themselves.
(18:25):
And if you need roof help, Erie Home has the newest,
best materials in the business. They look great, but they
last longer. Your new roof from Erie Home comes with
a 50-year transferable warranty. Schedule your free inspection at eriehome.com
slash buck today. That's E-R-I-E home.com. eriehome.com slash buck. Get
(18:46):
up to $ 1, 200 in a discount. eriehome.com slash buck. This
discount is maximized at $ 1, 200. Valid on new roofing installation.
Only minimum purchase required. Restrictions apply. See rep for warranty
or promotional detail. Welcome back in here to Clay and Buck.
Turns out that many of you are fired up about
our AI, artificial intelligence discussion. Let's continue on with that.
(19:09):
I really do think, just to put it clearly where
I stand on this. I think AI is transformational. I
think it's going to be generally enormously wealth producing in
a whole range of ways. I don't just mean like
money in your bank account, although there's that, but also
(19:30):
just the improvement in goods and services. The amount of
time that you will get back in your life from
a lot of things that are wasteful I also, Clay,
think how long is it before we can have AI
just sort of do our taxes, by the way? Wouldn't
that be nice? Think about all the time you spend
just going through finding this paperwork, things like that. I
think it's going to be enormously beneficial. That all said,
(19:53):
there's going to also be some dislocations and some problems,
and it's going to be used. It's a very powerful tool.
It will be used in good ways and bad ways.
But I think that the people who do this thing of,
I'm worried... Oh, I'm an AI researcher, and here's my
threat on how it's going to kill us all, maybe.
Notice how it's always like when you ask them percentages, 10%, 5%, 1%. Okay, well,
(20:16):
so what do we do with that? What are we
supposed to do if there's a 10% chance that this
kills all of us? Does that mean we stop AI research?
They'll say, oh, we need to have more governors, if
you will. I don't mean that. I mean that in
like the– isn't it a governor? Yeah, a governor is
on a car to slow down the speed. Thank you. Yeah,
that's what I mean. We need more governors on AI.
(20:37):
We need more guardrails put in. You know what's really
going to be a mess? If you decide that you're
going to have Chuck Schumer in charge of AI legislation
or something like that. The whole thing, we've seen this
with the Internet already, the whole thing becomes a giant
mess the second the people who have no understanding of
(20:58):
this whatsoever are coming up with kind of willy-nilly regulation,
which is going to be based on what the donor's
want them to do anyway, and there's going to be
a lot of backroom dealing. So my thing on this
is we'll be fine. It's going to be overall, I think,
a positive. And I also think that China really hopes
that we act like we're scared of our shadow on
(21:19):
this one.
Speaker 2 (21:20):
Bernie Sanders has called for the government to shut down
all AI. So if you're out there and you're just
not an AI person, I think it's worth asking how
often is Bernie Sanders been right in his political career
about anything? So the fact that he has come out
and said the government should insist that AI stop.
Speaker 1 (21:44):
He means our government. China's not going to stop.
Speaker 4 (21:48):
Imagine.
Speaker 1 (21:48):
Sorry. Yeah, no, I was just going to say, Buck.
Speaker 2 (21:50):
So imagine if you take it away from AI and
just say, imagine if we said, hey, our military is
strong enough as it is. We're not going to continue
to evolve at all in the military. That would be
a point of lunacy. That is, I think, what would
rapidly happen is the caller. We've got a couple more
(22:11):
callers here who I think want to weigh in. It's
not a situation where we can just say, hey, okay,
we're going to stop doing this. Whoever gets to generative AI,
and that's basically AI that can run itself and self-perpetuate,
is going to control the world.
Speaker 1 (22:29):
Yes, and to the point, they don't teach this in
school at all, so for a lot of people, even
if you're a history person, you really love history, you
may not know this, but the Soviets ran a long-running,
what they call, the KGB called Active Measures Campaign. So
that's sabotage, propaganda, assassination, things like that, all fall under
Active Measures. They had an active measures campaign to hinder U.S.
(22:51):
civilian nuclear energy by playing on Western fears, funding and
steering overlapping peace slash environmental groups, and turning any accident
into an indictment of all nuclear power while building a
large nuclear sector, Clay. That's the key part here, everybody.
The Soviets in our country, oh, nuclear. And you're like, oh,
(23:14):
how will they do? They had organizations like the World
Peace Council, the U.S. Peace Council, the National Council of
American Soviet Friendship, Communist Party USA, and then all kinds
of journalists and individuals and plants all over America for
decades to make everyone think, if there's a nuclear plant
three miles from me, my kids are going to have
three heads, you know? This was the whole, oh, we're
(23:37):
all going to die in some conflagration, while the Soviets,
now they messed up with Chernobyl, obviously, but while the
Soviets were building nuclear plants as fast as they possibly could. See,
it's the same. This is what I'm saying. We have
seen this before, everybody. And I know right now there's
this thing of like, oh, well, people on the right
are just, they're just taking, really? Clay and I were
the guys who weren't listening to anybody, among others, but
(23:57):
weren't listening to anybody about COVID, the science, the shots,
the masks, all that stuff. But we're now taking orders from,
on on china is an ai threat i mean give
me a break like this is this is very obvious
if you think it through tracy it works in ai
and he wants to talk to us tracy's a guy
(24:17):
tracy what's.
Speaker 4 (24:18):
Up hey what's going on guys i watch y'all or
listen every day sometimes i watch on youtube I just
have been losing my mind because I know for a fact,
in my lab, I code bare metal, as they call it,
on a Google server. I build live, and the first
thing I do is lock down the login so that
(24:38):
all your equipment can't be hacked or stolen. Your API
keys can't be stolen. But I'm here to tell you,
to make an AI do what they say it's done,
like hacking Hugging Face, and you can confirm this is
what they did. They turn off the safeties. Then they
give it a live Internet access so there's something to hack.
And they want to scare us so they can say
(24:59):
we're the only ones who know how to really regulate this,
and we let big AI regulate it. You won't have
to worry about censorship. They have the ultimate in censorship
at that point.
Speaker 1 (25:13):
I agree with him, by the way. I think they're
playing us for fools and they actually want more power.
I presume that you are optimistic. Give us, in your mind,
an optimistic vision. If you were calling back in in
five years, right? So you're calling back in in 2031.
Speaker 2 (25:33):
What do you expect AI to have done? What will
it look like for the average person out there? How
will their life be different?
Speaker 4 (25:42):
I've spent a lot of time worrying over that. It's
one of two visions. If we allow the government and
big AI to regulate AI so that we can never
have uncensored AI with filters on it telling us what
we can know and what we can think, then we
will take the biggest tool for and giving it to them,
(26:06):
and we will never be able to. When they do
these document dumps of a million documents, we now have
something that if you have a harness that tells it,
only give me output that's verified by fact, you just
filtered out all of their propaganda. You have solved the
problem of the information age. Too much noise, not enough signal,
but not if we let them regulate it to where
(26:28):
guys like me can't build platforms that will tell you
the truth ever. It'll only tell you what big AI says,
and the filters let through, and that is the danger. Now,
the other side is we win, and AI is a
free market, and guys like me get to keep building.
The big guys have to serve the customer, but if
(26:49):
they get to lock us all in like they do
with their software, oh, my good Lord.
Speaker 1 (26:54):
Thank you for the call. Buck, what I would say, too,
is this is so important for AI. I'm with him,
by the way. I just want to say it was
a great call. Go ahead.
Speaker 2 (27:02):
Where AI is getting its information, you and I use Grok.
Elon Musk has said that Grok's goal is to be
maximally truth-seeking. Where the input is coming in for AI,
when they're using Reddit forums and when they are scanning
for as much information as they are... The output is
(27:24):
only as good as the input. And so I'll use
as an example that isn't very significant, OpenAI, when you
asked OpenAI initially who would win a high school boys
team or a WNBA champion, that OpenAI said the WNBA
team would win. Because I bet it's been coded to
believe in girl power, yay, we're not going to be sexist.
(27:47):
Grok said, oh, the high school boys team is likely
to win by 30 or 40%. After OpenAI got ridiculed,
they went back in and they changed the answer, and
now OpenAI says a high school boys team is likely
to win. But it's only as good as the parameters
that are put in place, and so if you have
restrictive parameters, the information that you're going to be able
(28:08):
to get out of it is going to be less effective.
I just say, I've replaced, which I thought, when would
this ever happen, Clay? I've replaced Google it with Grok it.
Speaker 1 (28:17):
I agree. I go to Grok as my first line of,
of search engine because i don't want to be served
a whole bunch of ads and i don't want i
don't want the google teams uh you know secret algorithm
propaganda and i no i want an answer from something
that i trust and people might say oh but now
you're trusting grok i trust grok more than google yes
(28:38):
yes i do i trust elon's company and what its
mission is more than i trust uh whatever they're doing
over it over at uh alphabet which is google's parent
company so jeralyn in uh california.
Speaker 3 (28:55):
Before in Los Angeles, I am a software developer and
I don't work on AI directly myself. I'm not a
machine learning expert or anything like that, but I do
use it in my work. I have a little story
for you. When my dad was in college in the 70s,
his school counselors told him and all of his classmates
that if they wanted a career that was future proof,
(29:15):
they should become punch card operators because at that time
computers were running on punch cards. And computers were obviously
the wave of the future. And they thought that that
would be a secure career. And then fast forward, you know,
50 years, about 15 years ago, I was learning to
become a software developer. And I thought my career was future-proof,
(29:36):
unlike my dad's punch cards of the 70s, because everybody's
going to need a website. Everyone's going to need software.
There's no way that this career is going anywhere anytime soon. And,
you know, it's now 10 years later, and AI has
basically taken away my job. Not entirely, but there's a
lot of hiring managers out there who would prioritize AI
over hiring a junior software developer because because they think
(29:59):
that it's cheaper and more efficient.
Speaker 1 (30:01):
You bring up an excellent point, can I just say,
that it had been a decade ago, the snide phrase
of the left, remember this, was learn to code for
people who, you know, have a trade, you know, work
with their hands. It's learn to code. Like, that was
the answer. Well, now coding is doing the coding. And
people that can do electrical wiring or build something that
(30:24):
people need or, you They're in demand, and software engineers,
the software is doing the engineering. Gerilyn, what do you
think about AI?
Speaker 3 (30:37):
So here's the thing. AI, as it is developed right now,
does not think. And Clay, you were just talking about it.
I think you and Buck have a pretty good grasp
of how it works, but I think the general population
does not have a good idea of how it works,
and especially the scaremongers out there don't. It does not think.
The best it can do is look at proximity of
(30:58):
words that other people have thought of and then generate
answers based on percentages, which is why AI hallucinates. So
I also use AI instead of Google Now a lot
to ask it just like general knowledge questions. And a
lot of the time it's going to know, but a
lot of the time it's also wrong. And I have
to push back and say, no, you hallucinated that. And
(31:18):
then it will admit that it made it up because
it's just looking at patterns and saying, well, this is
the most likely answer based on the patterns of language
that I've seen other people, like that I'm seeing in
the internet or from my training data. But all of
that training data and everything it's reading off the internet
was made by a human at some point. The point
(31:39):
at which AI is going to get dangerous, if it
ever gets to that point, is when it can genuinely
be creative, when it can think of things without having
to look at patterns that humans have produced first. And
I view this as kind of parallel to, like, the
(32:00):
Ehrlich scare of the 60s, the population bomb idea of,
you know, Ehrlich said that the world was going to
starve to death in 30 years because the population was growing.
But he failed to take into account improved agricultural methods. AI,
it could go either way. I think we're sitting on
the cusp. It could either become incredibly dangerous, just like
(32:22):
everyone is predicting. Maybe we will be able to teach
it to learn to teach it to actually think. And
then we'll be living in the iRobot future where the
Cylons are going to take over and we all have
to bow to our robot overlords and or think of
a way to outthink it, just like, you know, Hal.
And I'm quoting all of these sci-fi movies and books because...
(32:46):
as Shelley said, poets are the unacknowledged legislators of the world.
I think science fiction writers are pointing us in the
right direction. There was already an Isaac Asimov story called
The Gods Themselves, the premise of which was that we
live in a not-too-distant future where all of our energy
(33:08):
needs are taken care of by siphoning it from a
different from a parallel universe, and this allowed us to
live in perfect luxury because that energy in his economy
was the stopping point, right? So we had enough energy
to power AI, power robots, and then we could live
(33:30):
lives of luxury because we had machines to do everything
for us. I don't know if the problem is going
to be energy. I don't know if the problem is
going to be we won't actually ever be able to
train AI the way we're hoping to, to think for itself.
I don't know what the issue is going to be,
or maybe none of that will be an issue. We're
standing on the cusp right now, and it could be
(33:51):
either AI does become so powerful and takes over, or
maybe this will be like the Y2K scare, and it
turns out to be a big nothing burger in 10 years,
and we're all like, well, what were we so afraid of?
But we won't know until 10 years from now, and
that's only if we keep developing AI, and you guys
are spot on.
Speaker 2 (34:10):
Thank you for the call, Gerilyn. I think the most
interesting aspect of all of this is energy. Energy, energy, energy.
Can we unlock energy that is far cheaper than what
we have so far? Can AI help us do it?
And what will power the AI revolution? To me, it
all boils down to energy at its baseline level. Look,
(34:34):
yesterday I was in Washington, D.C. I met with Frank Siller.
I met with Larry, the top... ad buyer there. I
met with Frank Siller's family. Rachel Compost Duffy was there
interviewing him about his brand new book. We were at
the Hayes Adams Hotel overlooking the White House. Frank was
with the president yesterday at the White House. It is
a somber week. Friday will be 25 years since the
(34:58):
September 11, 2001 terror attacks. As the audience knows, Frank
started Tunnel to Towers to honor his brother, Stephen, one
of the 343 firefighters.
Speaker 1 (35:09):
Who died on 9-11.
Speaker 2 (35:11):
Stephen's son, Stephen Siller Jr., was just nine months old
when he lost his dad. He spent the last four
months bringing a 16,900-pound steel beam from the World Trade
Center to dozens of cities across the U.S. in the
Tunnel to Tower Steel Across America Tour. Yesterday, that beam
was in D.C., President Trump there for a special ceremony.
(35:34):
The Tunnel to Tower Steel Across America Tour concludes in
New York City at Ground Zero on Friday, September 11th.
The Tunnel to Towers Foundation is hoping that seeing a
piece of history up close will help Americans remember, learn,
and never forget. To support Tunnel to Towers like both
Buck and I do, join us in donating $ 11 a
(35:54):
month at T2T.
Speaker 3 (35:54):
Org.
Speaker 2 (35:57):
Help America to never forget. Additionally, you can make a
meaningful impact by donating car or land as well. That's T,
the number two, T.org. Laugh.
Speaker 5 (36:10):
Learn.
Speaker 1 (36:11):
Hang with the guys. Right there when you need them most.
Speaker 4 (36:14):
Clay and Buck, just preset them on the iHeart app.
Speaker 1 (36:18):
Welcome back in.
Speaker 2 (36:18):
We will be back for the final hour.
Speaker 1 (36:20):
We'll take your calls.
Speaker 5 (36:21):
800-282-2882.
Speaker 2 (36:22):
I have used AI to dive into the jury question
that many of you asked that we talked about earlier
in the program. How did it end up being nine women,
three men? What are the procedures for selecting a jury?
We will discuss that. We'll continue to take your calls
on AI. And Ben Shapiro of The Daily Wire will
(36:44):
join us at the bottom of the third hour. All
that coming your way. You won't want to miss it.
We will be hanging out with you for another hour
next week.
The Clay Travis and Buck Sexton Show News
Hosts And Creators
Clay Travis
Buck Sexton
Show Links
WebsiteNewsletterPopular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.