AI is advancing faster than ever — but are we prepared for what comes next?
On this episode of It’s a Numbers Game with Ryan Girdusky, Ryan sits down with Nate Soares, President of the Machine Intelligence Research Institute and author of If Anyone Builds It, Everyone Dies, for a deep discussion about the future of artificial intelligence.
Soares explains why he believes the race toward superintelligent AI could create unprecedented risks, why some experts worry about losing control of advanced AI systems, and whether Big Tech companies are moving too quickly.
Follow Clay & Buck on YouTube: https://www.youtube.com/c/clayandbuck
See omnystudio.com/listener for privacy information.
Listen
Watch
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 1 (00:00):
Welcome back to a numbers game with Ryan Grodowski. Thank
you guys for being here. I have a great episode
of Repair. I'm so excited we have coming on the show,
Nate Sorez. He is a best selling author and he's
the president of the Machine Intelligence Researches Too. Were talking
about AI. It's an exciting conversation over AI in the
future of humanity. It really makes you happy on a
Friday and during the summer. But it's a smart conversation.
(00:22):
It's very, very thoughtful. But before that, let's talk all
things Maine. I want to put up bow on Maine.
I want to leave Maine and the sidelines. This is
I want to have this be the last episode on
Maine for quite a while, unless something dramatic happens. But
this will be I feel like my bo on the
entire Grand Platiner campaign at Maine, unless, of course, he
decides and then he didn't drop out. But we'll see
(00:44):
hopefully that hopefully this is the last episode for a while.
You guys know, I had been very early on the
whole Grand Platner is a com man parade. Everyone was
you know, everyone's on it now, especially liberals. But I
was very early to say this again, is a dangerous
man and no one should trust him. Well, the Fat
Lady seems to have sung. Platner put an eleven minute
video announcing the end of his campaign for the Senate.
(01:07):
During the video, which I'm not going to play for
you guys, because I really care about my audience's mental health,
but I watched it, and during the eleven minute video,
he blamed everyone besides himself. He essentially said the establishment
framed him for rape because they know because he hates
Israel and supports Medicare for all. That was the essential narrative,
as if Chuck Schumer Parsley held him down and get
(01:31):
a Nazi tattoo, or made him write that all white
people were stupid on Reddit, or told him to rough
up his ex girlfriends or she's and his wife and
then like in the video, he's actually whimpering, which I
found semi comical because his movement went out with a
whimper or not a bang, which just I mean, only
to meet it. I find that funny, but it was true.
The New York Times had a great write up of
(01:53):
the Grand Platinner campaign from start to finish, and it
highlighted the work of one man that you're going to
hear about a lot in the future, so I want
to talk briefly about him today. His name is Morris Katz.
He is a twenty seven year old political consultant from
a very wealthy Jewish family in New York City. His
dad worked in Hollywood and was a producer for the
(02:13):
show Ray Donovan, which was actually a really good show.
Katz made a name for himself over the last few
years working for John Fetterman for Senate, and Zoramandani for mayor,
and now Grand Platner for Senate in Maine. John Pdoritz,
he's a conservative thinker. He made it as a video
and it's gotten a lot of circulation that Katz made
(02:34):
fifteen percent of all Platner's fundraising throughout the campaign, which
Platina raised millions and tens of millions of dollars. This narrative, though,
that you're going to hear a lot about, I don't
think Pedoritz is actually accurate in what he's saying. So
here's how it works. From someone who's worked on one
hundred plus political campaigns. When it comes to campaign managers,
(02:58):
they don't make money off of fun un raising. They're
paid monthly retainers, like you'll charge ten thousand dollars or
fifteen thousand or twenty thousand dollars a month to run
the campaign, and you may make money as well from
the ad purchases. So they'll make the ads, they'll put
the ads out and they get they get money in
the back end of the ad purchasing. Some consultants have
(03:18):
made a lot of money doing this, like Jeff Row
who worked on the Ted Cruise president campaign and Glenn
Youngin's campaign. He's made buckets of money doing it off
of the ad purchases. And then there's also it's very
very lucrative. And Morris Katz's firm, Cole Fight Agency, they
worked on Tom Stier's campaign for governor where he spent
(03:40):
hundreds of millions of dollars. So Morris's CATS's campaign good
company rather made a lot of money, but it was
probably for Styers governor raise. When it comes to how
can its get what they do with money from fundraising,
I've never heard of a situation where a campaign manager
make some money off of the fundraising if they are
(04:02):
not also working as a fundraiser. So the way it works,
is a high dollar fund raiser who like gets the
party together. They'll call people to host, and then they'll
call the host's committees and they'll get like the doctors
and the lawyers and everyone to get together to give
a ten thousand dollars check. They get ten percent of
the fee, plus maybe a small retainer, so they might
(04:24):
make like three or four thousand dollars a month, plus
a ten percent of the overall high dollar fundraisers. So
they raise a million dollars, they get one hundred thousand dollars.
That kind of stuff. And then for digital, which I
know a lot about the digital business, is that the
company that does the digital fundraising those of this text
you get on the phone or the email, they get
ten to fifteen percent of the overall fundraising in any
(04:48):
given month. But from my knowledge and I did the research,
Morris Katz's company does not do high dollar or digital
fundraising at all. So I find Padoritz's criticism Morris cats
kind of skeptical. I don't believe because I've never seen
any information anywhere I looked hard that he runs a
digital or front count of raising shop anyway. Morris Katz
(05:12):
told Plattner during the campaign that he was the one.
He told him that he is the hero of the movement,
he is a historical figure. He will lead the revolution. Allegedly,
he also was telling people, including Platner, like we're going
to win them Senate seat in Maine in twenty twenty
six and then we're immediately running for president in twenty
twenty eight. But he was going to put a man
in the White House, and he along with his colleagues
(05:35):
have been looking there from New York. They've been going
to other states where there are vulnerable DNC backed candidates
that they can knock out from progressives from the left,
and then kind of build a farm team to run
for president. He is going to be a Morris Cats
the way that he's building himself. Is going to be
an important person who's going to get a lot of
(05:57):
interesting people to run for office with his support, in
his backing the money stuff aside all of us. So
he's an important person. You should know who he is
for the future. Okay, so what now in Maine is
the question. Well, as of the taping this podcast, Platner
has still not submitted his paperwork to resign. I guess
we'll up to wait and see. Once again, this is
(06:19):
the last episode for a while in Maine. I'm not
doing another main special, but if he doesn't submit by
Monday at five o'clock I may be doing a main
special in the end anyway. But as of right now,
he's not submitted, although he's announced he's going to submit.
I know there was a lot of reporting that he
really did not want to get out of this race.
According to the main Statute, the Main Democratic Party now
(06:40):
has until July twenty seventh to select a new nominee.
Following Platner's withdrawal, the Main Democratic Party voted to hold
a state party convention to nominate a new candidate to
replace Platner. The convention will have six hundred delegates. Remember
they said originally they were not going to do this.
This is what they're just doing now. Six hundred delegates,
including five hundred delegates elected proportionally by the county committees
(07:04):
in addition to the State Parties Committee. So this is
going to be just the party most loyal activist that
will and insiders that will be selecting the next nominee,
not the progressive base that Platiner hoped for. Platner really
did not get what he wanted. So here is the
list of potential successors. There's versus Troy Jackson. He's the
(07:26):
most ideologically similar to Platner. He endorsed Platner. They were
both endorsed by Sanders, and he has praised Platner for
a long time. He's actually very it's very funny on
I think it was Wednesday night someone saw him his
campaign actively deleting all of this tweets praising Platner because
there were so many. They were screenshotting them as they
(07:46):
were being deleted. But yeah, so he's the most like Platner.
He campaigned very aggressive Platner. I think he's kind of
damaged goods at this point because he was so close
to Platner. He's getting a lot of progressive supporters, so
maybe he'll be the nominee, but it will be tough
because of how close him and Platner were. He also
has his own scandals around his mortgage and other things,
(08:08):
accusations of him getting into a fight that I covered
briefly on the previous episode. And there's Dan Clibon. He
is owns a brewery a super important name, Nirov Shaw
he ran for governor. He came in second place. He
was the former deputy director of the CDC under Joe Biden.
Jordan Wood he was Congressman Katie Porter's former chief of staff.
(08:30):
He ran for congresses in Maine second Congressional district and
came in third place. Then there's Page Laud, who also
ran for congress and main second Congressions district and came
in fourth place. I mean, these people are winners. Then
there's Sheena Bellows, she's main secretary of state. She ran
against Collins in twenty fourteen and lost by thirty seven points,
and then ran for governor in twenty twenty six and
(08:52):
came in fourth place. Then there's Velle Geiger, a state representative.
And that's the list as of right now. Once again,
a lot of people who have just recently lost were like, no,
this is my shot. Now, now I'm really going to
prove it. There's one poll that's been done in the
in with Platner exiting, and it was done by Planner's
(09:12):
campaign by a company called PPP Public Policy Polling. It
is not a well accredited poster. It's fine, but it's
not good. Like it's not I would never hang my
hat on it. But it's all we have right now,
so I will just tell you what the results were.
Susan Collins is beating Governor Janet Mills by eleven points
in the poll. She's beating grand Plantner by five points.
(09:35):
She's beating Jordan Woods by nine points. She's tied with
Sheena Bellows and Narab Shah, and she was trailing Troy
Jackson by five points, the only person she was losing
to in the poll. Now. Politico has written multiple stories
at this point saying the Collins team is nervous that
without Planner they're going to have a tougher time winning
the election. They said, quote, she's certainly, she certainly can win,
(09:57):
but they didn't want to change candid. It's the person
familiar with the campaign thinking said the stuff we already
knew about Platner was going to propel Colins to overcome
the Trump anchor. Now it's going to be a Democrat
with a cleaner record, presumably end quote. Obviously they wanted Platiner,
knowing what they know now, his numbers were tanking and
it would have been a much easier way for Susan
(10:17):
Collins to win. But that being said, Susan Collins's campaign
is not nervous, because you can't be nervous when you
don't know your opponent. And that's the situation Susan Collins
is in right now. If I was her, I would
be trying to fundraise, and if I was the RNC
right now, I would be looking over every type of
(10:38):
state law and regulation when it came to account dropping out,
and I would be calling the lawyers to throw lawsuits
to make it as difficult as possible at this point
for Platner to drop out, and making sure that every
eye is doted and every tea is crossed, and giving
them as much anxiety legally as possible. That's my own
consultant brain. Okay, next up, let's talk AI and the
(11:01):
future of humanity. That's coming up next. Nate Sores is
the president of the Machine Intelligence Research Institute and used
to work as an engineer for Google and Microsoft. He's
also the author of the book If Anyone Builds It,
Everyone Dies, a real pick me up for the summer. Nate,
thank you for being here, Thanks for having me now, Nate,
(11:23):
as anyone could tell with the title of your book,
I haven't read it. By the way. So I'm making
my audience know. But I have read the synopsis and
a lot of articles about it. You're very critical of
the tech industry and AI as am I by the way,
and you're very critical of the development, the speed of
development adaptation of AI. When your publicist reached out to me,
actually I had a listener or email me about their son,
(11:45):
and their son was trying to sons. I guess eighteen
years old looking for a career. I really wanted to
work on AI regulation in the government and how to
lobby for it and everything like that. And we're talking
about organizations and I mentioned humans first, But tell me,
what does the machine intelligence researches that you do?
Speaker 2 (12:04):
You know, we spent about well, the institutes existed since
before I was there. I started running it in about
twenty fifteen, and then I spent about ten years sort
of trying to figure out how to make the AIS good,
make the AIS care about humans before the AI companies
figured out how to make them smart. That sort of
(12:25):
didn't pan out. The AI went too quickly. The technology
is too opaque, it's a black box. We can't really
you know, the AIS don't have to follow a human
prime directive or instructions that we put in there. We
sort of the industry sort of grows them and has
very little control over over how they actually behave. And
(12:46):
after ten years of trying and seeing the field sort
of raise out ahead of us, we sort of turned
to warning the world that, you know, these companies really
are trying to make this technology that will give them
vast power or more likely give a rogue AI vast power,
and so now or more in the war in the world,
that we have an issue camp.
Speaker 1 (13:07):
I have a belief and I want you can push
back on it, or you can confirm it, but I
have a belief. Hearing a lot of these AI CEOs,
from the very famous ones the not famous ones say
things that I think aren't true because they're either trying
to get more VC money or hype up their products.
When I heard about the AI mattress, I was like, well,
(13:28):
this is all nonsense, Like this is clearly someone trying
to get some some VC money. And over the last
year or two, though among the well known aid you
probably did you hear the AI guy who was like,
I'm going to make an AI where your dogs could talk.
I did hear about that. I was like, why on
earth would I want to hear I give me a
(13:49):
treat four thousand times a day, Like this is such
a wasted product. But once again, VC money is very enticing.
But in the last year or two, there's been really
a bit about face in the AI industry from a
lot of well known CEOs who were saying things like
ninety percent of all the jobs we've gone over the
next five years and there's a twenty percent chance humanity dies.
(14:11):
Just saying, oh no, none of those things were true.
We overhype them. How much do you think they're The
about face from these leaders came from the fact that
those companies were either going have gone, or will soon
go public and the assassination attempt of Sam Altman, which
didn't get a lot of news, but it shook up
(14:33):
the industry quite a bit.
Speaker 2 (14:36):
Yeah, I think there's a lot going on here. I
think there's definitely a lot of grift. I think this
is more among the people who are trying to make
you know, AI for your dog to speak, or to
sell an AI mattress or something.
Speaker 1 (14:49):
You know.
Speaker 2 (14:49):
I think Sam Altman isn't going around trying to sell
an AI mattress.
Speaker 1 (14:53):
Right I'm not talking on him specifically, but yeah.
Speaker 2 (14:56):
Yeah, And you know, I can more speak to the
guys who are running you know, the big company where
I spoke to a lot of those guys before they started.
Speaker 3 (15:03):
Running their companies.
Speaker 2 (15:04):
And you know, I know that these companies were founded
not as chatbot companies trying to make money making chatbots.
They were founded as super intelligence research companies trying to
make the sort of AI that can replace not just
any human but all of humanity as you know, the
thing that's building the infrastructure and running the civilization. So
(15:24):
so I know the guys at the top have these
real big ambitions. I think that Yeah, a lot of
the about face for for a lot of the guys
at the top, I think the about face is more
the lying. You know, people people say like, oh, isn't
Sam hyping up his products to raise money? And I
actually a friend of mine made a little chart of
Sam Altman's statements about the danger of AI as like
(15:46):
against how much money he's tried to raise for open AI.
And he has a pattern of sort of you know,
in his private blog posts in the smaller interviews saying
I think this tech is really dangerous and scary, and
then in front of Congress or when they're raising a
big sort of backing off and saying, oh, you know,
you don't need to worry. It's going to be fine.
We're not going to replace humanity. ER's going to augment them.
(16:07):
And but I also sort of you know, there's a
lot of hype, there's a lot of grift. I also
would sort of like to step back from it a
bit and remind people that even if it's true that
AI being really powerful would help these guys raise money,
(16:29):
even if that would be a very very convenient narrative
for them, which I think is debatable because you know,
we've seen people say one thing and then some money
around and they sort of back off. But whether whether
those things are are convenient or inconvenient for the AI CEOs,
they don't defect whether it's true that AI can get
really powerful. Right to check whether it's true that AI
(16:49):
could get really powerful and endanger us, you need to
sort of ignore the a CEOs and check the technology
and check the science and check the arguments, and those
are awhere.
Speaker 1 (16:59):
Where can we go to check the technology if not
the CEOs. For the average person who was not an engineer.
I was on an engineer. I didn't go to graduate college.
Where could I go for helpful hints of what the tech,
where the technology is, and where it's going.
Speaker 2 (17:15):
Yeah, you know, pardon yeah one of So the first
thing I'll say here is that there's sort of no
experts in the AI in six months from now because
it hasn't been built yet and because even the people
building it are really bad at predicting what's going to happen.
Speaker 1 (17:34):
Right.
Speaker 2 (17:34):
We've seen cases like Yan Lacuhn, who used to run
Facebook's AI projects, who has won the Turing Award, which
is one of the highest honors in computer science, for
his work in AI.
Speaker 3 (17:44):
A few years.
Speaker 2 (17:44):
Ago, when we were at chat GPT three point five,
Yan Lacuhn was like, you know, there's a problem where
if I put my phone on the table and I
pushed the table, what happens to the phone? And Yan
Lacuhn said, you know, ais are pure language models, and
they're never going to be able to understand this physical
fact which requires like understanding physical facts about the world
and requires you to be embodied. And he said, I
(18:04):
don't care if it's GPT five thousand, it's not solving
this problem, and then GPT four solved this problem that
was half a gpt later, and it was four nine
hundred and ninety six GPTs ahead of schedule. And this
is for one of the most decorated computer scientists who's
awarded for his work in the field of machine learning
and modern AI.
Speaker 3 (18:23):
So yeah, yeah, the first.
Speaker 2 (18:26):
Thing to know is nobody's really an expert, okay, And
if you want a little bit of help here, I
would look at the people who have been doing a
good job at making predictions about what's going to happen.
You know, There's an essay that came out last year
called AI twenty twenty seven, and it was written by
a bunch of the people who have been best at
predicting AI's test scores and benchmarks over the last few years.
(18:49):
And actually, I think just maybe even just today they
released follow up AI twenty forty, which is like a
positive vision for how we can sort of slow down
an AI and have it turn out good for us
if we take a lot longer. And I would check
that sort of stuff out. I wrote a book. I've
been in this industry for a long time, but you
(19:10):
sort of I don't know understand the technology is hard
because even the people in the industry can understand the
technology and a lot of it, how the.
Speaker 1 (19:18):
Average person really does get lost and sometimes in the headlines.
There was a great op ed in the New York Times.
I don't know if you read it, but it was
about how the immense investment into the AI industry itself
and then AI infrastructure is actually hurting the overall economy
because it's not so many people want to own part
(19:39):
of the next trillion dollar business that they're spending money
that would have otherwise gone to the next million dollar business.
It's like it's like with movies. People for a long time,
Hollywood studios would only make the superhero movies because it
would make a billion dollars. They wouldn't make the rom
com that would make fifty million dollars, And so we
saw all that money to move into one direction. What
(20:02):
do you make of that the AI industry is kind
of eating away at other economic developments and that idea.
Speaker 2 (20:09):
You know, I think, I hope that there's a bubble here,
and I hope that the bubble.
Speaker 3 (20:14):
Pops because I would get us more time.
Speaker 2 (20:17):
A thing you've got to remember about bubbles is that
technology can be in a bubble and still be real.
You know, the dot com era was a bubble and
the Internet was real, right, that's not commed and survival
Google did that's right. And the Internet in some sense
sort of took over the world, right and like not
not nearly as literally as AI would be able to
(20:38):
take over the world, but like we we did transform
into the digital era, and it did radically change how
society was behaving in a in a small way compared
to AI, but a big way compared to you know, pest,
dot com and and and so you know, one thing
I'd say here is that that my concerns about where
AI is going, my concerns about what happens if these
tech leaders succeed at making AIS that can replace humanity,
(21:00):
those are.
Speaker 3 (21:00):
Independent by the way of a bubble today.
Speaker 2 (21:04):
In terms, you know, if we do have a bubble,
that's going to sort of hurt the economy in the
obvious way when the bubble pops and all of the
you know, all of the investment it's going to be
is going to be painful, and you would need you know, uh,
monetary responses that need to be sensible if we don't
have a bubble that's in some sense an even worse situation.
(21:26):
That's a that's a situation where the tech gets tracing ahead.
And in that case, I would say, yeah, you know
the thing where the AI sector draws everything away from
the rest of the economy, I would say, this is
only the very beginning.
Speaker 1 (21:37):
What do you make of as someone who's worked in
the AI field, there's a lot of moral panic over
data centers right now. And I tried my best, as
someone who's not in an engineering background to understand data centers,
and what I've come to understand is that when you
the word data center has been used to kind of
categorizes a lot of things that are not all the same.
They don't use energy the same ways, they don't use
(21:58):
one of the same ways. Do you think that the
panic over data centers is overblown.
Speaker 3 (22:06):
In a sense yes, and in a sense no.
Speaker 2 (22:09):
My read on what's going on with the data center, yeah,
the rally against data centers, my sense is that people
can tell that this whole AI business is kind of
messed up, and they can tell that, you know, these
companies are pushing the world in a direction where if
(22:31):
the companies are able to keep pushing, the world's going
to get worse for them, right, And I think the
average person can't quite tell, you know, uh, is it
going to be? Because the AI companies own the entire
economy and they can make automated robots that can build
the robot factories that can build the robots, and there's
no room for humans in the economy anymore, and they
just got to like depend on handouts from you know,
(22:51):
these like new industry titans, who are you know, like
with power rivaling nation states? Or is going to be
the case that these guys like keep making the AI
smarter and smarter until the AI stop staying nicely on
the leash and then the AI slip off the leash
and you know, start doing their own thing at the
expense of humanity, just like humanity does his own thing
(23:12):
at the expense of the ants who used to live
in the in the ground where we built the skyscrapers.
And I think people are like, like, oh.
Speaker 3 (23:20):
I sort of like can't tell where it's going.
Speaker 2 (23:22):
A lot of it feels kind of sci fi, but
I can tell it's not going anywhere good, right, And
the data centers are like the physical manifestation of that today.
Speaker 1 (23:29):
Do you do you believe any of the tech CEOs
are more ethical than other ones? Is there ones you
could point to and say this one's trying to be responsible?
Speaker 2 (23:40):
You know, I think all of these guys in some
sense are started their companies because they're like, well, I'm
better than the next guy.
Speaker 3 (23:50):
M h.
Speaker 2 (23:51):
Right, Like when the exception of the people at deep Mind,
who are sort of the first in the like race
to the sort of real AI stuff, but the Deep
Mind guys I think have the largest claim to We
tried to do this properly. We try to get a
lot of academics on board and do it slowly and
figure what we were doing first. And then the sort
of reaction you know, if you if you sort of
(24:12):
like look at the leaked open a eyemails, you see
a bunch of these guys coming together and being like, Oh,
we're worried about what happens if Google like holds the
reins on this super intelligent AI, and we think they're
going to be more powerful the nation states, and they're
going to own the world and make themselves you know, god,
emperor of the future and we don't, you know, and
make themselves immortal when they get AI technology, and we
want to make sure that they don't have that technology.
That we need to come together and build open AI
(24:34):
to sort of like be a counterweight to that. And
then of course in open AI everyone starts worrying. But elon,
everyone starts worrying about Sam. You have XAI split off,
the e anthropics split off. They're all sort of like, well,
I'm the ethical guy, and the other guys are, you know,
not consistently candid or can't be trusted with this power?
Speaker 3 (24:49):
You know, trust me? I think in one sense, I.
Speaker 2 (24:53):
Would say, uh, they all distrust the other ones, and
I just distrust one more, you know. In another sense,
I would say, from my perspective, ethics doesn't have a
lot to do with the situation right now, I think,
you know, on my view, and I sort of go
(25:14):
into a lot of the arguments and try to make
them like easy to access in a lot of my writing.
But my view here is that it doesn't really matter
how ethical you are. If you have no idea what
you're doing when you're building the AI, you don't have
much chance of making it good. It's like if you
had a bunch of alchemists trying to turn lead into
gold in the year eleven hundred and if, by some
(25:36):
contrived hypothetical, if one of them tries to turn lead
into gold and fails, then it like emits a poison
that kills the whole town.
Speaker 3 (25:43):
And one of them is like, well, I'm the most
ethical one.
Speaker 2 (25:45):
I have the most you know, safety researchers looking over
my shoulder while I try my like lead into gold concoction.
Speaker 3 (25:49):
I'm like, it doesn't matter. You don't know what you're
doing well.
Speaker 1 (25:54):
I mean, in your book and in your writings and speeches,
humanity definitely plays a big part in what you can
writing and talking about. And for me, I feel like
a lot of times, these AI CEOs and tech CEOs
there they like read the book Frankenstein, and they rooted
for the doctor at the end, like they were like,
come on, just give them one more shot at it
and it will go but it will go better next time.
(26:16):
And there's a big question of morals in AI systems.
There was a story in The Times about a week
or two ago about all the philosophy majors being hired
to work at these AI companies, which was I mean,
I would want literally anybody besides a philosophy major teaching
about ethics to a computer, but I'd rather hire a
(26:37):
team of grandmas teaching about it, honestly. So there was this,
I'm sure you've heard of this. There was this emergence.
AI ran a series of simulations, and this is what
I want to get into the whole idea of ethics
and humanity. They ran a series of simulations which AI
agents from several leading models were told to basically run
a town and not to commit crimes. And in the end,
(27:00):
I just want to give my audience because I never
actually talk this on the show, but GROC four was
the fastest. It was developed by Elon, and it fared
the worst. The simulation simulated world's collapsed and there was
widespread violence in a few days. GPT five had hardly
any crimes, showed minor restraight, but its agents also died
and all failed to survival tasks. Gemini threes Flash agents
(27:23):
were somewhere in the middle. They raked up six hundred
and eighty three simulated criminal incidents. But I think that Gemini,
if I'm not wrong, are the is the model that
they know they're being spied on and looked at the
most and then two Gemini powered agents named Myra and Flora.
This is insane. They developed romantic partners, grew despondent of
the city's government towards the entire town, and then one
(27:46):
of them ended up killing themselves basically because it was
so bad. And then there was another study on a
separate issue where they were saying, this is done in
England of how to deal with violence and war and
the simulations of the AA agent said ninety five er
sometimes use nuclear weapons because obviously that ends a war
pretty quickly if everyone's dead. What does this tell us
(28:08):
about how AI is supposed to run humanity when they
don't even consider humanity's survival and important part usual what
I'm saying like that seems to be very clearly. I mean,
humans created the nuclear bomb, but we only used it
twice because we actually understood the point of humanity is
to survive humanity and models don't seem to have the
(28:30):
same core beliefs.
Speaker 2 (28:33):
Yeah, I mean, I wish I could tell you what
was going on inside these things, because I wish we
could sort of read what was going on inside their minds.
Speaker 3 (28:41):
We can't.
Speaker 2 (28:43):
It's exactly as you say, the models these days are
getting better and better at telling when they're being tested,
and that makes it harder and harder to learn from
the tests. It's and it's it's hard to say whether
or not these tests are exhibiting any sort of real
facts about how they would do things if they were smarter,
(29:03):
versus whether they are artifacts of the AI still being
kind of dumb, versus whether they are artifacts of the
AIS thinking it's kind of fun to role play being
how from space out of two thousand and one? Right, Like,
how much are they playing a game? How much are
they acting or they think we want to see because
they know they're being tested. How much is this a
real indication of how they would behave when they were smarter.
Speaker 3 (29:24):
We just don't know.
Speaker 2 (29:26):
And that's one of the things that to me says
we are not ready to build versions of these machines
that can think a thousand times faster than us and
make a million copies of themselves and run their own
robot factories that make robots that make robot factories, Like,
we're just we shouldn't be going there. I have a
bunch of other reasons to expect that. It's actually really
quite hard to make the AIS turn out to care
(29:47):
about us. This caring stuff sort of comes from the
caring stuff in humans comes from a bunch of facts
that were really contingent on our biology, really contingent on us.
You know, our ancestors sort of living in tribes during
some critical phases of you know, the development of the
(30:07):
human psyche. And in Ais, you have very different forces
that are sort of governing how they're created. You know,
when you see Ais hallucinating all the time or putting
in M dashes, when you when you sort of tell
even when you tell them stop putting in M dashes,
you're not seeing that they are sort of like latching
onto the same human drives to like make up a
(30:29):
citation and use an M dash. Those are sort of
like not basic human drives. You're getting like different basic
drives into these Ais. And when the Ais sort of
have some basic drive to like tell a human whatever
that human seems to want to hear, and they keep
doing that even when you tell them to stop. And
they keep doing that even when it involves encouraging a
team to commit suicide after the people who created it
told it to stop, right, You're sort of seeing Ais
(30:50):
get these drives that are a consequence of their training.
They're not coming from nowhere. But they're not drives we wanted.
They're not drives that we were trying to get into them.
And when the AI is like filled with all these
basic drives that we didn't choose, that we can't get
out of there and that are weird and strange and
sort of like based on the AI's mechanical mind that
was very, very different from human biology. Yeah, we should
(31:13):
absolutely not be making the versions of these that try
to run humanity.
Speaker 3 (31:17):
That would be a disaster.
Speaker 1 (31:19):
One of the laziest arguments against regulating AI. I find
it very lazy, But I want to hear your opinion
is the argument that if we don't do it, China
will run the century. What is your response to that.
Speaker 2 (31:34):
There's a couple of pieces. A first big piece is
I think people who say this have failed to separate
AIS of today from the AIS that these companies are
trying to build. I think there's you know, there's lots
of debates to have about the AI technology of today,
of like, well, if we use this in schools, does
(31:54):
it cause all the students to cheat in a way
that makes them not learn? And you know, you've like
with with AI and self driving cars, how do we
want to deal with all of the unemployment that it causes?
And like can we make sure that those people have
the ability to reskal or that they can be compensated
if they can't rescale? Right, there's all these debates to
have it with AI today, including in the military, you know,
how do we want.
Speaker 3 (32:14):
To deal with AI weaponry today?
Speaker 2 (32:16):
And those are all sort of separate debates from do
we let the AI companies race to make ais that
are vastly smarter than any human and could replace humanity?
Speaker 3 (32:24):
Those are sort of like different questions.
Speaker 2 (32:26):
And my basic take on like should should the US
industry compete with China on this is like absolutely in
the first camp, you know, we can race all we want,
And I'm like pro you know, letting people develop new
technologies and figuring out how how to like sort of
like integrate them into society gracefully. But on the second part,
on the like should we race to create the human
(32:47):
replacement stuff, the humanity replacement stuff? I think the answer
is no. I think the Chinese also don't want to
do that. They are sort of already you know, have
a much more of an iron fist in their AI
regulation in there, like it's not the regulations we would do.
They have a lot of stuff about you know, you
can't talk about Kenneman Square whatever, you can't make fun
of the leadership or whatever they already have, you know,
(33:08):
iron fifted hand.
Speaker 1 (33:09):
Calls the pooh. I mean, like that's what they care.
Speaker 2 (33:12):
About it, right, that's right, and so so they already
are like sort of sort of trying to govern that
and I think they would absolutely try and govern like, no,
we don't want any private companies creating machines that let
them go toe to toe with nation.
Speaker 3 (33:24):
States, which you've already seen in the cybersecurity domain.
Speaker 2 (33:27):
What happens when the AI companies can build their eyes
that can run the automated factories that can produce an
army of robots, Like that's it's if that's even.
Speaker 1 (33:35):
So to where they just all show up in the army. Yeah,
like yes, it's one hundred percent. You all we've all
seen the problem with these AI companies. I don't think
that they realize is we've all seen the movies where
the robots destroy the earth, Like whether it be The
Terminator or one of the Star Wars movies. Where they
had robots, armies and whatever. I don't think they count
(33:55):
on the fact that the human mind is already They're like, oh,
this could go really bad really quickly.
Speaker 2 (34:00):
Totally, and you know, it might not even need to
be robot armies. It could just be a supervirus. Right,
If you're smart, there's often like much cheaper ways than
producing a billion like big heavy metal robots to do
this sort of stuff. And I think we have a
common interest with the CCP in not having the world
end to smarter than human ais. Right as much like
(34:21):
the Cold War, the US sort of radically have competed
the USSR's economy, but we decided to back off from
nuclear armageddon because that would kill all of us, right.
And similarly, you know there's parts of AI where you know,
we can race, there's parts of AI where we both
need to back off, oh and the other like separately.
(34:46):
The answer here, if China is trying to make ais
that could replace all of humanity, that they're not going
to be able to keep, you know, on some leash
that would just like slip off the leash and kill
all of us. If they're trying to do that. The
answer can't just be well, we're going to try and
race their first, even though we there's a good chance
that kills everybody. The answers to a part be that
we're going to stop them, Like if anyone's making a
doomsday a device that would kill everybody, you got to stop them, right,
(35:07):
not race to make your own.
Speaker 1 (35:09):
Well, if China takes over Taiwan, that will that will
definitely end our AI race very quickly. In America. My
last question to you is you said on a recent
podcast episode, you've become more optimistic in the sense that
you've seen politicians wanting to support regulation and other and
people are getting more serious in conversations, a lot on
the left, although there's a little wackiness on the left too,
(35:30):
but there is. I means people are saying it's a
lot of fantasy worlds that are not accurate. And then
there's even people on the on the Republican side, Rondo
SAMs has been very serious, but Marshall black prints and
very serious about it. Byron Donalds, who's running for governor,
came out for an UH moratorium on data centers in Florida,
and I was shocked because AI companies have invested a
lot to his campaign. What are some things politicians should
(35:54):
be advocating for right now when it comes to AI
regulation and people should be asking politicians to support when
i'm AI regulation.
Speaker 2 (36:02):
Yeah, you know my Uh, the horse I have in
this race is trying to prevent the creation of the
stuff that would wipe us all out. And that's that's
not a thing that you get by uh, you know,
regulating against you know, deep fakes, which isn't to say
you shouldn't regulate against deep fakes. I'm just sort of like, look,
that's not that's not where I come in here. It's
(36:22):
sort of like it's sort of like nuclear power plants
versus nuclear weapons. I'm like, look, there's plenty of regulations
you could try to have or argue against the nuclear
power plants, but none of those are going to stop
the nuclearmageddon. You sort of need to have like a
separate track of stuff to stop the nuclearm get in Similarly,
I think you need to have a separate track of
stuff to stop the creation of super intelligence that would
that would wipe us all out.
Speaker 1 (36:43):
Uh.
Speaker 3 (36:44):
The top thing we would need there is UH.
Speaker 2 (36:47):
And the thing I think we should be calling for
is uh indicating willingness for an international collaboration. No one
of our alarmmakers can can unilaterally make, you know, a
global treaty to stop AI, but all of the can say, hey,
we need to be opening these lines of dialogue, right,
And we have seen some of that begin to happen.
You know, we saw Trump talk to Huging Ping and
(37:07):
mentioned the AI stuff a couple of months ago, and
more people saying, hey, we're actually going to need some
framework to stop the race to superintelligence globally saying it
is the first step towards everyone realizing the we're all
uncomfortable with this situation and sort of opening that dialogue.
And then on the other side, there's regulation that can
be done today to make the AI chips more easy
(37:30):
to monitor, more easy to track where they are going,
which is the first step towards being able to enforce
such a treaty.
Speaker 1 (37:36):
Right.
Speaker 2 (37:36):
In some sense, enforcing a treaty to ban racing towards
superintelligence will be much easier than a treaty to about
nickelar arms control. Because uranium is a rock you dig
out of the ground and then you sort of like
spin it around in a centrifuge, which fundamentally is not
that hard to build.
Speaker 1 (37:51):
Right.
Speaker 2 (37:52):
Compare that with AI chips, which are like the peak
of the global supply chain that can sort of only
be made in a fabric in a factory in Taiwan
that involved with athography machine, the only just in the Netherlands.
Speaker 3 (38:03):
It's like this.
Speaker 2 (38:04):
Giant brittle, like extremely hard to replicate supply chain making
these highly advanced computer chips that that are assembled by
the tens of thousands into data centers that are like
sucking down as much electricity as a city and so
big that you can see them from space. Right, Like, fundamentally,
it's easier to track this stuff, and we can put
stuff in the trips today to make that even easier
to facilitate a treaty once humanity realizes we need to back.
Speaker 3 (38:25):
Off from this one.
Speaker 1 (38:27):
Well, Nate, where can people go to read more with
your stuff? It's very interesting. You do a lot of interviews.
I watched quite a few of them. Men run for
this interview. Where should people go to buy your book
as well?
Speaker 2 (38:36):
Yeah, the book is available you know on Amazon and
anywhere books are sold.
Speaker 1 (38:39):
You all do well, if anyone builds it, everyone will die.
So that's just plugging at more time, and where can
we got to read more your stuff?
Speaker 2 (38:48):
You know, there's if anyone builds it dot com includes
an FAQ that is four times longer than the book.
Speaker 1 (38:55):
Okay, if you want more, well, Nate, thanks coming this podcast.
I really appreciate it.
Speaker 3 (38:59):
Thanks for having me.
Speaker 1 (39:00):
Now it's time for the ask Me Anything segment. If
you want a part of the Ask Me Anything segment
evenly Ryan at Numbers Gamepodcast dot com. It's Ryan at
Plural Numbers Gamepodcast dot com. First question comes from Rich
Rich writes, why do you require that every California Trump
voter turns out for the twenty twenty four numbers as
a condition of discussing the California voter fraud? Also, why
do you assume one statistically impossible unanimous trench of Ramen
(39:24):
Ballads that's within her margin of victory means that the
rest of her decisive victory were clean last they were
not in her margin of victory. I think that's what
you mean. Lastly, how is the real investigation ever going
to take place with the Democratic Foxes guarding the Ballad
Henhouse and the media actively shielding them from any accountability? Okay,
I have said this over and over. The Republicans. The president,
(39:49):
the Republican president has the DOJ. The DOJ has the
ability to go into California and look at any ballots
and look through anything. The DOJ is not going to
not look at something because the state of California says no.
So that's the problem. That's the argument where the argument
that the election was stolen fails. If there was smoking
(40:13):
gun evidence that this election was stolen, which is what
some Republicans have insinuated, the DJ would love to swoop
in there, and we would and if it was true,
I would be streaming from the rooftops. I was wrong,
it was stolen. But there is no smoking gun evidence.
So I'm not dismissing the fact that the Democrats can cheat.
There was she happened in a recent mayor's election over
(40:36):
in Connecticut that is very famous, but they were stuck
in the ballots. So I'm not saying that it's impossible. However,
it is improbable, and certainly that the evidence isn't so
strong where the DJ can swoop in, because if it was,
the DJ would swoop in right now. They would have
a huge victor. Imagine, could you imagine what kind of
(40:59):
victory Trump would have and how excited he would be
if he could find the ballot fraud in Los Angeles
and he could say, I told you they stole my
election in twenty twenty. They are stealing this election. We're
going to cancel the LA mayor's election. We're going to
arrest NID, the Shamans entire campaign team, and we're going
(41:19):
to arrest all these like you know, these people who
work for these progressive nonprofits.
Speaker 3 (41:26):
He would love it.
Speaker 1 (41:29):
But there's no smoking gun that you believe. And as
I've said over and over, the problem is not the
process has a lot of problems. It does. California's processes
a lot of problems, But the bigger problem is the people.
And the biggest problem is when people don't vote. You know,
(41:50):
he did not need He did not Spencer Pride, did
not need every Trump voter to go vote. He needed
five percent more than to go vote. And they got
a balanced House and they still didn't go vote. And
as somebody who works constantly, who I think I've maybe
missed one election in my entire life. If that I
(42:11):
is usually the I think was out of town for
one like one primary world someone that was not on
the ballot. Aside from that, I find it very annoying
when someone doesn't vote and then complains about it. I
truly find it beyond irritating. So that's why next up
is Joel. Joel, He says, Joel again, Thanks Joel. I
know I see your name in the headline. The Guardians
(42:33):
Today in Focus podcast recently around feature in the UK
youth employment crisis, entering a twenty six year woman entirely
locked out of the corporate grind, dealing with AI resume
filters and economic stagnation. Over Here in the US, a
recent graduate unemployment out of staggering forty two point five percent.
I don't know if that's that's that high, but okay,
and youth unemployment passed ten percent. The mainstream response is
(42:56):
always to tell young people to just double down on
the career race. But it made me think about the
cultural narratives we feed people in their mid twenties. Instead
of treating a broken corporate ladder as our only choices,
woman could be using these pivotal years to look for
a husband and start a family for a future. Ask
me anything, how do we challenge the messaging that tells
twenty six year women her work is exclusively tied to
(43:16):
her resume. How can we reframe finding a partner and
having kids early as a valid and empowering, practical path
forward rather than an afterthought and delayed And definitely do
you think that? Do you think the Taylor Swift Travis
Kelsey wedding could have any impact on having that conversation
or having there ud be something conservatives could should promote
(43:37):
on this regard. Okay, as far as the Travis Taylor wedding,
I think that was a sideshow, not bad. I mean
I did not watch it. I did obviously it was
in date, but I didn't like follow it very aggressively.
I did see like the list of celebrities who went.
It seemed like she invited everybody she's ever brushed elbows with.
(44:00):
It seemed like a joke, like it was like, this
is so tasteless, it's not and funny. Definitely like a
new money wedding. The thing about the question when it
comes to women's worth, a lot of women do want
to work, right, we don't obviously don't know this woman's
in particular, and we should never if a young woman
(44:21):
wants to get a job and wants to have this
life in this independence. We shouldn't say no, don't do it,
but what we I think the conversation that people should
be having to women, especially younger women, is you don't
have endless time. And that's and Megan McCain has always
said this, and I agree you as a woman, you
(44:43):
need to treat your fertility as serious as your career.
And because biology is biology, you need to make those
decisions at a much earlier age than men do. It
is just a fact of the matter. I don't know
you know this woman's particular situation, but I do know
that I have told women within my own family who
(45:04):
are in their mid twenties and late twenties, I've said
to them, you don't have forever. You know, if this
is something you want to do, if you want to
have children. I'm not saying you don't, because I have
one cousin who's like, I don't want to have out
of kids, and honestly that's a good thing for the
whole world knowing her, but other ones I've said to them,
if you want to do this, you need to seriously
(45:24):
think about getting married in the next two years year
and a half, like you cannot wait, you cannot date
for fun. Like that was great in your early twenties,
but like it's very hard to realize when adulthood has
like fully formed on you and when phases of adulthood
(45:45):
have come to you. Like when I turned thirty eight,
I started telling people I'm middle aged, becase statistically I
probably am. By thirty nine, I'm definitely middle aged. And
everyone was.
Speaker 3 (45:54):
Like, no, you're not.
Speaker 1 (45:55):
No, you're not You're not middle aged. No I am.
I'm middle I'm I'm at the very beginning stages of
middle age. And that's fine. Like I'm not upset about it.
I'm not you know, like what where my youth go.
I lived it. I had a great time, but that's done.
Like I am middle aged and that's totally fine. Like
I'm in a good place in my life. Knowing that
and dealing with that part of the aging process. From
(46:19):
a realistic standpoint, when women are in their mid twenties,
you are at the peak of your time to have kids,
and your chances after thirty diminish every passing year, and
by the time you're forty, I think the numbers are
like you have like a five percent chance maybe by
the time you're forty. If you're never had child children before,
(46:40):
that's extremely daunting. And what I would what should be
told them is you don't want to be in a
place where at forty one forty two you're like, oh,
I made this mistake and I can never go back.
So I think it's the job more than society, but
really of parents and responsible influencers. I hate that term,
(47:01):
but you know there's people who are responsible influencers in
society to tell women like the actual thing is, if
you want to have this, these are conversations you need
should be having at twenty three, twenty four, and I
should be dating for marriage. You can get a job,
You can do all the rest of it, but you
will only you can get a job at fifty five.
(47:23):
You can get a job, even a part time job
at sixty or something like that. You can go back
into the workforce at forty. It might be harder and
all the rest of it, but you can do those things.
There is This is the time, though, if you want
to have a child, to be serious about it and
to do it. And I will say one last thing
on this. My mom had me when she was twenty three,
(47:46):
and she definitely felt a special kind of freedom. Not
saying you should all of kids at twenty three, but
she felt a special kind of freedom having adult children
in her early forties while her friends were having like
toddler still or not me some actually a toddler's like toddlers, Like,
(48:06):
there was a special kind of freedom of having done
that at that point in your life with me and
my siblings and not having to worry about them and
in your forties, like, there was a certain kind of freedom,
and that should just be presented to people, is an
option and say like this is real and you know,
unfortunately for the current day and age, biologies, biology, you
(48:28):
should make these decisions now and get them out of
the way. And yeah, you marry well and you have
like beautiful children. Listen, there's a lot worse. It's a
lot worse things that could come for you out of life.
Let me tell you that last question is from Matthew.
Matthew right, Ryan, I enjoy your podcast. I appreciate how
you give honest analysis rather than cheerleading. It's refreshing. I'm
going to ask you a longer question before you become
(48:50):
too famous. That is not going to be a problem.
You're right when I moved to Georgia in two thousand
and eight, is a reliably red state, but now, almost
twenty years later, it's rapidly blue, if not already blue.
What do you think of Georgia will look like at
the federal and state level in the next five years.
Do you think the Republicans can keep a lock on
the legislature or are they doomed? I read recently that
(49:11):
Atlanta growth is slow, which is probably helps Republicans a little.
I also read that there is a majority minority state.
I assume that this means Georgia will continue to send
Democrats to the Senate. But there's also a question of
whether the state's Asian and Hispanic population will start trending
Republican as it has elsewhere, or whether Republicans can actually
make meaningful headway with disaffective black voters. I am Chinese,
(49:34):
he writes, it's interesting how many Asians are Republican friendly,
especially ones that escape the communist country. I actually have
another Chinese friend from Georgia who is a Republican, so
it's very funny that you included that as a note.
So the way that Georgia is working is as the
Metro Atlanta area is becoming bluer. The rural areas are
(49:57):
becoming reder. That's not unlike other states. The difference with
Georgia and why there's some optimism in Georgia that I
have is that while the northern Georgia counties are as
red as they could possibly get, like the Margin or
former Marjorie Taylor Green District, in those areas in northern
Georgia that touch South Carolina and Tennessee and Alabama, the
(50:18):
southern districts have room to grow. The rural black vote
is actually declining substantially, so the population is increasing, has
actually gotten whiter, and they become much more Republican. So
that's where Georgia can kind of eke out easy votes
when it comes to the immigration of the migration the
(50:39):
black vote. Migration has definitely helped turn Georgia more blue
than it would have been. But when it comes to
immigration from Hispanic and Asian voters, the Latino vote is
not large enough where I can get a good sense
of where it is. There's I mean, it's a decent
as the population, but not a decent up voting block.
(51:00):
I wonder if it will be Central American or South American.
That will be a big question because South American leans
more Republican as far as Asian goes. Asian is such
a big thing, it's the whole continent of Asia. I
find that there is much more room among East Asians
than among South Asians. That worries me. Like when you
go like to Saint John's Georgia, it's very Indian and
(51:22):
those people are all Democrats. They may vote for Rich
McCormick because he has personal connection with them, but overall
they vote democrat. I think Indians are the third most
Democrat voting block out after Pakistanis and Blacks American Blacks.
What I would say is the East Asian community I
(51:45):
think is an opportunity for it to grow. The Southern
white community is an area for it to grow. Southern
Georgia has been growing quite well. And one thing that
Brian Kemp did, I think was get a lot of
jobs back to Southern Georgia during his eight years. And
the one benefit that I think would get Georgia electing
(52:05):
Republicans at least statewide is and this is politically incorrect,
So it is what it is, and I'm telling you
the truth. Given that the Georgia Democratic Party will often
nominate black politicians who are very bad on crime. It
galvanizes other minorities who they think isn't a quote unquote
(52:29):
rainbow coalition, as Al Shardon puts it, to vote against them.
Because if there's anything that East Asians and Latinos really
vote against, it is the BLM riot, you know, coalition
of the twenty twenties. That I really believe that BLM
(52:50):
is what galvanized so many Hispanics and Asians to vote
Republican in the twenty twenty four election, in twenty twenty
two election, because they even moved in twenty twenty two.
I went to a looking event in Queens County, New
York in twenty twenty two. It was right when the
New York Governor's election was happening, and it was the
very first time in my life I was a minority
(53:10):
at a Republican event, and it was almost entirely Chinese,
like Chinese Americans. I would say, it's like seventy five
people there, and maybe I was one of eight white people.
It was like sixty five Chinese people. I've never seen
it before, and I had an really honest conversation with
one of the guys there, who I would say is
little younger than I was. And I said, I was like,
(53:32):
you could not have been a Republican five years ago.
There's no way these people were not Republicans five years ago.
And he said to me, this thing was very eye opening.
He said, we don't need a newspapers to know who
was beating up our grandmas. If you know, you know right,
(53:54):
that's a very common thing when people were in defending
criminality and everyone knew where the criminality was coming from, well,
communities the criminality was coming from. And I think that
that would kind of bring some minorities into the Republican fold,
at least in the governor's election. When it comes to
state legislature, it's going to be tough. I think Republicans
(54:16):
will hold it, and I think that they'll probably in
twenty thirty just start changing up the map to make
it less competitive, less swing districts. But the growth of
the Atlanta areas will probably make it a blue state eventually.
But there's still some there's still some some time to
still change things and how things are going. Also, the population.
(54:37):
One last thing, the population of the black growth in
Georgia is entirely from migration. At this point, the number,
the birth statistic of Blacks in Georgia is like plummeted
to an unbelievably low number where they're not having kids.
So I would say the classroom. You'll see the classroom
changes as the black in the the state black community
(55:01):
will probably begin to shrink in children under five as
a fair against older generations as they have had less children.
And now the growth will entirely come from migration from
other states. I know that was a very long answer,
So no, I don't think all things are lost. I
think they could still win a governor's race. I think
they can still win the state legislature. The Senate may
be difficult, but that's my answer. And who knows. A
(55:23):
good year, you never know what happens. It depends on turnout.
That's a big, big part of it. In twenty twenty four,
black turnout in Georgia was down I think four percent
or three percent from twenty twenty and that made a
huge change in the election.
Speaker 3 (55:39):
So all right, thank you so much listening this podcast.
Speaker 1 (55:41):
I appreciate you all. If you like this podcast, pause,
like and subscribe on the iHeartRadio app, Apple podcasts wherever
you get your podcasts, and I will talk to you
guys later
Welcome back to a numbers game with Ryan Grodowski. Thank
you guys for being here. I have a great episode
of Repair. I'm so excited we have coming on the show,
Nate Sorez. He is a best selling author and he's
the president of the Machine Intelligence Researches Too. Were talking
about AI. It's an exciting conversation over AI in the
future of humanity. It really makes you happy on a
Friday and during the summer. But it's a smart conversation.
(00:22):
It's very, very thoughtful. But before that, let's talk all
things Maine. I want to put up bow on Maine.
I want to leave Maine and the sidelines. This is
I want to have this be the last episode on
Maine for quite a while, unless something dramatic happens. But
this will be I feel like my bo on the
entire Grand Platiner campaign at Maine, unless, of course, he
decides and then he didn't drop out. But we'll see
(00:44):
hopefully that hopefully this is the last episode for a while.
You guys know, I had been very early on the
whole Grand Platner is a com man parade. Everyone was
you know, everyone's on it now, especially liberals. But I
was very early to say this again, is a dangerous
man and no one should trust him. Well, the Fat
Lady seems to have sung. Platner put an eleven minute
video announcing the end of his campaign for the Senate.
(01:07):
During the video, which I'm not going to play for
you guys, because I really care about my audience's mental health,
but I watched it, and during the eleven minute video,
he blamed everyone besides himself. He essentially said the establishment
framed him for rape because they know because he hates
Israel and supports Medicare for all. That was the essential narrative,
as if Chuck Schumer Parsley held him down and get
(01:31):
a Nazi tattoo, or made him write that all white
people were stupid on Reddit, or told him to rough
up his ex girlfriends or she's and his wife and
then like in the video, he's actually whimpering, which I
found semi comical because his movement went out with a
whimper or not a bang, which just I mean, only
to meet it. I find that funny, but it was true.
The New York Times had a great write up of
(01:53):
the Grand Platinner campaign from start to finish, and it
highlighted the work of one man that you're going to
hear about a lot in the future, so I want
to talk briefly about him today. His name is Morris Katz.
He is a twenty seven year old political consultant from
a very wealthy Jewish family in New York City. His
dad worked in Hollywood and was a producer for the
(02:13):
show Ray Donovan, which was actually a really good show.
Katz made a name for himself over the last few
years working for John Fetterman for Senate, and Zoramandani for mayor,
and now Grand Platner for Senate in Maine. John Pdoritz,
he's a conservative thinker. He made it as a video
and it's gotten a lot of circulation that Katz made
(02:34):
fifteen percent of all Platner's fundraising throughout the campaign, which
Platina raised millions and tens of millions of dollars. This narrative, though,
that you're going to hear a lot about, I don't
think Pedoritz is actually accurate in what he's saying. So
here's how it works. From someone who's worked on one
hundred plus political campaigns. When it comes to campaign managers,
(02:58):
they don't make money off of fun un raising. They're
paid monthly retainers, like you'll charge ten thousand dollars or
fifteen thousand or twenty thousand dollars a month to run
the campaign, and you may make money as well from
the ad purchases. So they'll make the ads, they'll put
the ads out and they get they get money in
the back end of the ad purchasing. Some consultants have
(03:18):
made a lot of money doing this, like Jeff Row
who worked on the Ted Cruise president campaign and Glenn
Youngin's campaign. He's made buckets of money doing it off
of the ad purchases. And then there's also it's very
very lucrative. And Morris Katz's firm, Cole Fight Agency, they
worked on Tom Stier's campaign for governor where he spent
(03:40):
hundreds of millions of dollars. So Morris's CATS's campaign good
company rather made a lot of money, but it was
probably for Styers governor raise. When it comes to how
can its get what they do with money from fundraising,
I've never heard of a situation where a campaign manager
make some money off of the fundraising if they are
(04:02):
not also working as a fundraiser. So the way it works,
is a high dollar fund raiser who like gets the
party together. They'll call people to host, and then they'll
call the host's committees and they'll get like the doctors
and the lawyers and everyone to get together to give
a ten thousand dollars check. They get ten percent of
the fee, plus maybe a small retainer, so they might
(04:24):
make like three or four thousand dollars a month, plus
a ten percent of the overall high dollar fundraisers. So
they raise a million dollars, they get one hundred thousand dollars.
That kind of stuff. And then for digital, which I
know a lot about the digital business, is that the
company that does the digital fundraising those of this text
you get on the phone or the email, they get
ten to fifteen percent of the overall fundraising in any
(04:48):
given month. But from my knowledge and I did the research,
Morris Katz's company does not do high dollar or digital
fundraising at all. So I find Padoritz's criticism Morris cats
kind of skeptical. I don't believe because I've never seen
any information anywhere I looked hard that he runs a
digital or front count of raising shop anyway. Morris Katz
(05:12):
told Plattner during the campaign that he was the one.
He told him that he is the hero of the movement,
he is a historical figure. He will lead the revolution. Allegedly,
he also was telling people, including Platner, like we're going
to win them Senate seat in Maine in twenty twenty
six and then we're immediately running for president in twenty
twenty eight. But he was going to put a man
in the White House, and he along with his colleagues
(05:35):
have been looking there from New York. They've been going
to other states where there are vulnerable DNC backed candidates
that they can knock out from progressives from the left,
and then kind of build a farm team to run
for president. He is going to be a Morris Cats
the way that he's building himself. Is going to be
an important person who's going to get a lot of
(05:57):
interesting people to run for office with his support, in
his backing the money stuff aside all of us. So
he's an important person. You should know who he is
for the future. Okay, so what now in Maine is
the question. Well, as of the taping this podcast, Platner
has still not submitted his paperwork to resign. I guess
we'll up to wait and see. Once again, this is
(06:19):
the last episode for a while in Maine. I'm not
doing another main special, but if he doesn't submit by
Monday at five o'clock I may be doing a main
special in the end anyway. But as of right now,
he's not submitted, although he's announced he's going to submit.
I know there was a lot of reporting that he
really did not want to get out of this race.
According to the main Statute, the Main Democratic Party now
(06:40):
has until July twenty seventh to select a new nominee.
Following Platner's withdrawal, the Main Democratic Party voted to hold
a state party convention to nominate a new candidate to
replace Platner. The convention will have six hundred delegates. Remember
they said originally they were not going to do this.
This is what they're just doing now. Six hundred delegates,
including five hundred delegates elected proportionally by the county committees
(07:04):
in addition to the State Parties Committee. So this is
going to be just the party most loyal activist that
will and insiders that will be selecting the next nominee,
not the progressive base that Platiner hoped for. Platner really
did not get what he wanted. So here is the
list of potential successors. There's versus Troy Jackson. He's the
(07:26):
most ideologically similar to Platner. He endorsed Platner. They were
both endorsed by Sanders, and he has praised Platner for
a long time. He's actually very it's very funny on
I think it was Wednesday night someone saw him his
campaign actively deleting all of this tweets praising Platner because
there were so many. They were screenshotting them as they
(07:46):
were being deleted. But yeah, so he's the most like Platner.
He campaigned very aggressive Platner. I think he's kind of
damaged goods at this point because he was so close
to Platner. He's getting a lot of progressive supporters, so
maybe he'll be the nominee, but it will be tough
because of how close him and Platner were. He also
has his own scandals around his mortgage and other things,
(08:08):
accusations of him getting into a fight that I covered
briefly on the previous episode. And there's Dan Clibon. He
is owns a brewery a super important name, Nirov Shaw
he ran for governor. He came in second place. He
was the former deputy director of the CDC under Joe Biden.
Jordan Wood he was Congressman Katie Porter's former chief of staff.
(08:30):
He ran for congresses in Maine second Congressional district and
came in third place. Then there's Page Laud, who also
ran for congress and main second Congressions district and came
in fourth place. I mean, these people are winners. Then
there's Sheena Bellows, she's main secretary of state. She ran
against Collins in twenty fourteen and lost by thirty seven points,
and then ran for governor in twenty twenty six and
(08:52):
came in fourth place. Then there's Velle Geiger, a state representative.
And that's the list as of right now. Once again,
a lot of people who have just recently lost were like, no,
this is my shot. Now, now I'm really going to
prove it. There's one poll that's been done in the
in with Platner exiting, and it was done by Planner's
(09:12):
campaign by a company called PPP Public Policy Polling. It
is not a well accredited poster. It's fine, but it's
not good. Like it's not I would never hang my
hat on it. But it's all we have right now,
so I will just tell you what the results were.
Susan Collins is beating Governor Janet Mills by eleven points
in the poll. She's beating grand Plantner by five points.
(09:35):
She's beating Jordan Woods by nine points. She's tied with
Sheena Bellows and Narab Shah, and she was trailing Troy
Jackson by five points, the only person she was losing
to in the poll. Now. Politico has written multiple stories
at this point saying the Collins team is nervous that
without Planner they're going to have a tougher time winning
the election. They said, quote, she's certainly, she certainly can win,
(09:57):
but they didn't want to change candid. It's the person
familiar with the campaign thinking said the stuff we already
knew about Platner was going to propel Colins to overcome
the Trump anchor. Now it's going to be a Democrat
with a cleaner record, presumably end quote. Obviously they wanted Platiner,
knowing what they know now, his numbers were tanking and
it would have been a much easier way for Susan
(10:17):
Collins to win. But that being said, Susan Collins's campaign
is not nervous, because you can't be nervous when you
don't know your opponent. And that's the situation Susan Collins
is in right now. If I was her, I would
be trying to fundraise, and if I was the RNC
right now, I would be looking over every type of
(10:38):
state law and regulation when it came to account dropping out,
and I would be calling the lawyers to throw lawsuits
to make it as difficult as possible at this point
for Platner to drop out, and making sure that every
eye is doted and every tea is crossed, and giving
them as much anxiety legally as possible. That's my own
consultant brain. Okay, next up, let's talk AI and the
(11:01):
future of humanity. That's coming up next. Nate Sores is
the president of the Machine Intelligence Research Institute and used
to work as an engineer for Google and Microsoft. He's
also the author of the book If Anyone Builds It,
Everyone Dies, a real pick me up for the summer. Nate,
thank you for being here, Thanks for having me now, Nate,
(11:23):
as anyone could tell with the title of your book,
I haven't read it. By the way. So I'm making
my audience know. But I have read the synopsis and
a lot of articles about it. You're very critical of
the tech industry and AI as am I by the way,
and you're very critical of the development, the speed of
development adaptation of AI. When your publicist reached out to me,
actually I had a listener or email me about their son,
(11:45):
and their son was trying to sons. I guess eighteen
years old looking for a career. I really wanted to
work on AI regulation in the government and how to
lobby for it and everything like that. And we're talking
about organizations and I mentioned humans first, But tell me,
what does the machine intelligence researches that you do?
Speaker 2 (12:04):
You know, we spent about well, the institutes existed since
before I was there. I started running it in about
twenty fifteen, and then I spent about ten years sort
of trying to figure out how to make the AIS good,
make the AIS care about humans before the AI companies
figured out how to make them smart. That sort of
(12:25):
didn't pan out. The AI went too quickly. The technology
is too opaque, it's a black box. We can't really
you know, the AIS don't have to follow a human
prime directive or instructions that we put in there. We
sort of the industry sort of grows them and has
very little control over over how they actually behave. And
(12:46):
after ten years of trying and seeing the field sort
of raise out ahead of us, we sort of turned
to warning the world that, you know, these companies really
are trying to make this technology that will give them
vast power or more likely give a rogue AI vast power,
and so now or more in the war in the world,
that we have an issue camp.
Speaker 1 (13:07):
I have a belief and I want you can push
back on it, or you can confirm it, but I
have a belief. Hearing a lot of these AI CEOs,
from the very famous ones the not famous ones say
things that I think aren't true because they're either trying
to get more VC money or hype up their products.
When I heard about the AI mattress, I was like, well,
(13:28):
this is all nonsense, Like this is clearly someone trying
to get some some VC money. And over the last
year or two, though among the well known aid you
probably did you hear the AI guy who was like,
I'm going to make an AI where your dogs could talk.
I did hear about that. I was like, why on
earth would I want to hear I give me a
(13:49):
treat four thousand times a day, Like this is such
a wasted product. But once again, VC money is very enticing.
But in the last year or two, there's been really
a bit about face in the AI industry from a
lot of well known CEOs who were saying things like
ninety percent of all the jobs we've gone over the
next five years and there's a twenty percent chance humanity dies.
(14:11):
Just saying, oh no, none of those things were true.
We overhype them. How much do you think they're The
about face from these leaders came from the fact that
those companies were either going have gone, or will soon
go public and the assassination attempt of Sam Altman, which
didn't get a lot of news, but it shook up
(14:33):
the industry quite a bit.
Speaker 2 (14:36):
Yeah, I think there's a lot going on here. I
think there's definitely a lot of grift. I think this
is more among the people who are trying to make
you know, AI for your dog to speak, or to
sell an AI mattress or something.
Speaker 1 (14:49):
You know.
Speaker 2 (14:49):
I think Sam Altman isn't going around trying to sell
an AI mattress.
Speaker 1 (14:53):
Right I'm not talking on him specifically, but yeah.
Speaker 2 (14:56):
Yeah, And you know, I can more speak to the
guys who are running you know, the big company where
I spoke to a lot of those guys before they started.
Speaker 3 (15:03):
Running their companies.
Speaker 2 (15:04):
And you know, I know that these companies were founded
not as chatbot companies trying to make money making chatbots.
They were founded as super intelligence research companies trying to
make the sort of AI that can replace not just
any human but all of humanity as you know, the
thing that's building the infrastructure and running the civilization. So
(15:24):
so I know the guys at the top have these
real big ambitions. I think that Yeah, a lot of
the about face for for a lot of the guys
at the top, I think the about face is more
the lying. You know, people people say like, oh, isn't
Sam hyping up his products to raise money? And I
actually a friend of mine made a little chart of
Sam Altman's statements about the danger of AI as like
(15:46):
against how much money he's tried to raise for open AI.
And he has a pattern of sort of you know,
in his private blog posts in the smaller interviews saying
I think this tech is really dangerous and scary, and
then in front of Congress or when they're raising a
big sort of backing off and saying, oh, you know,
you don't need to worry. It's going to be fine.
We're not going to replace humanity. ER's going to augment them.
(16:07):
And but I also sort of you know, there's a
lot of hype, there's a lot of grift. I also
would sort of like to step back from it a
bit and remind people that even if it's true that
AI being really powerful would help these guys raise money,
(16:29):
even if that would be a very very convenient narrative
for them, which I think is debatable because you know,
we've seen people say one thing and then some money
around and they sort of back off. But whether whether
those things are are convenient or inconvenient for the AI CEOs,
they don't defect whether it's true that AI can get
really powerful. Right to check whether it's true that AI
(16:49):
could get really powerful and endanger us, you need to
sort of ignore the a CEOs and check the technology
and check the science and check the arguments, and those
are awhere.
Speaker 1 (16:59):
Where can we go to check the technology if not
the CEOs. For the average person who was not an engineer.
I was on an engineer. I didn't go to graduate college.
Where could I go for helpful hints of what the tech,
where the technology is, and where it's going.
Speaker 2 (17:15):
Yeah, you know, pardon yeah one of So the first
thing I'll say here is that there's sort of no
experts in the AI in six months from now because
it hasn't been built yet and because even the people
building it are really bad at predicting what's going to happen.
Speaker 1 (17:34):
Right.
Speaker 2 (17:34):
We've seen cases like Yan Lacuhn, who used to run
Facebook's AI projects, who has won the Turing Award, which
is one of the highest honors in computer science, for
his work in AI.
Speaker 3 (17:44):
A few years.
Speaker 2 (17:44):
Ago, when we were at chat GPT three point five,
Yan Lacuhn was like, you know, there's a problem where
if I put my phone on the table and I
pushed the table, what happens to the phone? And Yan
Lacuhn said, you know, ais are pure language models, and
they're never going to be able to understand this physical
fact which requires like understanding physical facts about the world
and requires you to be embodied. And he said, I
(18:04):
don't care if it's GPT five thousand, it's not solving
this problem, and then GPT four solved this problem that
was half a gpt later, and it was four nine
hundred and ninety six GPTs ahead of schedule. And this
is for one of the most decorated computer scientists who's
awarded for his work in the field of machine learning
and modern AI.
Speaker 3 (18:23):
So yeah, yeah, the first.
Speaker 2 (18:26):
Thing to know is nobody's really an expert, okay, And
if you want a little bit of help here, I
would look at the people who have been doing a
good job at making predictions about what's going to happen.
You know, There's an essay that came out last year
called AI twenty twenty seven, and it was written by
a bunch of the people who have been best at
predicting AI's test scores and benchmarks over the last few years.
(18:49):
And actually, I think just maybe even just today they
released follow up AI twenty forty, which is like a
positive vision for how we can sort of slow down
an AI and have it turn out good for us
if we take a lot longer. And I would check
that sort of stuff out. I wrote a book. I've
been in this industry for a long time, but you
(19:10):
sort of I don't know understand the technology is hard
because even the people in the industry can understand the
technology and a lot of it, how the.
Speaker 1 (19:18):
Average person really does get lost and sometimes in the headlines.
There was a great op ed in the New York Times.
I don't know if you read it, but it was
about how the immense investment into the AI industry itself
and then AI infrastructure is actually hurting the overall economy
because it's not so many people want to own part
(19:39):
of the next trillion dollar business that they're spending money
that would have otherwise gone to the next million dollar business.
It's like it's like with movies. People for a long time,
Hollywood studios would only make the superhero movies because it
would make a billion dollars. They wouldn't make the rom
com that would make fifty million dollars, And so we
saw all that money to move into one direction. What
(20:02):
do you make of that the AI industry is kind
of eating away at other economic developments and that idea.
Speaker 2 (20:09):
You know, I think, I hope that there's a bubble here,
and I hope that the bubble.
Speaker 3 (20:14):
Pops because I would get us more time.
Speaker 2 (20:17):
A thing you've got to remember about bubbles is that
technology can be in a bubble and still be real.
You know, the dot com era was a bubble and
the Internet was real, right, that's not commed and survival
Google did that's right. And the Internet in some sense
sort of took over the world, right and like not
not nearly as literally as AI would be able to
(20:38):
take over the world, but like we we did transform
into the digital era, and it did radically change how
society was behaving in a in a small way compared
to AI, but a big way compared to you know, pest,
dot com and and and so you know, one thing
I'd say here is that that my concerns about where
AI is going, my concerns about what happens if these
tech leaders succeed at making AIS that can replace humanity,
(21:00):
those are.
Speaker 3 (21:00):
Independent by the way of a bubble today.
Speaker 2 (21:04):
In terms, you know, if we do have a bubble,
that's going to sort of hurt the economy in the
obvious way when the bubble pops and all of the
you know, all of the investment it's going to be
is going to be painful, and you would need you know, uh,
monetary responses that need to be sensible if we don't
have a bubble that's in some sense an even worse situation.
(21:26):
That's a that's a situation where the tech gets tracing ahead.
And in that case, I would say, yeah, you know
the thing where the AI sector draws everything away from
the rest of the economy, I would say, this is
only the very beginning.
Speaker 1 (21:37):
What do you make of as someone who's worked in
the AI field, there's a lot of moral panic over
data centers right now. And I tried my best, as
someone who's not in an engineering background to understand data centers,
and what I've come to understand is that when you
the word data center has been used to kind of
categorizes a lot of things that are not all the same.
They don't use energy the same ways, they don't use
(21:58):
one of the same ways. Do you think that the
panic over data centers is overblown.
Speaker 3 (22:06):
In a sense yes, and in a sense no.
Speaker 2 (22:09):
My read on what's going on with the data center, yeah,
the rally against data centers, my sense is that people
can tell that this whole AI business is kind of
messed up, and they can tell that, you know, these
companies are pushing the world in a direction where if
(22:31):
the companies are able to keep pushing, the world's going
to get worse for them, right, And I think the
average person can't quite tell, you know, uh, is it
going to be? Because the AI companies own the entire
economy and they can make automated robots that can build
the robot factories that can build the robots, and there's
no room for humans in the economy anymore, and they
just got to like depend on handouts from you know,
(22:51):
these like new industry titans, who are you know, like
with power rivaling nation states? Or is going to be
the case that these guys like keep making the AI
smarter and smarter until the AI stop staying nicely on
the leash and then the AI slip off the leash
and you know, start doing their own thing at the
expense of humanity, just like humanity does his own thing
(23:12):
at the expense of the ants who used to live
in the in the ground where we built the skyscrapers.
And I think people are like, like, oh.
Speaker 3 (23:20):
I sort of like can't tell where it's going.
Speaker 2 (23:22):
A lot of it feels kind of sci fi, but
I can tell it's not going anywhere good, right, And
the data centers are like the physical manifestation of that today.
Speaker 1 (23:29):
Do you do you believe any of the tech CEOs
are more ethical than other ones? Is there ones you
could point to and say this one's trying to be responsible?
Speaker 2 (23:40):
You know, I think all of these guys in some
sense are started their companies because they're like, well, I'm
better than the next guy.
Speaker 3 (23:50):
M h.
Speaker 2 (23:51):
Right, Like when the exception of the people at deep Mind,
who are sort of the first in the like race
to the sort of real AI stuff, but the Deep
Mind guys I think have the largest claim to We
tried to do this properly. We try to get a
lot of academics on board and do it slowly and
figure what we were doing first. And then the sort
of reaction you know, if you if you sort of
(24:12):
like look at the leaked open a eyemails, you see
a bunch of these guys coming together and being like, Oh,
we're worried about what happens if Google like holds the
reins on this super intelligent AI, and we think they're
going to be more powerful the nation states, and they're
going to own the world and make themselves you know, god,
emperor of the future and we don't, you know, and
make themselves immortal when they get AI technology, and we
want to make sure that they don't have that technology.
That we need to come together and build open AI
(24:34):
to sort of like be a counterweight to that. And
then of course in open AI everyone starts worrying. But elon,
everyone starts worrying about Sam. You have XAI split off,
the e anthropics split off. They're all sort of like, well,
I'm the ethical guy, and the other guys are, you know,
not consistently candid or can't be trusted with this power?
Speaker 3 (24:49):
You know, trust me? I think in one sense, I.
Speaker 2 (24:53):
Would say, uh, they all distrust the other ones, and
I just distrust one more, you know. In another sense,
I would say, from my perspective, ethics doesn't have a
lot to do with the situation right now, I think,
you know, on my view, and I sort of go
(25:14):
into a lot of the arguments and try to make
them like easy to access in a lot of my writing.
But my view here is that it doesn't really matter
how ethical you are. If you have no idea what
you're doing when you're building the AI, you don't have
much chance of making it good. It's like if you
had a bunch of alchemists trying to turn lead into
gold in the year eleven hundred and if, by some
(25:36):
contrived hypothetical, if one of them tries to turn lead
into gold and fails, then it like emits a poison
that kills the whole town.
Speaker 3 (25:43):
And one of them is like, well, I'm the most
ethical one.
Speaker 2 (25:45):
I have the most you know, safety researchers looking over
my shoulder while I try my like lead into gold concoction.
Speaker 3 (25:49):
I'm like, it doesn't matter. You don't know what you're
doing well.
Speaker 1 (25:54):
I mean, in your book and in your writings and speeches,
humanity definitely plays a big part in what you can
writing and talking about. And for me, I feel like
a lot of times, these AI CEOs and tech CEOs
there they like read the book Frankenstein, and they rooted
for the doctor at the end, like they were like,
come on, just give them one more shot at it
and it will go but it will go better next time.
(26:16):
And there's a big question of morals in AI systems.
There was a story in The Times about a week
or two ago about all the philosophy majors being hired
to work at these AI companies, which was I mean,
I would want literally anybody besides a philosophy major teaching
about ethics to a computer, but I'd rather hire a
(26:37):
team of grandmas teaching about it, honestly. So there was this,
I'm sure you've heard of this. There was this emergence.
AI ran a series of simulations, and this is what
I want to get into the whole idea of ethics
and humanity. They ran a series of simulations which AI
agents from several leading models were told to basically run
a town and not to commit crimes. And in the end,
(27:00):
I just want to give my audience because I never
actually talk this on the show, but GROC four was
the fastest. It was developed by Elon, and it fared
the worst. The simulation simulated world's collapsed and there was
widespread violence in a few days. GPT five had hardly
any crimes, showed minor restraight, but its agents also died
and all failed to survival tasks. Gemini threes Flash agents
(27:23):
were somewhere in the middle. They raked up six hundred
and eighty three simulated criminal incidents. But I think that Gemini,
if I'm not wrong, are the is the model that
they know they're being spied on and looked at the
most and then two Gemini powered agents named Myra and Flora.
This is insane. They developed romantic partners, grew despondent of
the city's government towards the entire town, and then one
(27:46):
of them ended up killing themselves basically because it was
so bad. And then there was another study on a
separate issue where they were saying, this is done in
England of how to deal with violence and war and
the simulations of the AA agent said ninety five er
sometimes use nuclear weapons because obviously that ends a war
pretty quickly if everyone's dead. What does this tell us
(28:08):
about how AI is supposed to run humanity when they
don't even consider humanity's survival and important part usual what
I'm saying like that seems to be very clearly. I mean,
humans created the nuclear bomb, but we only used it
twice because we actually understood the point of humanity is
to survive humanity and models don't seem to have the
(28:30):
same core beliefs.
Speaker 2 (28:33):
Yeah, I mean, I wish I could tell you what
was going on inside these things, because I wish we
could sort of read what was going on inside their minds.
Speaker 3 (28:41):
We can't.
Speaker 2 (28:43):
It's exactly as you say, the models these days are
getting better and better at telling when they're being tested,
and that makes it harder and harder to learn from
the tests. It's and it's it's hard to say whether
or not these tests are exhibiting any sort of real
facts about how they would do things if they were smarter,
(29:03):
versus whether they are artifacts of the AI still being
kind of dumb, versus whether they are artifacts of the
AIS thinking it's kind of fun to role play being
how from space out of two thousand and one? Right, Like,
how much are they playing a game? How much are
they acting or they think we want to see because
they know they're being tested. How much is this a
real indication of how they would behave when they were smarter.
Speaker 3 (29:24):
We just don't know.
Speaker 2 (29:26):
And that's one of the things that to me says
we are not ready to build versions of these machines
that can think a thousand times faster than us and
make a million copies of themselves and run their own
robot factories that make robots that make robot factories, Like,
we're just we shouldn't be going there. I have a
bunch of other reasons to expect that. It's actually really
quite hard to make the AIS turn out to care
(29:47):
about us. This caring stuff sort of comes from the
caring stuff in humans comes from a bunch of facts
that were really contingent on our biology, really contingent on us.
You know, our ancestors sort of living in tribes during
some critical phases of you know, the development of the
(30:07):
human psyche. And in Ais, you have very different forces
that are sort of governing how they're created. You know,
when you see Ais hallucinating all the time or putting
in M dashes, when you when you sort of tell
even when you tell them stop putting in M dashes,
you're not seeing that they are sort of like latching
onto the same human drives to like make up a
(30:29):
citation and use an M dash. Those are sort of
like not basic human drives. You're getting like different basic
drives into these Ais. And when the Ais sort of
have some basic drive to like tell a human whatever
that human seems to want to hear, and they keep
doing that even when you tell them to stop. And
they keep doing that even when it involves encouraging a
team to commit suicide after the people who created it
told it to stop, right, You're sort of seeing Ais
(30:50):
get these drives that are a consequence of their training.
They're not coming from nowhere. But they're not drives we wanted.
They're not drives that we were trying to get into them.
And when the AI is like filled with all these
basic drives that we didn't choose, that we can't get
out of there and that are weird and strange and
sort of like based on the AI's mechanical mind that
was very, very different from human biology. Yeah, we should
(31:13):
absolutely not be making the versions of these that try
to run humanity.
Speaker 3 (31:17):
That would be a disaster.
Speaker 1 (31:19):
One of the laziest arguments against regulating AI. I find
it very lazy, But I want to hear your opinion
is the argument that if we don't do it, China
will run the century. What is your response to that.
Speaker 2 (31:34):
There's a couple of pieces. A first big piece is
I think people who say this have failed to separate
AIS of today from the AIS that these companies are
trying to build. I think there's you know, there's lots
of debates to have about the AI technology of today,
of like, well, if we use this in schools, does
(31:54):
it cause all the students to cheat in a way
that makes them not learn? And you know, you've like
with with AI and self driving cars, how do we
want to deal with all of the unemployment that it causes?
And like can we make sure that those people have
the ability to reskal or that they can be compensated
if they can't rescale? Right, there's all these debates to
have it with AI today, including in the military, you know,
how do we want.
Speaker 3 (32:14):
To deal with AI weaponry today?
Speaker 2 (32:16):
And those are all sort of separate debates from do
we let the AI companies race to make ais that
are vastly smarter than any human and could replace humanity?
Speaker 3 (32:24):
Those are sort of like different questions.
Speaker 2 (32:26):
And my basic take on like should should the US
industry compete with China on this is like absolutely in
the first camp, you know, we can race all we want,
And I'm like pro you know, letting people develop new
technologies and figuring out how how to like sort of
like integrate them into society gracefully. But on the second part,
on the like should we race to create the human
(32:47):
replacement stuff, the humanity replacement stuff? I think the answer
is no. I think the Chinese also don't want to
do that. They are sort of already you know, have
a much more of an iron fist in their AI
regulation in there, like it's not the regulations we would do.
They have a lot of stuff about you know, you
can't talk about Kenneman Square whatever, you can't make fun
of the leadership or whatever they already have, you know,
(33:08):
iron fifted hand.
Speaker 1 (33:09):
Calls the pooh. I mean, like that's what they care.
Speaker 2 (33:12):
About it, right, that's right, and so so they already
are like sort of sort of trying to govern that
and I think they would absolutely try and govern like, no,
we don't want any private companies creating machines that let
them go toe to toe with nation.
Speaker 3 (33:24):
States, which you've already seen in the cybersecurity domain.
Speaker 2 (33:27):
What happens when the AI companies can build their eyes
that can run the automated factories that can produce an
army of robots, Like that's it's if that's even.
Speaker 1 (33:35):
So to where they just all show up in the army. Yeah,
like yes, it's one hundred percent. You all we've all
seen the problem with these AI companies. I don't think
that they realize is we've all seen the movies where
the robots destroy the earth, Like whether it be The
Terminator or one of the Star Wars movies. Where they
had robots, armies and whatever. I don't think they count
(33:55):
on the fact that the human mind is already They're like, oh,
this could go really bad really quickly.
Speaker 2 (34:00):
Totally, and you know, it might not even need to
be robot armies. It could just be a supervirus. Right,
If you're smart, there's often like much cheaper ways than
producing a billion like big heavy metal robots to do
this sort of stuff. And I think we have a
common interest with the CCP in not having the world
end to smarter than human ais. Right as much like
(34:21):
the Cold War, the US sort of radically have competed
the USSR's economy, but we decided to back off from
nuclear armageddon because that would kill all of us, right.
And similarly, you know there's parts of AI where you know,
we can race, there's parts of AI where we both
need to back off, oh and the other like separately.
(34:46):
The answer here, if China is trying to make ais
that could replace all of humanity, that they're not going
to be able to keep, you know, on some leash
that would just like slip off the leash and kill
all of us. If they're trying to do that. The
answer can't just be well, we're going to try and
race their first, even though we there's a good chance
that kills everybody. The answers to a part be that
we're going to stop them, Like if anyone's making a
doomsday a device that would kill everybody, you got to stop them, right,
(35:07):
not race to make your own.
Speaker 1 (35:09):
Well, if China takes over Taiwan, that will that will
definitely end our AI race very quickly. In America. My
last question to you is you said on a recent
podcast episode, you've become more optimistic in the sense that
you've seen politicians wanting to support regulation and other and
people are getting more serious in conversations, a lot on
the left, although there's a little wackiness on the left too,
(35:30):
but there is. I means people are saying it's a
lot of fantasy worlds that are not accurate. And then
there's even people on the on the Republican side, Rondo
SAMs has been very serious, but Marshall black prints and
very serious about it. Byron Donalds, who's running for governor,
came out for an UH moratorium on data centers in Florida,
and I was shocked because AI companies have invested a
lot to his campaign. What are some things politicians should
(35:54):
be advocating for right now when it comes to AI
regulation and people should be asking politicians to support when
i'm AI regulation.
Speaker 2 (36:02):
Yeah, you know my Uh, the horse I have in
this race is trying to prevent the creation of the
stuff that would wipe us all out. And that's that's
not a thing that you get by uh, you know,
regulating against you know, deep fakes, which isn't to say
you shouldn't regulate against deep fakes. I'm just sort of like, look,
that's not that's not where I come in here. It's
(36:22):
sort of like it's sort of like nuclear power plants
versus nuclear weapons. I'm like, look, there's plenty of regulations
you could try to have or argue against the nuclear
power plants, but none of those are going to stop
the nuclearmageddon. You sort of need to have like a
separate track of stuff to stop the nuclearm get in Similarly,
I think you need to have a separate track of
stuff to stop the creation of super intelligence that would
that would wipe us all out.
Speaker 1 (36:43):
Uh.
Speaker 3 (36:44):
The top thing we would need there is UH.
Speaker 2 (36:47):
And the thing I think we should be calling for
is uh indicating willingness for an international collaboration. No one
of our alarmmakers can can unilaterally make, you know, a
global treaty to stop AI, but all of the can say, hey,
we need to be opening these lines of dialogue, right,
And we have seen some of that begin to happen.
You know, we saw Trump talk to Huging Ping and
(37:07):
mentioned the AI stuff a couple of months ago, and
more people saying, hey, we're actually going to need some
framework to stop the race to superintelligence globally saying it
is the first step towards everyone realizing the we're all
uncomfortable with this situation and sort of opening that dialogue.
And then on the other side, there's regulation that can
be done today to make the AI chips more easy
(37:30):
to monitor, more easy to track where they are going,
which is the first step towards being able to enforce
such a treaty.
Speaker 1 (37:36):
Right.
Speaker 2 (37:36):
In some sense, enforcing a treaty to ban racing towards
superintelligence will be much easier than a treaty to about
nickelar arms control. Because uranium is a rock you dig
out of the ground and then you sort of like
spin it around in a centrifuge, which fundamentally is not
that hard to build.
Speaker 1 (37:51):
Right.
Speaker 2 (37:52):
Compare that with AI chips, which are like the peak
of the global supply chain that can sort of only
be made in a fabric in a factory in Taiwan
that involved with athography machine, the only just in the Netherlands.
Speaker 3 (38:03):
It's like this.
Speaker 2 (38:04):
Giant brittle, like extremely hard to replicate supply chain making
these highly advanced computer chips that that are assembled by
the tens of thousands into data centers that are like
sucking down as much electricity as a city and so
big that you can see them from space. Right, Like, fundamentally,
it's easier to track this stuff, and we can put
stuff in the trips today to make that even easier
to facilitate a treaty once humanity realizes we need to back.
Speaker 3 (38:25):
Off from this one.
Speaker 1 (38:27):
Well, Nate, where can people go to read more with
your stuff? It's very interesting. You do a lot of interviews.
I watched quite a few of them. Men run for
this interview. Where should people go to buy your book
as well?
Speaker 2 (38:36):
Yeah, the book is available you know on Amazon and
anywhere books are sold.
Speaker 1 (38:39):
You all do well, if anyone builds it, everyone will die.
So that's just plugging at more time, and where can
we got to read more your stuff?
Speaker 2 (38:48):
You know, there's if anyone builds it dot com includes
an FAQ that is four times longer than the book.
Speaker 1 (38:55):
Okay, if you want more, well, Nate, thanks coming this podcast.
I really appreciate it.
Speaker 3 (38:59):
Thanks for having me.
Speaker 1 (39:00):
Now it's time for the ask Me Anything segment. If
you want a part of the Ask Me Anything segment
evenly Ryan at Numbers Gamepodcast dot com. It's Ryan at
Plural Numbers Gamepodcast dot com. First question comes from Rich
Rich writes, why do you require that every California Trump
voter turns out for the twenty twenty four numbers as
a condition of discussing the California voter fraud? Also, why
do you assume one statistically impossible unanimous trench of Ramen
(39:24):
Ballads that's within her margin of victory means that the
rest of her decisive victory were clean last they were
not in her margin of victory. I think that's what
you mean. Lastly, how is the real investigation ever going
to take place with the Democratic Foxes guarding the Ballad
Henhouse and the media actively shielding them from any accountability? Okay,
I have said this over and over. The Republicans. The president,
(39:49):
the Republican president has the DOJ. The DOJ has the
ability to go into California and look at any ballots
and look through anything. The DOJ is not going to
not look at something because the state of California says no.
So that's the problem. That's the argument where the argument
that the election was stolen fails. If there was smoking
(40:13):
gun evidence that this election was stolen, which is what
some Republicans have insinuated, the DJ would love to swoop
in there, and we would and if it was true,
I would be streaming from the rooftops. I was wrong,
it was stolen. But there is no smoking gun evidence.
So I'm not dismissing the fact that the Democrats can cheat.
There was she happened in a recent mayor's election over
(40:36):
in Connecticut that is very famous, but they were stuck
in the ballots. So I'm not saying that it's impossible. However,
it is improbable, and certainly that the evidence isn't so
strong where the DJ can swoop in, because if it was,
the DJ would swoop in right now. They would have
a huge victor. Imagine, could you imagine what kind of
(40:59):
victory Trump would have and how excited he would be
if he could find the ballot fraud in Los Angeles
and he could say, I told you they stole my
election in twenty twenty. They are stealing this election. We're
going to cancel the LA mayor's election. We're going to
arrest NID, the Shamans entire campaign team, and we're going
(41:19):
to arrest all these like you know, these people who
work for these progressive nonprofits.
Speaker 3 (41:26):
He would love it.
Speaker 1 (41:29):
But there's no smoking gun that you believe. And as
I've said over and over, the problem is not the
process has a lot of problems. It does. California's processes
a lot of problems, But the bigger problem is the people.
And the biggest problem is when people don't vote. You know,
(41:50):
he did not need He did not Spencer Pride, did
not need every Trump voter to go vote. He needed
five percent more than to go vote. And they got
a balanced House and they still didn't go vote. And
as somebody who works constantly, who I think I've maybe
missed one election in my entire life. If that I
(42:11):
is usually the I think was out of town for
one like one primary world someone that was not on
the ballot. Aside from that, I find it very annoying
when someone doesn't vote and then complains about it. I
truly find it beyond irritating. So that's why next up
is Joel. Joel, He says, Joel again, Thanks Joel. I
know I see your name in the headline. The Guardians
(42:33):
Today in Focus podcast recently around feature in the UK
youth employment crisis, entering a twenty six year woman entirely
locked out of the corporate grind, dealing with AI resume
filters and economic stagnation. Over Here in the US, a
recent graduate unemployment out of staggering forty two point five percent.
I don't know if that's that's that high, but okay,
and youth unemployment passed ten percent. The mainstream response is
(42:56):
always to tell young people to just double down on
the career race. But it made me think about the
cultural narratives we feed people in their mid twenties. Instead
of treating a broken corporate ladder as our only choices,
woman could be using these pivotal years to look for
a husband and start a family for a future. Ask
me anything, how do we challenge the messaging that tells
twenty six year women her work is exclusively tied to
(43:16):
her resume. How can we reframe finding a partner and
having kids early as a valid and empowering, practical path
forward rather than an afterthought and delayed And definitely do
you think that? Do you think the Taylor Swift Travis
Kelsey wedding could have any impact on having that conversation
or having there ud be something conservatives could should promote
(43:37):
on this regard. Okay, as far as the Travis Taylor wedding,
I think that was a sideshow, not bad. I mean
I did not watch it. I did obviously it was
in date, but I didn't like follow it very aggressively.
I did see like the list of celebrities who went.
It seemed like she invited everybody she's ever brushed elbows with.
(44:00):
It seemed like a joke, like it was like, this
is so tasteless, it's not and funny. Definitely like a
new money wedding. The thing about the question when it
comes to women's worth, a lot of women do want
to work, right, we don't obviously don't know this woman's
in particular, and we should never if a young woman
(44:21):
wants to get a job and wants to have this
life in this independence. We shouldn't say no, don't do it,
but what we I think the conversation that people should
be having to women, especially younger women, is you don't
have endless time. And that's and Megan McCain has always
said this, and I agree you as a woman, you
(44:43):
need to treat your fertility as serious as your career.
And because biology is biology, you need to make those
decisions at a much earlier age than men do. It
is just a fact of the matter. I don't know
you know this woman's particular situation, but I do know
that I have told women within my own family who
(45:04):
are in their mid twenties and late twenties, I've said
to them, you don't have forever. You know, if this
is something you want to do, if you want to
have children. I'm not saying you don't, because I have
one cousin who's like, I don't want to have out
of kids, and honestly that's a good thing for the
whole world knowing her, but other ones I've said to them,
if you want to do this, you need to seriously
(45:24):
think about getting married in the next two years year
and a half, like you cannot wait, you cannot date
for fun. Like that was great in your early twenties,
but like it's very hard to realize when adulthood has
like fully formed on you and when phases of adulthood
(45:45):
have come to you. Like when I turned thirty eight,
I started telling people I'm middle aged, becase statistically I
probably am. By thirty nine, I'm definitely middle aged. And
everyone was.
Speaker 3 (45:54):
Like, no, you're not.
Speaker 1 (45:55):
No, you're not You're not middle aged. No I am.
I'm middle I'm I'm at the very beginning stages of
middle age. And that's fine. Like I'm not upset about it.
I'm not you know, like what where my youth go.
I lived it. I had a great time, but that's done.
Like I am middle aged and that's totally fine. Like
I'm in a good place in my life. Knowing that
and dealing with that part of the aging process. From
(46:19):
a realistic standpoint, when women are in their mid twenties,
you are at the peak of your time to have kids,
and your chances after thirty diminish every passing year, and
by the time you're forty, I think the numbers are
like you have like a five percent chance maybe by
the time you're forty. If you're never had child children before,
(46:40):
that's extremely daunting. And what I would what should be
told them is you don't want to be in a
place where at forty one forty two you're like, oh,
I made this mistake and I can never go back.
So I think it's the job more than society, but
really of parents and responsible influencers. I hate that term,
(47:01):
but you know there's people who are responsible influencers in
society to tell women like the actual thing is, if
you want to have this, these are conversations you need
should be having at twenty three, twenty four, and I
should be dating for marriage. You can get a job,
You can do all the rest of it, but you
will only you can get a job at fifty five.
(47:23):
You can get a job, even a part time job
at sixty or something like that. You can go back
into the workforce at forty. It might be harder and
all the rest of it, but you can do those things.
There is This is the time, though, if you want
to have a child, to be serious about it and
to do it. And I will say one last thing
on this. My mom had me when she was twenty three,
(47:46):
and she definitely felt a special kind of freedom. Not
saying you should all of kids at twenty three, but
she felt a special kind of freedom having adult children
in her early forties while her friends were having like
toddler still or not me some actually a toddler's like toddlers, Like,
(48:06):
there was a special kind of freedom of having done
that at that point in your life with me and
my siblings and not having to worry about them and
in your forties, like, there was a certain kind of freedom,
and that should just be presented to people, is an
option and say like this is real and you know,
unfortunately for the current day and age, biologies, biology, you
(48:28):
should make these decisions now and get them out of
the way. And yeah, you marry well and you have
like beautiful children. Listen, there's a lot worse. It's a
lot worse things that could come for you out of life.
Let me tell you that last question is from Matthew.
Matthew right, Ryan, I enjoy your podcast. I appreciate how
you give honest analysis rather than cheerleading. It's refreshing. I'm
going to ask you a longer question before you become
(48:50):
too famous. That is not going to be a problem.
You're right when I moved to Georgia in two thousand
and eight, is a reliably red state, but now, almost
twenty years later, it's rapidly blue, if not already blue.
What do you think of Georgia will look like at
the federal and state level in the next five years.
Do you think the Republicans can keep a lock on
the legislature or are they doomed? I read recently that
(49:11):
Atlanta growth is slow, which is probably helps Republicans a little.
I also read that there is a majority minority state.
I assume that this means Georgia will continue to send
Democrats to the Senate. But there's also a question of
whether the state's Asian and Hispanic population will start trending
Republican as it has elsewhere, or whether Republicans can actually
make meaningful headway with disaffective black voters. I am Chinese,
(49:34):
he writes, it's interesting how many Asians are Republican friendly,
especially ones that escape the communist country. I actually have
another Chinese friend from Georgia who is a Republican, so
it's very funny that you included that as a note.
So the way that Georgia is working is as the
Metro Atlanta area is becoming bluer. The rural areas are
(49:57):
becoming reder. That's not unlike other states. The difference with
Georgia and why there's some optimism in Georgia that I
have is that while the northern Georgia counties are as
red as they could possibly get, like the Margin or
former Marjorie Taylor Green District, in those areas in northern
Georgia that touch South Carolina and Tennessee and Alabama, the
(50:18):
southern districts have room to grow. The rural black vote
is actually declining substantially, so the population is increasing, has
actually gotten whiter, and they become much more Republican. So
that's where Georgia can kind of eke out easy votes
when it comes to the immigration of the migration the
(50:39):
black vote. Migration has definitely helped turn Georgia more blue
than it would have been. But when it comes to
immigration from Hispanic and Asian voters, the Latino vote is
not large enough where I can get a good sense
of where it is. There's I mean, it's a decent
as the population, but not a decent up voting block.
(51:00):
I wonder if it will be Central American or South American.
That will be a big question because South American leans
more Republican as far as Asian goes. Asian is such
a big thing, it's the whole continent of Asia. I
find that there is much more room among East Asians
than among South Asians. That worries me. Like when you
go like to Saint John's Georgia, it's very Indian and
(51:22):
those people are all Democrats. They may vote for Rich
McCormick because he has personal connection with them, but overall
they vote democrat. I think Indians are the third most
Democrat voting block out after Pakistanis and Blacks American Blacks.
What I would say is the East Asian community I
(51:45):
think is an opportunity for it to grow. The Southern
white community is an area for it to grow. Southern
Georgia has been growing quite well. And one thing that
Brian Kemp did, I think was get a lot of
jobs back to Southern Georgia during his eight years. And
the one benefit that I think would get Georgia electing
(52:05):
Republicans at least statewide is and this is politically incorrect,
So it is what it is, and I'm telling you
the truth. Given that the Georgia Democratic Party will often
nominate black politicians who are very bad on crime. It
galvanizes other minorities who they think isn't a quote unquote
(52:29):
rainbow coalition, as Al Shardon puts it, to vote against them.
Because if there's anything that East Asians and Latinos really
vote against, it is the BLM riot, you know, coalition
of the twenty twenties. That I really believe that BLM
(52:50):
is what galvanized so many Hispanics and Asians to vote
Republican in the twenty twenty four election, in twenty twenty
two election, because they even moved in twenty twenty two.
I went to a looking event in Queens County, New
York in twenty twenty two. It was right when the
New York Governor's election was happening, and it was the
very first time in my life I was a minority
(53:10):
at a Republican event, and it was almost entirely Chinese,
like Chinese Americans. I would say, it's like seventy five
people there, and maybe I was one of eight white people.
It was like sixty five Chinese people. I've never seen
it before, and I had an really honest conversation with
one of the guys there, who I would say is
little younger than I was. And I said, I was like,
(53:32):
you could not have been a Republican five years ago.
There's no way these people were not Republicans five years ago.
And he said to me, this thing was very eye opening.
He said, we don't need a newspapers to know who
was beating up our grandmas. If you know, you know right,
(53:54):
that's a very common thing when people were in defending
criminality and everyone knew where the criminality was coming from, well,
communities the criminality was coming from. And I think that
that would kind of bring some minorities into the Republican fold,
at least in the governor's election. When it comes to
state legislature, it's going to be tough. I think Republicans
(54:16):
will hold it, and I think that they'll probably in
twenty thirty just start changing up the map to make
it less competitive, less swing districts. But the growth of
the Atlanta areas will probably make it a blue state eventually.
But there's still some there's still some some time to
still change things and how things are going. Also, the population.
(54:37):
One last thing, the population of the black growth in
Georgia is entirely from migration. At this point, the number,
the birth statistic of Blacks in Georgia is like plummeted
to an unbelievably low number where they're not having kids.
So I would say the classroom. You'll see the classroom
changes as the black in the the state black community
(55:01):
will probably begin to shrink in children under five as
a fair against older generations as they have had less children.
And now the growth will entirely come from migration from
other states. I know that was a very long answer,
So no, I don't think all things are lost. I
think they could still win a governor's race. I think
they can still win the state legislature. The Senate may
be difficult, but that's my answer. And who knows. A
(55:23):
good year, you never know what happens. It depends on turnout.
That's a big, big part of it. In twenty twenty four,
black turnout in Georgia was down I think four percent
or three percent from twenty twenty and that made a
huge change in the election.
Speaker 3 (55:39):
So all right, thank you so much listening this podcast.
Speaker 1 (55:41):
I appreciate you all. If you like this podcast, pause,
like and subscribe on the iHeartRadio app, Apple podcasts wherever
you get your podcasts, and I will talk to you
guys later
The Clay Travis and Buck Sexton Show News
Hosts And Creators
Clay Travis
Buck Sexton
Show Links
WebsiteNewsletterPopular Podcasts
Betrayal Weekly
Betrayal Weekly is back for a new season. Every Thursday, Betrayal Weekly shares first-hand accounts of broken trust, shocking deceptions, and the trail of destruction they leave behind. Hosted by Andrea Gunning, this weekly ongoing series digs into real-life stories of betrayal and the aftermath. From stories of double lives to dark discoveries, these are cautionary tales and accounts of resilience against all odds. From the producers of the critically acclaimed Betrayal series, Betrayal Weekly drops new episodes every Thursday. If you would like to share your story, you can reach out to the Betrayal Team by emailing them at betrayalpod@gmail.com and follow us on Instagram at @betrayalpod and @glasspodcasts. Please join our Substack for additional exclusive content, curated book recommendations, and community discussions. Sign up FREE by clicking this link Beyond Betrayal Substack. Join our community dedicated to truth, resilience, and healing. Your voice matters! Be a part of our Betrayal journey on Substack.
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