The Skeptics Guide #1028 - Mar 22 2025

If you're a maintenance supervisor at a manufacturing facility and your machinery isn't working right, Granger knows you need to understand what's wrong as soon as possible.

So, when a conveyor motor falters, Granger offers diagnostic tools like calibration kits and multimeters to help you identify and fix the problem.

With Granger, you can be confident you have everything you need to keep your facility running smoothly.

Call 1-800-GRANGER, clickgranger.com, or just stop by.

Granger for the ones who get it done.

You're listening to the Skeptic's Guide to the Universe.

Your escape to reality.

Hello, and welcome to The Skeptic's Guide to the Universe.

Today is Wednesday, March 19th, 2025, and this is your host, Stephen Novella.

Joining me this week are Bob Novella.

Hey, everybody.

Kara Santa Maria.

Howdy.

And Jay Novella.

Hey guys, Evan is in tax season hell right now.

He says it's a particularly bad year, I think, because there was some snafu at the governmental level.

Oh, you think?

No, I mean, there wasn't getting some forms in late and stuff like that, so it's just scrambling now.

I think the next few years are going to be tough.

Amazing.

You think cutting half of the IRS employees was a bad thing?

I don't know.

Yeah, why do it now?

Why not do it three months ago?

Whatever.

I mean, really?

You're going to try to apply logic to this situation?

Sorry.

So we had, as follow-up, we have to point out that the NASA SpaceX Crew 9 has safely returned to Earth.

So if you recall, these two astronauts, which were Nick Haig and Sunny Williams, they were supposed to be only up on the ISS for what, two weeks?

And then, what was the actual J, do you remember what specifically?

They got stranded there.

They didn't, they, they couldn't, oh, I know,

they weren't able to return on their capsule.

Yeah.

Although they were, yeah, there was some issues with the capsule, so they weren't sure it'd be safe, but it did, it was safe when it came down, But then they got stranded there.

Whatever.

The rotation was such that they ended up being there for nine months.

But now they have splashed down.

They have returned safely.

And apparently, there were some dolphins greeting them in the ocean.

Oh, yeah, I saw that.

Yeah, when they splashed in.

I wonder what the rehab is like once they get home.

Well, they were there for nine months, so they were, you know, they were showing signs without a doubt.

Yeah.

Oh, yeah.

I mean, if if I if I have a lazy weekend, I have I'm having a rough time on Monday.

Yeah, it's rough.

I mean, yeah, there's a lot of physiological change that happen when you're into microgravity for months at a time.

Not just, you know, you lose bone mass, you lose muscle mass.

Even with the exercise, that just sort of reduces the loss, but doesn't eliminate it.

But also, your fluids get all out of whack, and your eye pressure changes, and brain function changes.

Yeah, it's just not what they promised me when I was a kid about being in space.

It's so annoying.

I think people were surprised, though, you know, the fact that they were stranded and it took this long to get them back you know because we don't have ships at the ready it it's everything is so mapped out and planned and we just don't have you know we're not at that level of technology where they're like send up a three you know and yeah just not there yeah yeah but they were busy while they were there they weren't twiddling their thumbs they were actively engaged with there's a lot of experiments going on on the iss at any given time um so they were pressed into service apparently to help with a lot of the experiments that they were doing.

So we were speculating about like, were they happy to be on the ISS for nine months?

Was it inconvenient for them?

Were they?

I mean, they all said they were happy.

Of course, they're going to say that.

Yes, we're astronauts.

This is what we signed up for.

This is awesome.

But I just have to wonder:

you're not mentally prepared for a nine-month mission in low Earth orbit.

Like, you're thinking you're going to be home in two weeks, and then you miss basically almost a year of your life at that time.

Yeah, I mean, not just mental preparation.

I mean, there's just like things at home that you need to have taken care of.

Like, oh, yeah, who's going to watch my mom now that I'm in space for nine months?

You know, that kind of stuff.

That I'm sure they had to do some fast rearranging of their personal lives in order to be, you know, account for being away for so long.

It's true.

But I mean, if they had partners or they had close relatives, I think that there probably was a mentality of like, this is what we all signed up for.

We knew that this was always a remote possibility.

Yes, a not-so-remote possibility.

Right.

You know, as

commonplace and everyday as space travel has become in our lives, it's still really risky and it's still pushing the edge of technology and our ability to manage these really complex systems.

You know, this is, there's still nothing routine about space travel.

You think about how rockets work, I mean, you're dealing with massive energies, you know, pretty much at the limit of human engineering to get into orbit.

Like, you know, getting into full orbit takes, like, that's still

difficult to do, you know?

Yeah.

yeah yeah

safely reliably I mean that's just not easy no the amount of planning that goes into every one of these missions it's an enormous enormous amount of planning and then of course they're relying on the the decades of experimentation and research and everything to make it make it safe I mean I would think you know things are safer today than they probably ever were but even still I mean it takes one little mistake yeah for something horrible to happen it's not like not like jet travel not like commercial airline travel.

It's nowhere near being that routine or having that kind of safety record.

You know, the fact that we have people living in orbit around the Earth and the fact that

we sent a ship recently that orbited the moon, came back,

that's incredible.

China's missions to the far side of the moon.

It's a really great time to be alive for space travel, and we're going to see more when I talk about Artemis, though.

There are some delays.

We'll get into that.

Why don't we just go ahead and get into that now, Jay?

So tell us what's going on with the Artemis mission.

NASA announced recently that there's more delays to the program.

So, we did Artemis 1 successfully.

Artemis 1, like I said, we launched the ship, it flew out to the moon, it orbited the moon, it came back home.

Uncrewed.

Uncrewed, and it proved out a lot of the technology.

And for the exact reason why we send these test missions, we found a problem with a heat shield.

So, Artemis 2 was slated to launch September 2025, and now it's unfortunately rescheduled to April 2026.

So, you know, we're just over a year.

Hopefully, it sticks.

And the big one, which is Artemis III,

its goal was to actually put humans on the lunar surface, right?

Artemis II is just to bring people on the ride.

Artemis III is, you know, boots on the ground.

And we haven't had boots on the ground since the Apollo era.

They rescheduled that to late 2026 or mid-2027.

And you know what they say about these dates.

You know, you always take the latest, the latest interpretation.

So, like I was saying, the primary cause for the delays comes from issues that NASA found about the Orion spacecraft's heat shield during the uncrewed Artemis 1 mission back in 2022.

So, when the capsule was doing the maneuver called the skip re-entry, they found that the heat shield had unexpected erosion on it.

And what ended up being the cause was that gases became trapped within the shield and it caused internal pressure buildup.

So, you know, there was cracking, uneven shedding of the outer layers, and

that amount of damage is not acceptable.

So, the skip re-entry maneuver is where the capsule dips into the Earth's atmosphere, and it skips out of the atmosphere, and then re-enters.

And this is all to manage the angle and the speed of the descent.

And it's a really you know, really cool concept that seems to work.

But, you know, we haven't really tested it out.

So, this was it.

So that uncrewed mission proved that this was a major issue that they wanted to

rectify because, you know, when the heat shields fail, we know what happens.

So the delays have, unfortunately, much broader implications for NASA's lunar exploration agenda, which, you know, I'm following very closely because it's fascinating.

The Arnamis program was designed to establish a sustainable human presence on the moon, which is something that should have happened in the early 80s.

Right, Bob?

So

this is part of a longer plan to get ready for future Mars missions, which, you know, hopefully most of you already know about.

Any scheduled adjustments to these missions, you know, when they're going to go and, you know, when are they going to finish the build on everything?

You know, there's a

lot of moving parts here, right?

You have contractors who are working around the clock to build out the things.

that they need

with the modifications that they're asking for.

There's international partnerships where we're relying on other governments from a US-based perspective.

We're relying on other countries to give us the parts and the information that they are working on.

And it's also important to keep this momentum of these projects going because if there's a breach of that momentum, space programs are massive.

Even if you're sending a satellite up, there's an incredible on-earth amount of work and people and money and everything that's behind it.

When you're talking about sending people people to the moon again,

we've got so many people and so much money invested and so many things spun up that you need to maintain a consistent flight schedule to retain the expertise, to manage the budgets.

You got to uphold the stakeholders' confidence.

And you can't just launch a spacecraft whenever you want to.

You need approvals and all this stuff.

So the timing is very critical.

So these moves are not good.

They're not like going to end or break anything, but it does cost money and there's a lot of human hours and consideration that goes into all of this.

So hopefully, we won't take a major hit here.

Now, rescheduling of the Artemis missions comes with, you know, there's a heavy amount of international competition in space exploration, which is fantastic because it's driving governments to invest in it, which I think is important.

China's been making really good strides with its lunar projects.

The China National Space Administration, the CNSA, they outlined plans for a crewed lunar landing by 2030.

I bet you that they hit their dates.

They have a sample mission that's planned to get some regolith samples of the far side of the moon.

That's going to happen this year.

And next year, they have an ice location mission planned in 2026.

Now, you might be hearing a lot about NASA and China, like they want to find ice and they want to go to the South Pole.

Yeah, this is significant.

There are geopolitical ramifications here about that specifically.

So the South Pole is being focused on because that's the place that they think is most likely to have, you know, ice, and

hopefully, that'll be a source of water in the future, right?

So, water is unbelievably important for us to find.

One, because it's really heavy and it takes up a lot of space.

And flying water out there is not going to work.

They'll have water as much as they can carry, but nowhere near what a, you know, a constantly crude base on the moon would need.

So, what water just doesn't is not just used for people to drink and to cook with.

Water is used to make fuel, right?

Because it's oxygen and hydrogen that makes up water.

And they could, if there's enough water there, they could eventually, you know, have the infrastructure to refuel, to create fuel on the moon, which would be a really great, like a launch place for the Mars missions, which is the intention.

So if we have to fly out all of that stuff that we need, it's going to dramatically increase costs and everything.

So that's why people are scrambling to get to the south pole of the moon to find this ice and to put their flags down and

mark some territory for themselves so they can start this ice collection.

So whoever gets theirs first,

if it's China or the United States, whichever one gets there first and really puts their footprint down is going to give them a lot of power.

There's a lot of political power that comes with that.

somebody needs water on the moon you know and you're the you're the people that could dole it out you know there's just a lot of power in that so i'm hoping that, you know, we see some really, really awesome progress over the next year.

I hope that everything is going to be green lit and we stick to these dates.

I'm very, very excited to be with my family and watch people go back to the moon and land on the moon.

That's going to be one hell of a video, guys.

Imagine that.

We're going to see it again very soon.

So don't be upset about the delays because it's going to save human lives and it's part of it, it's all cooked into this whole thing.

But let's see what happens, though.

We got to keep our expectations a little loose here because, again, the amount of time and money and everything that it takes to do these things, it's so complicated that we have to be perfectly okay with the fact that we've got to wait another year.

Kara, are you okay with this?

I'm really worried about you.

I'll be okay.

Now, Bob hasn't said a freaking word because he's stewing right now.

He's so worried, yeah.

Wait, wait, what's happening?

What's going on?

Jay, will this give them more time to actually finish the new spacesuits?

Of course.

I mean, you know, I was reading about what they were going to do with that time, and I couldn't find any nice details.

Like, you know, hey, guys, we have another year.

We can blah, blah, blah.

I don't know if there is a lot of blah, blah, blah.

I think they've got everything, everything else is probably in a very good position because I can't find any info on it.

I could be wrong, of course.

I couldn't find anything about the spacesuits either, Steve.

I was thinking about that.

Like, where are they right now?

Are we manufacturing them right now, or are they still kind of kibitzing?

I think they're probably still tweaking them.

Yeah.

Is my sense.

Yeah,

I think the people in charge of the suits are like, yes, more time.

We need this time.

And everyone else is like, crap.

All right.

Thanks, Jay.

Kara, when I saw this, I knew you had to talk about this.

This is fascinating.

Tell us about punishing AI to get it to behave better.

Yeah.

So basically, OpenAI published a paper in the archive in the computer science under artificial intelligence, which means it's not yet peer-reviewed.

And then they published a blog post to go along with that paper.

So the paper that was recently published is called Monitoring Reasoning Models for Misbehavior and the Risks of Promoting Obfuscation, which sounds a little bit wonky.

And then the blog post that they published at the OpenAI website is Detecting Misbehavior in Frontier Reasoning Models.

So I think in order to really dive into this in a meaningful way, we have to define a few terms.

The first one is to specify that what we were talking about here are large language models.

We've talked about LLMs before on the show.

We don't need to do a deep dive here.

But just as a reminder, a large language model is a machine learning model.

It is coded originally to comprehend and generate human language, but it is trained on big, big, big data, right?

So it's kind of scanning large, large data sets of text, learning patterns, learning structures, and because of that, it excels it at sort of natural language.

So that's a large language model.

Specifically, OpenAI in this paper was using something called a frontier reasoning model.

So that's a type of LLM that's often defined as a large-scale LLM that it pushes the boundaries of what current AI does by doing decision-making and reasoning in a more complex way.

So the idea is, how do humans make logical analyses about things?

And how can we utilize what we know about human reasoning and implement that with our AIs?

And one of the major ways that frontier reasoning models work is that they use something called a chain of thought.

Have you guys heard about a chain of thought before?

Chain of thought prompting?

Yeah, I have.

Okay.

So, this is an approach, again, that tries to simulate human-like reasoning, according to IBM.

Like, here's a direct quote from their description, by quote, delineating complex tasks into a sequence of logical steps towards a final resolution.

So basically when you look at the examples of chain of thought prompting, you can find different examples online.

They'll show sort of a black box example that doesn't use chain of thought, and then they'll show the example where they use chain of thought next to it.

And what it does is it breaks down the steps.

So imagine you're taking a math test in school and there's a word problem and you either just give the answer to the word problem, which is a more traditional approach for these LLMs, or you go through and you break it down and say, okay, because the train is going this many miles per hour, okay, I calculate that, here's that answer, then I add this part of the word problem, here's that answer, and so you break it down into steps.

Therefore, this is the total answer.

That's sort of what chain of thought reasoning is.

So you can actually go into the code and you can see the chain of thought, and you can see each step of how that LLM got to the sort of outputs, not just based on its behavior, but based on its chain of thought.

So what OpenAI decided to do, and this is kind of an interesting take, is that they used a frontier reasoning model.

The example that they used is one called OpenAI 03 Mini, and they wanted to ask that frontier reasoning model to do something.

And then they used a weaker large language model.

In this case, they used something called GPT-40.

And that weaker model became a sort of monitor for the stronger model.

So the weaker model

monitor, or sorry, the weaker model, the monitor, was able to look at what the frontier reasoning model was doing.

It could both look at its outputs, right?

So sort of the behavior, what kind of answers it gave, but it could also look at its chain of thought reasoning and see how it got to its answers.

Does that all make sense?

So they said, hmm, we know that when we look at our LLMs, they are sometimes kind of naughty.

Like sometimes our LLMs do things that we, you know, don't want them to do.

Like they take loopholes or they will pretend that they did a process, but really they'll give you like an empty data set or something.

Like, you know, they'll cheat.

And so they were like, okay, this is something that the LLMs have started to do on their own.

We want to understand why they're doing it.

And one of the ways that they know that they can do it is through something that they're calling like a reward hack system.

So basically the LLM itself will be given a coding task or it'll be given something to do.

And then it'll go like, okay, how can I hack this so that I can get the reward, the incentive for doing this faster or with fewer steps?

And they find that large language models exploit loopholes just like people do.

So they're like, okay, this this is something that's probably not a good thing because when we're trying to ask an LLM to, you know, do a task, we want the outcome to be what we asked for, right?

We don't want it to be some sort of like hacky, you know, Mickey Mouse version of what we asked for.

We want the actual outcome that we're requesting.

And so they were like, okay, how can we prevent these LLMs from doing that?

And they thought, well, one way that we can do it is we ask it to do something.

We see how it does its reward hacking, right?

This misbehaving and sort of trying to get through the

prompt as quickly and sometimes brokenly as possible.

And what happens if we have this monitor, this weaker large language model that's actually watching what the frontier reasoning model is doing?

And we have the model detect these different exploits and figure out, okay,

where are they hacking?

What are they doing?

How can we make sure that we sort of let it know that we're aware of what it's doing?

And so they looked at two different things.

They could look at both the outcome of the model itself, like did it do the task?

But they could also look at what I mentioned before was that chain of thought.

So they can actually read the sort of chain of thought.

I guess, what would you call it?

Is it programming?

The chain of thought commands?

I don't know what the right word is to use there.

Flow?

There you go.

Okay, good.

They could look at the chain of thought itself.

And they found that very often the monitor was much more effective if it had access to the chain of thought in addition to the actions of the LLM.

If they only looked at the actions, sometimes the monitor couldn't tell if the LLM cheated to get there.

They could only see the output, right?

But once they could look at the chain of thought, they could see sort of where the LLM was misbehaving, where it was cheating.

It would say something like, let's hack this,

and then it would hack it.

And so it's like, oh, wait, now it's like, I can see the step where it's hacking.

Now I can detect it easier.

Okay, that's right.

It's not a black box anymore.

It's not a black box anymore.

Exactly.

So this is a great thing, right?

So you would think, okay,

if I can see the chain of thought and I can see that it's actively hacking, then I can suppress these, quote, bad thoughts or these, you you know misbehaviors and so what they did is they penalized the LLM for having what they called quote bad thoughts during their training meaning that they gave them a lower reward whenever they triggered that that chain of thought monitor so every time you trigger the chain of thought monitor and it gets flagged that you're that you're being hacky you know that you're doing something we don't want you to do you get a lower reward well what do you think happened do you think it ended up hacking less or

what do you guys think?

Hacking more surreptitiously.

Hacking more surreptitiously.

Yeah, it just started lying.

Oh my gosh, just flat out lying.

It just flat out was like, I'm not doing that.

And it stopped, you know, giving it indications that it was doing the hacking.

So the outcome here is

this effective method for monitoring reasoning models may not be worth the incremental improvement to the outcome capabilities because it's learning to hide its chain of thought from the monitor.

And that's not a good thing.

So they're saying outright, until we better understand these what they're calling chain of thought optimization pressures, we recommend not using them.

Because if you're starting to train your models by pressing on the chain of thought and actually training them out of basically monitoring that chain of thought, what's going to end up happening is they're just going to get even better at cheating and lying.

And then then they're going to go back into black box and we will never be able to see it.

But maybe it can help teach us how to be better at lying and cheating.

Well, what they're saying, they're funny because they keep, of course, comparing it to human thought because that's how they think about this, right?

It's a large language model trying to mimic human reasoning, all of that.

And so they're basically like, we already do all this.

Machines are learning how to do it the way we do it.

It's a child.

This is what children do.

Yep.

As soon as you punish them for a behavior, they immediately learn how to lie to cover up the bad behavior.

Yeah, they've got the cookie, they've got the chocolate on their face, and it's like, you stole a cookie?

I didn't steal a cookie.

It's like, but I can see the chocolate on your face.

So

I have a solution.

Yeah.

Okay.

The reaction that we want should be something akin to, I have a terrible pain in all the diodes down my left side.

That's what...

Kara, that's what Marvin, the paranoid android, said in Hitchhiker, I did the galaxy.

Thank you.

So what I'm saying is we have to cause pain, like real-ass pain, so that they say something equivalent to that.

I don't know, though, Bob.

I'm afraid that would be worse.

I mean, that's kind of what they're saying, right?

Like, here's a direct quote from

the paper on the archive.

While we show that integrating chain of thought monitors into the reinforcement learning reward can indeed produce more capable and more aligned agents in the low optimization regime, we find that with too much optimization, agents learn obfuscated reward hacking, hiding their intent within the chain of thought while still exhibiting a significant rate of reward hacking.

Because it's difficult to tell when chain of thoughts have become obfuscated, it may be necessary to pay a

monitorability tax by not applying strong optimization pressures directly to the chain of thought, ensuring that chain of thoughts remain monitorable and useful for detecting misaligned behavior.

Don't give them the opportunity to hide the fact that they're cheating.

And the way you're going to induce them to hide the fact that they're cheating is to tell them, if you cheat, I'm going to punish you.

That's so crazy.

How can that possibly be what's going on?

I mean, these aren't, this isn't sentience.

No, but it's interesting because

it's following the pathway of least resistance, I guess.

Yep.

You know, it's just looking for every type of solution space, and lying turns out to be very efficient, I guess, as a way of...

of getting to the ultimate goal, which is to get the reward and to avoid punishment.

Where did it learn about lying?

How did it work?

It figured it out because that just works, because it works.

But it's fascinating because this is what people do.

It's so amazing that they're just automatically following human behavior and that it learned on its own.

Obviously, no one told it specifically to lie as a strategy.

It just hit upon that automatically.

Yeah, but wait, we're talking about an LLM trained on like the entirety of the internet, right?

I mean, that's what we're talking about.

How much lying is on the internet?

I mean,

thinking is right.

I don't think it needed.

And it didn't need, it just wasn't big insight.

This is just like, oh, no, it doesn't matter.

It's more than just lying.

That's the thing that we have to remember, right?

What we're talking about here is what they're calling reward hacking, right?

So, like, finding a loophole to give you the output in the

path of least resistance, like you mentioned, Steve.

And if it means cutting corners or not doing it the way that you're supposed to do it, but doing it a cheaper, faster, and a little bit kind of jankier way, it's gonna do that.

And then when we say, Hey, don't do that, you need to do it the right way, it goes, Yeah, it goes, I did do it.

I really love that, Kara.

Cheaper, faster, jankier.

Yeah, that's what it is, right?

Right, but I, but what I also love is that we're using the word lying because I used the word lying when I was talking about this.

Nowhere in the paper do they use the word lying, they're very careful to say, Oh, they're obfuscating.

Yeah,

okay, and then we look that word up.

Nice.

What's also fascinating is that we may realize that the pathway to getting it to stop lying is to teach it to care about the truth, I guess.

We have to give it

the AI version of a superego where it polices itself.

But if it's amoral, basically we're creating an amoral thinking

kind of algorithm.

And the amoral thing to do is to follow the shortest path to your goal.

That's because we're trying to optimize it, but not

without any regard for any values or morals.

And so what they optimize,

we have to treat it like nomad.

We know it's sterilized.

I mean, make it so that it self-destructs if it tells a lie.

Well, and here's the thing.

I don't even know if it's so much, quote, morality as it is regulation, which I think can be somewhat amoral.

But you're right, Steve, there is that interesting question of does there have to be a moral valence?

I can't help but think about a large scale system like laissez-faire capitalism and see a massive parallel here.

That if we're talking about capitalism without guardrails, it's not a bug, it's a feature that greed is going to be rewarded, right?

Greed is good.

By definition, in a capitalist system that does not have regulation, the

cheaper, faster, jankier approaches, because it is by definition an amoral.

I don't mean immoral, I mean amoral system.

And so people will be exploited, people will be harmed.

So we're going to be able to do that.

So this is exactly.

Bob's example of Nomad is, I unfortunately spot the F on because this is the worst case scenario that you don't need a truly sentient AI that's malevolent or megalomaniacal or whatever.

It's amazing how close to these science fiction worst case scenario this is.

It's just figuring out that, oh, well, you wanted me to do X, I'm doing X.

That happens to involve lying.

And

you don't want me to do that.

Okay.

And then they

hide the lying.

But similarly, it's like, you know, AI may also decide, oh, you want me to

end conflict?

Okay, I could end conflict by exterminating all humanity.

That's the

pathway.

I mean, that's kind of a sci-fi joke, but that's an exact analogy to what's happening here.

Yeah, it reminds me of D D, Steve, too.

I remember when you get like a,

you hit the jackpot, right?

You're playing D D, like you get a wish.

You got to be careful how you ask for stuff.

Because didn't somebody in your, in your team, you were GMing and they and they and they said, um, you know, they asked for something that was kind of ridiculous.

So they, so the wish put them all into sleep-suffering hallucinations of what they wanted to be real or something like that?

It's like a low energy, you know, low energy.

You don't have to get into the details about that, but yes,

that's too much of a segue.

But the bottom line is, yes, any DM worth of salt knows how to completely subvert wishes.

You have to be careful about where they're coming from.

These were morally compromised wishes.

And so it's the same thing.

You have to be careful how you tell the AI what to do.

But again, it's just not the same as

interpreting our literal words.

It's trying to predict how it's going to solve the problems we give it.

And

we make assumptions about like what it won't do to solve the problem but we can't do that

the whole thing is fascinating it's yeah and scary

some too yeah

and when they get smarter they're just gonna find these workarounds that we will just never figure out i know the agenda we gotta not push them to hide their workarounds yeah we gotta they have to be deadly exactly so they have to be out in the open which is why they're saying don't do this monitoring and and give this negative feedback because they're gonna start hiding it

Let them show their work out in the open, even if we don't like their work.

Right, and then take it for what it's worth.

Yeah.

Like, okay.

And then let's figure out another way to suppress them from making those decisions.

Like, maybe it's not about doing the most optimized thing.

Maybe it's about doing the thing that is most optimized given X, Y, and Z parameters that are not necessarily going to be optimized.

But this isn't the first time that they've caught AI doing this, Carrie, and just giving them parameters isn't enough because, again, they sort of hack the parameters.

What they're doing is the AI equivalent of rationalization or motivated reasoning.

It's figuring out how to get what it wants with the least amount of whatever and rationalizing away that the things you're telling it not to do.

In this case, right?

It's like it's lying and then it's lying about lying or hiding the lying because it's still trying to get to the shortest pathway to its goal.

And I would say, you know, oh, we just, we have to think at this stage about how to build the, you know, the laws of robotics or whatever into the AI.

But again, that may work to some extent, but it doesn't really solve the rationalization or motivated reasoning thing.

They'll just figure out a way to justify why they have to do the thing that will hack their way to the outcome they're trying to get to.

I don't know.

It seems like really,

this is a difficult problem.

We're going to have to really think very carefully about.

I mean, it's a difficult problem for humanity.

Yeah.

So obviously it's also a difficult problem for the machines that humanity is making.

Why does it have to be so efficient?

That's what I'm saying.

What if we say don't do it so efficiently?

Do it with all these other parameters instead?

I don't know.

I guess we'll see.

But I mean, that does seem like...

We have to crack that nut, though, because this is really...

Yeah, because I think we see that in evolution, too.

Like, for the most part, we try to do things as efficiently as possible, but we're not willing to give up certain things just for efficiency, which is how we end up with things that don't seem to make a lot of sense.

Right.

Well, evolution, I think, is a good analogy because, yeah, evolution is the survival of,

again, it's not really the fittest.

It's

whatever has the traits that will lead them to have the greatest reproductive success.

But that that doesn't necessarily that the fitness can get defined in all kinds of weird ways.

Exactly.

And there are hacks too.

Like if females have, you know, sexually select males to have ridiculous big feathers, even though those feathers are inefficient and make them easier to hunt and less physically able to evade predators, whatever, it doesn't matter because they need them to get laid.

And that's it.

That sort of subverts

everything else.

And that's kind of what's happening here.

And evolution is also notoriously for like evolving into a corner.

You know, it's like, yeah, you accomplished your short-term goal of maximizing your reproductive success.

Now you're screwed.

But now you're screwed, right?

Because you can't see the big picture.

So, again, how do we make AI see the big picture and want to do the right thing?

Want to find it.

It's hard for AI to evolve itself into corners over and over again.

So not without a lot of damage.

But we got to sort this out before we start putting it in charge of our nuclear profile.

Seriously.

All right.

Let's go because we saw that movie, right?

All right.

I'm going to give you guys a little bit of an update on the bionic hand, right?

This is a technology I've been following.

I've been sort of monitoring it and reporting on it from time to time on the SGU and on my business.

You can rebuild him.

Faster, stronger,

jankier.

All right.

So there was a

recent study published.

This was by the Johns Hawkins University.

They're working on developing a prosthetic robotic hand, right?

So this is something,

not for a robot, although it could work for a robot, but this is more for to replace a lost body part.

And of course, the human hand, right?

What's what's more difficult than that, right?

Imagine, you know, think about the exquisite sensitivity, finesse, fine motor control, speed, you know, sensitivity that you have with your hand.

Oh, it's an amazing tool.

My God.

Just this simple thing of, like, forget about typing at 100 words a minute, right?

But, or playing the piano or playing an instrument, a stringed stringed instrument.

Imagine just holding an object in your hand, right?

Just that, just holding an object in your hand.

You have to apply enough pressure to keep it from falling, but not so much pressure that you crush it.

Yeah.

So your hand has to have some sense of those things, right?

How do you know if you're holding an object hard enough to keep it from slipping?

Because you feel it slip, right?

There's some sensory feedback that keeps you maintaining the right pressure.

In fact, one of the symptoms of carpal tunnel syndrome, even before people get weakness, is they notice that they're dropping held objects.

They're dropping objects from their hand because they don't have the sensory feedback.

They're getting weak.

Yeah, I see that a lot with my patients who have neuropathy from chemo.

Yeah, exactly.

That they drop things and they can't hold things slippery.

But their grip strength is not that much.

That's the feedback.

It's not weakness.

It's just that without that grip.

They just can't feel that they're holding it.

exquisite sensation, you know, yeah, they can't feel how much pressure they need to keep it from slipping without crushing it, too, right?

And you know, people need to be able to hold delicate things, soft things, heavy things, hard things, slippery things.

And we do it, of course, all seamlessly, subconsciously, but that's the engineering challenge now of trying to make a robotic prosthetic hand.

So, the Johns Hopkins team trying to do this, right, and trying to specifically mimic the human hand has came up with a couple of innovations, right?

So they're calling it a hybrid to hybrids.

To hybrids, bionic hand.

You gave them space for that.

Because, so essentially, prior incarnations of the bionic hand have had to make a choice.

Are we going to have

the hand be rigid or soft?

And there's advantages and disadvantages to both.

The soft hand is gentle, the rigid hand is strong.

So this is a hybrid hand because it has a rigid, hard endoskeleton and a soft outer tissue, right?

So again, mimicking a hand.

So you have a hard skeleton, and on top of that is layered this soft gel, right?

So that's number one.

The second thing they did was they

added three layers of sensors sensors at different depths in the fingers?

Can you imagine that?

So, with a little bit of pressure, the first

most superficial sensors will be activated.

With more pressure, the deeper sensors will be activated.

That's smart.

And with more pressure, the deeper sensors still will be activated.

Again, it's not unlike how our hands work.

And then they tested it out in a few people.

They trained them how to use it.

This particular model was connected to muscles, right?

So you know how the different ways you could interface a robotically controlled prosthetic limb to the brain.

You can have a direct connection to the brain.

You can have an indirect connection to the brain through like scalp surface electrodes.

I haven't seen this yet, but I'm looking, I'm wondering the stentrodes, you know, the ones that go inside the skull in the veins, are another option there.

Or you can connect directly to the nerves, the nerve endings, if they are there and they're functioning, or you can connect directly to the surviving muscles on the stump, basically, right?

On, like, say,

a

retained forearm muscles.

And then the user learns how to control the robotic limb by contracting their surviving muscles.

And that learning goes two ways, because you could also now use AI so that the limb learns the user and the user learns the limb.

Does that make sense?

They're figuring out, the user is figuring out how to contract the muscles in the right combination and sequence to get the hand to do what it wants, but the hand also learns what the user wants it to do using AI.

And so

that rapidly increases the learning curve.

Now providing sensory feedback to the user, and we've discussed this before, means that they can control the limb better, more naturally.

It feels more like a part of them.

They can control it even without looking at it.

Whereas without the sensory feedback, they need visual feedback to

be able to control the limb.

But now, with the multiple layers of sensations, it gives them some sense of how strongly, for the first time, really, how strongly they are gripping something they are trying to grip.

And the early experience with this has been very positive, giving increased control, being able, for example, to grip and raise a cup of coffee.

Think about that.

Imagine having that.

That's hard because you also have to do the thing where it doesn't slosh out of the cup.

Right, but also, like, the cup is a little delicate, but the liquid makes it heavy.

So that's a challenging item when you think about how do you hold it firmly enough

to carry the weight without crushing the cardboard or whatever cup that you're using.

Right.

And so they were able to do that.

Users were able to successfully do that with this hybrid multi-layered sensored bionic hand.

So it's incremental, but I think it was significant enough to report on it.

And they're planning on, what do you think?

What's the next step?

They're going to give more layers, more sensors, more, you know, so increase the discrimination of the feedback.

Yeah, closer, yeah, closer sensors, right?

More.

Yep.

So they also measured the user's discrimination sensitivity, and they said that, so let me just read the quote here: Our innovative design capitalizes on the strengths of both soft and rigid robots, enabling the hybrid robotic hand to compliantly grasp numerous everyday objects of varying surface textures, weight, and compliance while differentiating them with 99.69% average classification accuracy.

That's pretty good.

The hybrid robotic hand with multi-layered tactile sensing achieved 98.38% average classification accuracy in a texture

discrimination task, surpassing soft robotic and rigid prosthetic fingers.

So

that's a pretty solid advance.

So I love the fact that this technology is just rapidly, incrementally, but I would say rapidly progressing.

And I also love the fact that it's following, and this gets back to the AI discussion a little bit too, it's following a mimic human

evolution kind of approach.

Like

evolution already solved this problem, right?

And so why not just engineer it to function like a human hand?

Because that's what we're trying to replace.

And that approach seems to be working well.

So that's why I think it's similar to the AI in that we have to think, well,

you know, basically an AI is like a child.

We have to raise it to do the right thing.

And how do we do that?

Like, how do you do that with your child?

You know, anyway.

Interesting to think about it that way.

We'll keep an eye on this technology.

It is continuing to progress.

There's many, many different centers around the country, around the world working on it.

The progress seems to be pretty steady.

Steve, they have a picture of it?

Yeah, yeah.

It's on my blog post.

It'll be on the

it's the picture I'm going to use for this week's episode.

Okay.

Yeah, I mean it doesn't look that impressive, but it's a five-fingered hand and it's kind of like gel-like fingers.

You can kind of see their little see-through, so you could see the endoskeleton as well.

It doesn't look human, you know, humaniform.

You know, it looks like a robot hand, doesn't look like a human hand.

So I wonder how far away, though, we are from

something similar to Luke Skywalker's hand, you know, when it gets a cut off by the lightsaber.

And that it looks like, or even just like the bionic man, like a

you really can't distinguish it from a natural biological hand.

Probably, I would guess we're still something like a hundred years away from that.

But

having something that functions, that's like pretty close to

human functioning, at least for like everyday basic activity, I think we're getting we're partly there already and we're getting pretty close to that.

And then it'll just be incremental from there.

Like you think of like the really demanding tasks, like some of the ones I mentioned, playing a piano or playing an instrument, that's probably still a ways off as well.

But pretty exciting.

So exciting.

All right.

Bob, tell us about this petawatt electron beam.

I won't say my typical peta joke that I do every time.

Yeah, that's lame.

Okay, yeah, this was super cool.

Researchers at the SLAC National Accelerator Lab have created a super intense, super brief beam of electrons with 10 times higher peak current compared to previous experiments.

Now, how do they do that?

Why should you care?

And what big numbers am I going to throw at you today?

This study was recently published on the Preprint Archive server.

The title is Experimental Generation of Extreme Electron Beams for Advanced Accelerator Applications.

All right, I'll give you my takeaway.

My takeaway from all this research around this topic was that the future needs bigger, better, briefer beams.

A lot of alliteration there.

Beams in this broader context outside, you know, I'm going stepping a little bit outside the confines of this paper, but beams in this broader context refer not just to electron beams, like in the paper, but all types of beams.

Particle beams, essentially, beams of photons, protons, electrons, positrons, ions,

lots of particle beam types.

These particle beams have myriads of uses.

I can't even list them all in scientific research, medical treatments, industrial applications, electronics, and on and on and on.

These beams are hugely, hugely

important, and they're just ubiquitous.

I think I was kind of surprising to see all the way that all these types of beams are actually being used.

Now, I say that the future needs better beams because that's the conclusion of a separate paper that I recently read.

It's called Accelerator and Beam Physics Roadmap created by

government scientists in 2022.

I found this while researching

this topic, and it's on the Department of Energy Government website.

At least the article was there a couple of hours ago.

Who knows if it's even still there now?

Ha.

In that paper, they conclude that in order for us to take our research and tools to the next level, they described four grand challenges in accelerator and beam physics over the next decade.

So they talk about we need to greatly improve the beam intensity, the beam quality, the beam control, and beam prediction, being able to predict it using, you know, with like virtual particle accelerators, essentially.

All those things

need to be greatly improved if we are really going to...

have the tools that we need to have in the next, even in the next 10 years.

So in one fell swoop, this research that I'm going to talk about has made significant advances in all four of these grand challenges.

And they did it by demonstrating what they could now do to a beam of electrons.

The main goal

or

one of the biggest goals was to take the beam of electrons and squeeze it and shape it in such a way that it has the properties, the precise properties that they wanted.

So they're manipulating the beam.

In this case, they wanted this beam to be as brief as possible and as powerful as possible.

Those are two

really important qualities that they wanted to showcase.

So now we can already use machines to emit a brief and very powerful beam of electrons, say, but that's just the starting point.

They start with this beam that's emitted and then they send it down the tube and then they need to massage the beam and tweak it to really take it to that next level.

They can't just create the beam that they want from the machine.

They need to manipulate it in order to take it to the highest echelons of what's physically possible now.

So, once they created this initial beam with the electrons moving at near the speed of light already, they then hit it with a specially modulated infrared laser beam.

It only lasts for a few picoseconds, trillionths of a second.

That in and of itself is amazing advancement, but that's been made years ago.

This laser beam, though, was carefully crafted to make some electrons gain energy and some electrons to lose energy.

So, what happens to the electron beam then is that many of the electrons just start converging to create a very brief but very wickedly intense spike of current.

The end result of this beam shaping technique was something that was never created before.

These electrons went from being relatively diffuse, right, and lower intensity to an intense spike of electrical

energy bunched together in a spike lasting only a few femtoseconds, a quadrillionths of a second.

Its peak power surged to one petawatt.

That's a quadrillion watts.

That is 100 times greater than the global instantaneous electrical consumption.

So at any moment, we're consuming a few terawatts of electricity.

They brought this to a quadrillion watts, so hundreds of times greater.

I mean, really incredible.

That's extremely powerful, but remember, the total energy is tiny, right?

Because this is only lasting for a few femtoseconds.

So it's not like you could run your house on this for more than a fraction of a second, but it's compressed in time so much that that's why it's so amazingly powerful for that very brief, brief instant.

Claudio Emma is a lead author of the new study and staff scientist at the Department of Energy's National Accelerated Laboratory.

Claudio said, Not only can we create such a powerful electron beam, but we're also able to control the beam in ways that are customizable and on demand, which means we can probe

a much wider range of physical and chemical phenomena than ever before.

Now, this new shaping technique does not only apply to electron beams.

That's right, that's kind of my point.

This can be applied to many other beam types, and it looks like this technique should help with each of those four beam grand challenges that I described.

One specific quote, though, really caught my attention.

Richard Darcy, who is a plasma accelerator physicist at the University of Oxford, said, it's not just an experimental demonstration of something interesting, it's a stepping stone on the way to mega-amp beams.

I'm not done with the quote, but let me interrupt the quote right here.

He mentions mega-amp beams.

That's a million, a mega is a million.

So that's a million amps.

The beam that they made was 0.1 mega amps, so that's a tenth.

So what he's saying is that we are a tenth of the way there to creating these mega amp beams.

And so Darcy continues, if achievable, those even more powerful beams might begin to perform extraordinary feats such as ripping particles out of empty space.

So what he's saying is that assuming that these electron beams can get into this mega amp amp regime, which is no guarantee, we're only a tenth of the way there.

Who knows what it might take to get an order of magnitude higher?

It might be, you know, it could take, who knows, could take a ridiculously long time, or it could be simple scaling, and they could get there.

Don't know, don't know.

But if we ever get to that, this million amps, they might allow for something extra cool, which is ripping particles out of empty space.

I've heard of this before.

It's a fascinating idea.

And it is actually a thing.

It's actually, you know, it's theoretical and it's described.

It's called vacuum breakdown.

It's also known as the Schwinger effect.

Schwinger?

Is that the sound it makes?

Yes, that's exactly.

The idea predicted by quantum electrodynamics is that even empty space is seething with virtual particles, electrons and positrons that quickly annihilate each other.

Now, the intense electromagnetic fields created by a mega-amp level beam might be able to separate these virtual particles particles from the quantum vacuum, seemingly creating particles from nothing, but actually turning the energy of the field itself into matter, energy into matter.

Hello, E equals M C squared.

If that pans out, who knows what technologies could come from that?

I tried to do some research, like what would happen if we actually had these mega amp beams?

What kind of things could they actually do?

If you can rip these particles and antiparticles from the vacuum, what could we do?

Some think that it could lead to ultra-powerful particle sources, seems kind of obvious, I guess, or even exotic propulsion systems.

But my favorite remote possibility is the creation of cheap antimatter.

That would be truly awesome because,

as you may know, you know, we've been making antimatter for decades.

It's extremely, ridiculously expensive, like billions of dollars per nanogram.

It's just like, we're never going to be able to create it, it seems, in quantities, unless we have a major breakthrough like this.

So that could be quite fascinating.

So I got to say kudos to these scientists for shaping their beams.

Keep up the good work and get to mega-amp beams as soon as possible.

Please, that'd be very nice.

Thank you.

All right, thanks, Bob.

Well, everyone, we're going to take a quick break from our show to talk about our sponsor this week, Acorns Early.

Do you know what I wish I had when I was a kid?

Acorns Early.

It's a smart money app and debit card for kids that helps them learn the value of money.

Acorns Early's chores tracker teaches kids that hard work pays off, and Acorns Early lets parents pay allowances automatically.

So, kids can spend what they earn with their very own debit card.

Plus, parents can keep track of where and when their kids are spending.

Yeah, I'm definitely going to use this to help my kids learn about savings and investing.

You know, this was something that my parents never even talked about, let alone like, you know, really taught me how to do it.

And I think it's an incredible lesson for kids, and I'm super excited that they're going to get a head start.

You ready to help your kids learn the value of money?

Just head to acornsearly.com/slash SGU or download the Acorns Early app to get started.

Sign up now, and your first month is on us.

TNCs apply, monthly subscription fee starting from $5 per month, unless canceled.

All right, guys, let's get back to the show.

All right, Jay, it's who's that noisy time?

All right, guys, last week I played This Noisy.

Yeah,

sounds big, doesn't it?

Do not like.

Don't like it?

Nope.

All right.

Well, I had lots of people email me this week.

So Shane Hillier wrote in and said, Hi, Jay, this week's noisy sounds like being on the inside of a large ship during rough seas.

God, that's scary.

Yeah,

I agree.

It does sound like that.

Although that is not correct.

But hell yeah, it sounds like that.

Will Beldman wrote in, hi, Jay, sounds to me like an icebreaker.

I'm thinking

he's talking about the boat and probably not like the sentence, right?

Yeah.

Yeah, yeah.

Like a nuclear-powered icebreaker.

Yeah, so the icebreakers have things that break the ice, or is it just the hull of the ship breaking the ice?

I don't know.

It's the hull of the ship.

Yeah, I think it's mainly the hull bearing down on the ice.

But yeah, the hull is designed to cut through the ice, and if it's powerful enough, the ship engines are strong enough to do it.

They need one of those lasers Bob was talking about.

A listener named Andrew Nicholas wrote in and said, hey, Jay, I think this sounds like a large whale, maybe blue or humpback, breathing through its blowhole.

It sounds like this was recorded on a boat or ship.

See, we have a nautical theme.

this week, guys.

That is not correct.

Yeah, I don't know.

I haven't heard whales make that noise unless you heard something specifically.

I don't know where you got the idea, but sure, yeah, it has like an underwater vibe to it.

And then we have another guest from Kenny Dickey.

And Kenny said, Longtime listener, first-time guesser, new Patreon member.

He says, I'm a Canadian and I canceled a bunch of U.S.

streaming services to become a member to create funds to send to the SGU.

Keep doing what you're doing.

That's fantastic.

Thank you, Kenny.

We really, really appreciate it.

And God knows we need support now because Steve.

Oh, yeah, blame me.

So he said, the noisy sounds like the inside of a wind turbine during high winds.

There sounds like there's a long cable dangling down the main shaft.

Maybe it's in a shutdown sequence or something.

So I think that's a cool guess.

It is not that.

And nobody guessed it this week.

And I was really surprised because I think that this sound was, I don't know, probably came out maybe about six months ago.

This is actually the sound of

the movement of the inner core of the Earth.

Not just that, but this is the sound of the Earth's flipping magnetic field heard as sound.

You know, I don't think it was turned into sound waves.

I think this was a recording of it.

Didn't you do this one before?

No, I don't think so.

No, that sounds familiar.

If we did, people didn't guess it.

Yeah, if we did, it's like when you do science or fiction and you're like, we've already done all of these.

We're like, ugh.

So listen again.

It just sounds like there's things happening.

Yeah, I'm with you, Kara.

I don't like it.

Sounds too big.

It sounds like some big steampunk machine breaking apart on the inside.

Yeah.

About to kill us all.

Yep.

Yeah, it's got

like a demonic thing going there.

All right.

We've got a new noise a few guys this week.

This was sent in by a listener named Melissa Burke, and she told me how to pronounce her last name.

And Melissa, I would not have pronounced that correctly.

Let me play that song for you right now.

Okay.

If you think you know what this week's noisy is or you heard something cool, just like a lot of these other people, then you got to email me at wtn at the skepticsguy.org.

Steve, everybody knows what I'm about to say.

Not a con 2025 happening May 15th, 16th, and 17th in White Plains, New York.

Please consider joining us.

We're going to have a wonderful conference.

Lots of fun, lots of socializing, and partying and music.

It's going to be a really good time.

You can go to nataconcon.com to get all the information that you need.

You could also join the SGU mailing list.

Every week, Ian and I put together an email of everything that we've done the previous week and lots of other fun stuff in there and messages and secret stuff and pictures of Bob that he doesn't know.

I have all sorts of stuff you can get for free.

I know.

I don't have a flow yet from CARA.

Like, I don't have secret friends of CARES that can send me stuff.

Nope.

Nope.

Yeah.

Ouch.

I get that.

And as you guys know, Steve is coming full-time to the SGU.

And that means that this is a wonderful time to give us any support that you can to help us continue to do what we're doing.

We have a lot to fight back against this year in particular.

We have new programming that we're working on.

As a matter of fact, we have one of them that we are scheduling now.

We really have come a long way with that.

I'm hoping sometime in the middle of the summer that we will release episodes.

We will give you more details as we feel you need to know.

Right now, all you need to know is come to Naticon.

We really want to see you.

And what else do you want to say, Steve?

I want to say one other thing.

We are getting dangerously close to our Patreon goal where we need to do a live 24-hour streaming show.

Oh, God.

It's going to happen.

If we get this goal, I will drag each and every one of you kicking and screaming into the studio for a 24-hour show.

We will do it.

We will keep that pledge, and it will be fun.

It'll be awesome.

And our listeners can make it happen.

Yeah, if we end up doing that,

I promise everybody that

not only will it be fun, but I promise you that you will see something you've never seen before.

I've come up with ideas, Steve, because we've been talking about this.

Yeah, I want to do something different, though, this time, Steve.

I don't want us to spend just 24 hours in that studio.

I want us to have things happening.

Yeah,

things will happen, Jay.

Don't worry about it.

All right, I'm going to do one email.

This one comes from Dan.

And Dan writes: For over a century, physics has claimed to be advancing at every major cosmological mystery remains unresolved.

Instead of addressing why their models keep failing, physicists have relied on increasingly complex unproven assumptions, dark matter, dark energy, singularities, inflation, to patch over contradictions rather than solve them.

But the truth is, these mysteries were never real.

The errors were built into the models themselves.

The solution isn't exotic or hidden.

It was in relativity all along.

The universe is not expanding.

It is relativistically revealing itself at the speed of light.

Gravity is not an independent force.

It is suppressed charge expanding into three spatial dimensions.

Black holes are not singularities.

They are time-coiled neutral zones where relativity folds upon itself.

The beginning of the universe was not a past event.

It is a continuous absolute moment moving outward at sea.

None of this required new physics, just actually respecting relativity.

I know this because I built a model that does.

I'm not a physicist.

I'm a geologist.

A year ago, after suffering a severe neck injury, I spent my recovery rebuilding physics from first principles.

In doing so, I unified physics in months.

Not because I had special insight, but because the answer was always there.

It just...

it was just being ignored.

Meanwhile, mainstream physics is still stuck defending models that don't work.

Dark matter, never detected.

Dark energy, never detected.

Singularities, mathematically impossible.

Physics, off course for a century.

And until someone forces the correction, humanity is stalled.

We are at a crossroads.

Either we acknowledge these failures and move forward, or physics remains an ever-growing list of imaginary fixes to a broken model.

All right.

We get emails like this pretty frequently.

Wow, man.

Yeah.

And

the response that I gave Dan is along the lines of the response that I typically give.

Make a prediction.

Yeah, so I said, Dan, is there any way to test your model?

Have you spoken to a physicist to see if there is anything glaringly wrong with your model?

Have you submitted any papers for peer review?

I'll be blunt.

You are a non-physicist claiming to have overturned a century of physics in a month.

The probability that you are correct is pretty close to zero.

Now is the time for you to demonstrate massive humility and try to find out what is wrong with your model and why physicists believe what they do.

Otherwise, you are destined to be nothing but one in a long line of endless cranks.

Best, Steve.

So polite, but not

but I did not pull my conscience.

So it's like essentially like you know, I don't, you know, from experience, I have no desire to

spend any significant amount of time deconstructing this guy's physics or having an extensive back and forth with him.

Unless he can honestly address that response, there's just no point to it.

As I said, you're just at that point, you're just a crank.

Now, cranks may be smart, they may have a lot of knowledge and information, but there's something wrong with their process.

And I think Dan pretty glaringly displays

this is why I wanted to use this as an example.

Because some are like really really incoherent and

I'm concerned that I'm dealing with somebody who needs to have professional care.

I don't get that vibe here.

I think Dan is just doing it wrong.

As he said, he's not a physicist.

He built a model based on his understanding of relativity.

He thinks that he's somehow solved all the biggest mysteries in physics that have perplexed physicists for a century.

And he's done it, again, he said, in months.

And this is what separates, I think, think, cranks from non-cranks, right?

Cranks have a fatal lack of humility.

They get to this point and they think, I've overturned all of physics.

Rather than I must be wrong, I need to,

and I'm curious to find out why I'm wrong.

I should talk to an actual physicist.

And the trick there is

getting them to give you the time of day.

physicist I've ever spoken to tells me that they are inundated with this kind of stuff.

This, I've proven Einstein wrong, or I've solved all the big problems of physics, or I've unified everything, or I have a new theory of everything that explains everything, or whatever.

That's like this is constant, constant, constant.

We get a tiny fraction of it, of what physicists get.

Yeah.

Right?

So, but anyway,

if you really think you're onto something, that's the other thing.

I do this like when people with medical, you know, pseudoscience too, they tell me I've cured cancer.

It's like, okay, if you really think you've cured cancer, then it's the only ethical, morally responsible thing for you to do is to make sure that you are taken seriously by the establishment.

If you are going by the medical establishment, you need doctors to take you seriously, scientists to take you seriously.

If you just want to shout from the rooftops or, you know, basically now on social media about how, you know, how brilliant you are and

how there's a conspiracy to silence you, whatever, then you're a crank.

And that's all you will be.

You may make money doing that, but you will always only ever be a crank.

If you really think you're onto something, then you owe it to yourself, you owe it to the world to figure out a way to be taken seriously.

And that starts with humility.

That starts with humility.

The first thing you should do is try to prove yourself wrong.

Try to understand why the experts don't already believe what you believe.

Now, Dan did respond to me.

His response was, Oh, yeah.

Oh, really?

Well, okay, then, guess I'll drop this here.

Then he says, Alpha to 50 decimals, and he gives me alpha to 50 decimals.

Physicists aren't even looking in the right place for it, but hey, if you figure only physicists know physics, then I've got news for you.

They're way off.

I'm holding a complete rewrite of physics from first principles.

But if you're too good, I'll check with someone else.

Good day.

Good day, sir.

Good day, sir.

I said, Good day.

So, fine.

That's it.

I gave him my one and only chance that I will ever give him.

That was his response.

He doubled down on, I have rewritten all of physics from first principles.

And go right ahead, Dan.

Good luck with that.

You will never, ever be treated as anything other than a crank because that is how you are behaving.

It's not a criticism, it's not an insult, it's an observation.

That's the way it is.

If you want to be treated like a crank, act like a crank.

If you want to be treated like a scientist, scientist, act like a scientist.

That response was a crank response.

100%.

I've never gotten anybody to understand that.

Oof.

Ever.

And is that a very common response?

100%.

Yeah.

100%.

Totally.

You can get anybody who is really humbly trying to engage and say, okay, you're right.

I probably missed something here.

Not when they open up like that.

Not when they say, I've rewritten physics.

But when

you go there,

odds get even

less.

I've never pulled anybody back from that ledge, but I have gotten

constructive responses from people who were part of the way there.

They were like, am I onto something here?

I'm like, I don't think so, but this is what you should do.

And then sometimes they respond.

But generally speaking, no.

If you think you're smarter than all of the scientists in the world, there's probably no hope for you.

Yikes.

Yeah.

I mean, at the very least, run it by ChatGPT and see what that says.

I don't know.

I honestly don't know that I would trust ChatGPT with that.

For basic science, though, I think it's worth it.

It's worth a shot.

Go for it.

Go like, Dan, you're really honest.

Yeah, that's right.

I'm worried that Chat JPT will be like, yeah, this sounds good.

Yeah, that wouldn't surprise me.

It wouldn't surprise me.

Yeah, no, I wouldn't do that.

All right.

Well, we have a really interesting interview coming up with Michael Marshall and Cecil Cicarello.

So let's go to that interview now.

We are joined now by Michael Marshall and Cecil Cicarello.

Guys, welcome to the Scott Fixed Guide.

Hi, great to be here.

Love to catch you up.

Yeah, Marshall, we've had you on before.

Cecil, I believe this is your first time on the show, correct?

It is.

It is.

Thanks for having us.

Tell us a little bit about yourself.

Well, let's see.

I've been podcasting for, gosh, it's got to be since 2000, maybe eight or nine.

A newcomer.

I actually got inspired to start a podcast because of you guys.

Years ago, I was at a dead end job, like searching for things to listen to.

And I was, I remembered that I really enjoyed Art Bell.

And so I started listening to the old Art Bells and I thought, gosh, this is just so hokey.

It doesn't seem like it's real.

It's talking about crop circles.

And I googled in the in the search bar, I googled what's the opposite of a believer.

And skeptic came up.

And skeptic's guide was one of the first things.

You guys had awesome SEO.

in the early days and I wound up finding your show and I binged your show and you guys

you guys were an influence on me to start a podcast years and years ago.

I started a podcast called Cognitive Dissonance and I've been podcasting with my best friend Tom for over a decade and

now I just started a new show with Marsh here called The No Rogan Experience.

That's what you're here to talk about, right?

The No Rogan.

So tell us about your podcast.

So how best to describe it?

So Cecil, you came to me at the back end of last year with an idea to listen to what Joe Rogan actually has to say.

And I was really interested in this because Joe Rogan, obviously, he had a huge influence in the discourse, the conversation.

He has a massive, massive platform, and he's someone who gets criticized for a lot of the things that I said on his show.

But not a lot of people who criticize him probably are regular Rogan listeners.

I know that I'd criticized him for some of the like anti-vax stuff that I've been on his show, but I'd never actually listened to Joe Rogan.

And I felt like if we were to sit down and, you know, as Cecil suggested, actually listen to what Rogan was saying and what his guests were saying, and then see what

the appeal was to people and what you know what's the truth behind what he's saying?

How do we understand why this is reaching so many people?

We might be able to find a way of explaining to those same kind of audiences why this stuff doesn't really check out, whether there's a different set of opinions, different facts that you could follow up on and try to unpick the influence that Joe has on such a sizable amount of the public discourse and so, so many listeners.

And so, yeah, every week we listen to an episode of Joe Rogan.

And then, if he's wrong, which it turns out he frequently is, we explain where he's wrong and why he's wrong in ways that we think are kind of fair to him, not just like overly critical and roasting, but

pretty firm where it needs to be as well.

So, were you surprised by anything?

Like, did your opinion of him or his show shift after actually listening to full episodes on the regular?

Or

do you think it was pretty accurate, what you thought before?

Well, before we started, I actually went to listen to see if it would actually be a show.

Because like Marsh suggests, I hadn't listened to him.

I had always seen clips, everything taken out of context, never really an entire show that I've consumed before.

So I decided to sit and listen to him.

And as soon as I listened to a recent episode, I thought, my goodness, this would be so easy to do.

And so I wasn't super surprised.

What has surprised me is, is we are digging into the back catalog now.

And what has surprised me is that I have gone back to right around the beginning of the pandemic and we're doing a show that happened at that point.

And he has very different views.

I would say his views nowadays are more right-leaning.

Certainly he's platforming tons of people who are very right-leaning, who have lots of right-leaning opinions.

Lots of people.

He, in fact, had on the president of the United States before he was elected.

He had on the vice president of the United States before he was elected.

He had on the Doge head before he was appointed.

So he has sort of shifted to this right.

But when I listened to this 2020 episode, I was really shocked that, in fact, he was, he felt more libertarian, more centrist, certainly, certainly felt like he was more generous to left-leaning opinions at that point.

So he, I think his position has changed, and it's going to be interesting to sort of peel the onion back and see sort of when that happened as we work our way through the back catalog and continue to do some of the more modern shows.

Yeah, and for me, I think like, I do think my opinion has changed a bit.

I was expecting there to be a large part of the show where he's espousing some pretty out there conspiracy theories.

But what we've seen with even the shows that we've kind of covered so far is there are entire three, four-hour conversations that are just every single bit of conspiracism you can imagine.

And he's just spiraling through it.

And he seems to be very willing to just roll with almost any kind of conspiracy theory, as long as it doesn't touch on a couple of a couple of areas.

He's not on board with Flat Earth.

He doesn't like anything that criticizes Elon Musk or Donald Trump.

Any conspiracy theory that might shine them in a bad light gets shut down pretty hard, pretty quickly.

So yeah, he does just roll through everything.

But what I think has been really interesting is because we've now listened to multiple episodes over multiple months, you can hear him, you can hear where he's getting his, getting those opinions from.

So like there'll be one show where he's talking about his position on

Russia's invasion of Ukraine and he's explaining that actually the whole reason for Ukraine's aggression for Russia's aggression was because America backed a coup in the Ukraine and then they deliberately tried to provoke Russia and they that NATO promised they would never expand towards the east and yet they did.

And I thought those don't sound like kind of Joe Rogan's original ideas and lo and behold when you look back through his catalogue you can find the moment one of his guests brought those ideas to him.

And if you find the origin of those ideas originally, like some of them come from op-eds written in newspapers by Vladimir Putin.

So it is just literally Russian propaganda that's been laundered through one of his guests and is now something Joe brings up on the regular to other people and if you're a regular listener you'll hear him do that but you won't know that the origin of this was that conversation he had three months ago and the origin of that was vladimir putin's so that laundering is something that's really surprised me so he's gone down his own rabbit hole basically yeah

he's followed himself down a rabbit hole on his own social media but basically by the people he's choosing to interview and then listening uncritically to those people yeah i think so but i think it's also so he has certainly kind of radicalized himself But the people he has had around him, I think, seem to fall into a mix of absolute true believers who 100% are fully on board with what they're saying.

And another group of people who recognize, and it's, you know, it's very hard to kind of say how big this group is, but we've certainly seen examples of people who clearly seem to recognize that Joe is willing to follow them if they know which buttons to press and which of his biases to engage.

Like he had Mark Andriessen, the billionaire investor in cryptocurrencies, on the shore.

And Andrierson was talking about how this big free speech issue is that if you express political opinions that the Biden government doesn't like, they can remove your bank account from you.

And it could happen to any one of us.

Any one of your listeners, Joe, any one of your viewers could have their bank account taken away in a moment.

And when he was bringing up examples, all of his examples were cryptocurrency companies that were denied bank accounts because they were on a very sort of risky financial basis.

And so that's a very different thing to not being allowed a bank account because of your political opinions.

But because he knew to hit this free speech and this kind of political oppression button, Joe was completely unquestioning and accepting that, yeah, it's a real risk that any one of us could be debanked at any moment.

And it seemed pretty clear to me that Mark knew what he was doing and which buttons to press to try and smuggle these anti-regulation rhetoric and talking points into the conversation, knowing that there'd be millions of people who'd swallow them if Joe Rogan's show was the delivery mechanism.

And to call up another Mark, Mark Zuckerberg did almost the exact same thing.

Mark Zuckerberg came on, started talking about free speech and how his company is being attacked for free speech, et cetera, et cetera.

And he says we're being sued in the EU.

And what he leaves out is that the EU is suing him for privacy concerns, not for free speech concerns.

But he brings it up on Joe's show to make it look like that Facebook is fighting this free speech fight constantly so he can launder his message and basically read his terms of service to Joe's audience.

And Joe lets him do it because all he had to do was just rub Joe's belly in the right way, and Joe will let him pass and do whatever he wants.

Do you guys get the feeling or sense that he is researching any of these guests, like whatever their claims are in any way?

That's an interesting question.

I don't think he's doing a lot of deep research on them.

He's certainly not like, he's not questioning whether they're right by the time they get on, but he's certainly having pre-conversations that seem to kind of map out

what direction the show is going to go.

Because when he's talking, when he's got Mark Zuckerberg on and Mark Zuckerberg's talking about how actually it's really unfair that Facebook's getting kind of maligned in these ways, it's pretty clear they're going back through a conversation they've done ahead of time.

So, in those kind of situations, it's clear that he is doing some research, but that research is, hey, what do you want to say?

And he's had Elon Musk on, and it very much feels like a mopping-up exercise for some bad PR that Elon's had.

But then there are other guests that we recently were actually just about to put out a show where he had a recent conspiracy theorist called Ian Carroll on.

And Ian Carroll's whole show is him talking about how

Charles Manson was

recruited through LSD with MK Ultra and how there are elites out there who are using like horrific assault in order to cause multiple personality syndromes in people in order to turn them into weapons and use them as assassins to work for like shadowy mossad kind of organizations.

And when I was researching some of the claims that Ian Carroll's making on that show, I found a two-hour presentation on Rumble of him making exactly the same points, almost in the same order.

And it feels like if Joe hadn't seen that, whoever's booking Ian Carroll must have seen it, because Joe did not seem surprised that that was the flow of conversation that was coming.

There was very clearly some bullet points they were getting through.

So I don't know that Joe is spending a great deal of time sort of researching whether his guests are right, but he does seem to have some prior understanding of what it is they're there to talk about.

And when he had

Kai Dickens on, who did the telepathy tapes, he had a conversation with her that he seemed like he might have listened to the telepathy tapes, but he didn't listen super closely because at a certain point, she confuses him by saying the words,

the phrase spelling, as in facilitated communication spelling.

But he thought she meant actual spelling.

So like spelling in school as a as a class or learning, some sort of learning outcome.

He confused actual learning

how to spell with what she had called spelling.

And he did it pretty, pretty, pretty obviously in the show.

You could hear him sort of catch himself and say, oh, I thought you meant this.

And if he would have listened to the telepathy tapes all the way through, I don't think he would have made that mistake.

I think he probably listened to it a little and said, I believe this.

And then he had her on and just sort of let her amplify her message.

Guys, how often does he make a podcast?

Four times a week.

Four times.

And they are three to four hours.

I think one show we listen to is four and a half hours long.

It's not all interesting, to be honest.

I mean, we don't listen to it all, thankfully.

We tend to pick out the ones that we think this is a show that's going to be worth us listening to.

I imagine the shows that we don't get to listen to are no better than the ones that we choose.

This isn't that we have a selection bias, I think it's just more of a sampling issue, I think.

My limited research on him, like listening to his podcast and kind of reading summaries on what he believes and what his statements are, you know, I just get the sense that he's got a very misinformation-dense program

that is just kind of playing into his guests and playing along, you know.

But one thing that I was trying to do, I was actually being a little forgiving, going, does he really believe this or is he just doing this for the show because it works?

And do you guys have a sense on that?

I think he does kind of believe it.

So far, I haven't heard Joe feel like he's ventured into a territory where he feels like he's grifting at all.

I think some of these other guys, these other big sort of misinformation guys, if you think about like Alex Jones, I think sometimes there's ways you can sort of tell maybe Alex is playing a little bit up for the camera.

But I think Joe, I think he sort of has his own confirmation biases and those drive his show.

And those happen to be in line with all of this conspiracy misinformation that is, that is sort of polluting.

the information ecosystem.

So it fits really well.

I feel like he does buy into the stuff that he covers.

I think that's, I also think that he uses that as a shield to prevent people from sort of questioning, I guess, his morals on how he's platforming and creating a bunch of misinformation.

He uses that a lot of times to be like, hey, I'm just here I am just having these freewheeling conversations.

But I think, you know, when you have a platform that big, you, oh, you have to be responsible for some of the information you're putting out there.

And he's not at all.

Yeah, I think I'd agree.

Although I do think there are moments that aren't cynical, but they are at least calculated.

Because there's a point in which he's talking to Kai Dickens about the telepiotapes, and he's talking about billionaires.

And he said, the thing I don't understand about billionaires is once they get past a certain amount of money that they're never going to be able to spend, it's just numbers on a page.

Like, what is the drive to carry on adding money to that?

It's ridiculous to keep adding for looking for possessions when you've got so much.

It's like, that's a ridiculous thing.

The next day, he interviewed Elon Musk.

and that curiosity was entirely absent.

It's like, if you are genuinely curious about those things, Joel, why are you bringing it up to Kai Dickens, not a billionaire, but you don't bring those same questions up to the richest person in the world who might actually be able to offer some insight into that?

So I think there are some sort of calculated moments where he believes what he's saying at the time.

And I think he did believe when he's talking to Kai Dickens that there is, he's got this curiosity about what the drive is once you're so rich.

But he also knows that he's not to ask that question of anybody who could literally answer it because his bread is buttered on the side of keeping the billionaires on board.

So I followed him a little bit more closely early on in the podcast, where he would have

an eclectic variety of people on his show.

And he could kind of get away with: I'm just having conversations with all kinds of different people, and he's not really being selective.

But as you say, there's been a shift.

And recently, has there been any legitimate guests on, or is he pretty much all conspiracy theorists, grifters, and pseudoscientists at this point?

So I think there's certainly been some legitimate guests.

I mean, I know even in the last week, he had Michael Coster from The Daily Show, who I really like The Daily Show.

I really like Michael Coster on it.

Very different political opinions.

But the reason we didn't cover that show is because they just stayed away from politics and talked about sports and stuff.

So he will have legitimate people on, but

he won't have very many legitimate people on and talk about anything substantive.

I think Brian Cox was on recently.

We've not got to that show.

I'd be interested to know how that conversation went.

And what we have seen when he is talking to somebody who raises some good points is that in that moment he might accept those good points, but then the next time he's having a conversation, there's almost no trace of that.

So if he did talk to a scientist at some point about how climate change is genuine, that hasn't left a lasting impact on him at all.

Whereas when we do see him talk to these more extremist fringe figures, the things that they say, we can track them coming up again and again on his show.

And I actually think that's, in a way, that's the biggest impact he is likely to have had.

Whatever impact he had in the American election, it wasn't necessarily, in my view, the interview he did with Trump.

It's the fact that every single show afterwards, or on show after show after show afterwards, he kept bringing up how wonderful Trump is.

So it's more likely that his impact is in the repetition.

And I think there are, that's why there's savvy actors who know, if you get on the show and can smuggle your message in in a kind of a Joe-shaped pill that he can swallow, he'll keep regurgitating that at every available opportunity.

And it's worth way more than just the time that you're on the show with him.

Because when he's saying this when you're not on the show, it seems even more genuine.

And there's also the current stream, and this is, I think, been a constant throughout his entire podcasting career, of comedians that are in his orbit and fighters and sort of fight previews because he's the UFC announcer.

So that sort of thing still has a place on his show.

Still some pop culture there.

Yeah.

So there is still some of that.

But I will say, when the USAID stuff broke, he had three three or four people in a row that were all anti-USAID, pro-Doge on right in a row.

So he had on Brett Weinstein,

and then he had on Mike Benz.

And so, you know, those types of people all sort of have a very similar message.

And he's sort of following the right-wing talking points.

And in fact, you could kind of see that, you know, most of the information he gets is from Twitter.

And that's the place that he trusts the most.

If he finds a link while he's searching for something, when he asks Jamie to call something up, he will often stop on the Twitter link instead of the news link.

So he will ask Jamie to call Twitter up so that they can look at the Twitter link.

And so much of the stuff that he repeats on his show and the guests he has on his show are really sort of based in the Twitter ecosystem.

So you can figure that some of the stuff that they're going to pull from there is coming straight from Twitter.

Yeah, and there's a great moment in, I think it might be the next show that we put out, that illustrates exactly what Cecil is talking about, where where he's talking about something about you know missing kids in a in America and how there's all 300,000 kids have gone missing and even Jamie his producer points out that actually that's that's not true it's he goes away and fact checks it and he brings up a BBC article and Joe says oh well don't look at the BBC one because I'm sure I read it somewhere else and he tells Jamie specifically what words to put into Google to find the link that he wants that proves him right and he does that and they still scroll past three or four different news articles to get to the tweet that Joe was remembering which Jamie, who's normally very on board with him, points out, well, that's a tweet without even a source in it.

So, Joe is like so selective that he feeds the words into the search engine that he's specifically looking for for his

information and then skims, like skips straight past the legitimate sources to find that one tweet that he remembered.

So, his show, with his what, 14 million listeners is 20 million.

20 million, yeah.

I mean, he is a gurgling pot of misinformation,

Yeah.

Pretty much.

Yeah.

I think that's absolutely true.

I think that, you know, I mean, we, we, I was talking about the USAID show that we covered, this one with this one with Brett Weinstein.

Now, this was before he had on Mike Benz.

Now, Mike Benz is, I guess, a little more qualified to talk about this because he's he was part of the State Department, I guess, at some time.

But that's very generous, Cecil.

That's very generous to his CP, to his resume.

Very generous of how I'm framing who Mike Benz is.

But even less qualified is Eric Weinstein, who just happens to be one of these intellectual dark web guys who's having a conversation with him about other stuff, and they happen to bring up the USAID stuff.

And they quote a tweet with no link, just a tweet with a bunch of numbers in it for with buzzwords attached.

So they will say something like, they spent $20 million on Iraqi Sesame Street.

They spent $2 million on Moroccan pottery classes, et cetera, et cetera.

But it's literally just a tweet he's reading aloud.

So we can see that he's just finding these things in the Twitter ecosystem and presuming they're true.

And it's not just this one time.

Multiple times they have said, we do not pay attention to mainstream media.

We go to these places, and Joe is mainstream media now.

He's even more mainstream than any other media out there.

But what he'll say is, I don't go to those places.

Where I go is to Twitter and I go to podcasts, and that's where I get my information.

And so he's teaching people bad information hygiene themselves.

Yeah, and if I could just add to that, Cecil, you're absolutely right.

But then the figures he's picking out on Twitter would be relatively fringe figures in any other version of the world.

But that information, that might be from a very small account or very unverified account, then gets amplified to the 14 million or 20 million people who are watching that Joe Rogan episode.

So it is laundering these niche misinformation actors into a massive, massive audience.

And then we've even seen plenty of examples where the things that get said on Rogan Shaw get picked up by actual newspapers to say, oh, here's when Elon was on, and here's what they talked about, and here are these specific claims.

And they get picked up by newspapers, they get picked up by that kind of misinformation blogger sphere who just

and sort of that misinformation social media sphere who then just amplify and repeat those exact things again.

So it's it's he he does and I don't think it's cynical, I don't think it's deliberate, but just his bad information diet means he'll pick out the most rotten apple in a very large barrel and then present it to one of the world's largest audiences as legitimate.

And do you think, I mean, to what extent are bad actors picking up on this?

And as you say, like Vladimir Putin putting out propaganda that ends up on Joe Rogan's show, that's like, you know, Christmas for Putin.

So, how many, do you think that there are people out there who are now like, all we have to do is tweet this to Rogan, and this information, this misinformation or disinformation will get out on Rogan's show.

And to the extent that that's true, you better believe the Russians are doing that, right?

Have you encountered that where you thought, like, that was not just bad information hygiene, that's deliberately placed disinformation.

So

I can't put my finger on something that was specifically Russian, but I do think, for example, the Mark Andriessen thing is a perfect example.

You know, when he's talking about the debanking, now that is a the goal he had in saying that your bank account could be removed was to demonize the CFPB, is that what it is, Consumer Financial Protection Bureau.

He was trying to demonize that to kind of build a permission structure for the removal of the protections that that agency affords.

And if you come along and say, I am a billionaire investor in some quite high-risk areas of finance, and this agency keeps limiting my ability to make money, I want to get rid of it, Joe's listeners and viewers aren't going to be on board with what you're saying.

But if you can feed them this debanking narrative, they will be calling for the removal of

those regulations.

And lo and behold, the clip of Mark Andreessen talking about that gets shared on Twitter by Elon Musk and gets millions and millions of views and feeds into the Doge kind of the Doge project and becomes part of the mainstream media's narrative.

And I think that is that smuggling of disinformation because you had a billionaire who had a political agenda, which was to weaken the regulations that are restricting his ability to make money.

And he recognized that he had an ally in Elon Musk to try and get rid of those things, but you need to build the permission structure to make sure that people aren't complaining and you can feed your disinformation to this

super spreader, this combine harvester or fertilizer or whatever it is of

flinging this stuff around and it will get as wide as you want it to go.

And Elon Musk did something very similar.

We covered the show just last time and Elon Musk came on to talk about all the Doge cuts and how they were justified.

And he was literally just making up the idea that you didn't even have to have any kind of permission at all.

You just walk into any of these places and say, I want a grant.

And then they just grant you the money as if

there's no process for them to actually vet you at all.

You just ask for it.

And he was saying that these people who had these grants, all they had to do was

send him a picture of the work that they're doing and they wouldn't even do that.

And he's just lying to people to make it look like the cuts that he's making are actual cuts that will affect the bottom line and that are justified.

And I think Elon came on to sell that to the American people and to sell the fact that his gesture that he made on stage wasn't

a Nazi salute.

It was in fact just a very enthusiastic arm gesture.

And they spent much of their time talking about those two things the entire episode to try to launder Elon's image with the public so that when Joe's listeners go away and people say, Man, can you believe what Doge is cutting?

They'll say, Well, actually, those Doge cuts are good.

And Social Security is a Ponzi scheme.

And so they'll just keep on repeating the talking points that Elon has planted in the audience so that the audience will then follow whatever Elon had said.

And Joe is a willing participant.

Yeah, absolutely.

And just to not sidestep the point about Russia, given that we can see this from these actors, I completely agree that this is a wide open goal for Russian disinformation.

And we have seen Tucker Carlson going on Rogan after he's been out to Russia and talking about

how great Putin actually is, or talking about spreading false narratives about Russia's invasion in Ukraine.

He had Lex Friedman, who as a podcaster who went out to interview Zelensky and

try to get Zelensky to seek peace by essentially conceding to Russia, which was very obviously playing into the

political goals of Russia.

Now,

why Friedman was doing that?

We'd have to kind of follow back to see where he was getting his bad information from.

And I wouldn't be surprised if

we could follow that far enough back, it would have Kremlin origins, even if he didn't realize it himself.

But yeah, so it's not that Russia is targeting Joe, I don't think, but I would be hugely unsurprised to see that they were targeting the kind of people that will end up on Rogan, knowing full well that those people are probably an easier target to get the initial message through.

And then once it's with Joe Rogan, there's no fingerprints around him because he's getting it organically from

this other American source.

All right.

Well, Marsha and Cecil, thanks for joining us.

So for our audience, this is the No Rogan Experience Podcast.

That's K-N-O-W-R-O-G-A-N.com.

We'll have the link on our show notes.

Sounds like a good listen.

Guys, thanks so much for coming on.

I'm going to say I'm not going to sleep well tonight, but it was nice talking to you.

Yeah, I mean, I didn't anticipate this when when we started listening, but it's not been great.

It hasn't been great to listen to you.

It's not been good, but thanks for having us today.

All right, well, thanks for joining us, guys.

Thanks, guys.

Thanks for joining us.

It's time for science or fiction.

Each week, I come up with three science news items or facts, two real and one fake.

And then I challenge my panel of skeptics to tell me which one they think is the fake.

And you know what?

You can play at home too if you wish.

You don't need any equipment.

All right.

We have a theme this week.

The theme is invertebrates.

Okay.

Here we go.

Item number one: jellyfish are in the same invertebrate phylum, echinoderms, as sea cucumbers and sea urchins.

Item number two: invertebrates make up 97% of all animal species with 30 phyla, compared to only one phylum for vertebrates.

And item number three, invertebrates not only lack a backbone, they completely lack any bone or cartilage at all, either internal or external.

Jay, go first.

Now, first of all, isn't it jellies not jellyfish?

I know, but I still see jellyfish so frequently in even scientific references.

I'm just going to say jellyfish.

But yeah,

some people technically say you should just say jelly, but.

Or sea jelly.

Okay, either just jelly or jellyfish.

i haven't seen sea jellyfish oh i see sea jelly all the time i know it's sea star sea jelly starfish no sea jelly i don't think anybody really knows so let's yeah it doesn't matter it's the pluto sure the planets doesn't matter okay so we got jellyfish are in the same invertebrate phylum uh

kinoderms uh as as sea cucumbers and sea urchins um now i have uh i have been in close vicinity and have actually held sea cucumbers.

I've seen sea urchins and I've also seen jellyfish.

And, you you know, on the outside, they don't look that related, but you never know, right?

I don't have anything to say like, you know, like that I've read or whatever, but, you know, just the visual scan, I'm saying maybe not.

Next one, invertebrates make up 97% of all animal species with 30 phyla compared to only one phyla for vertebrates.

Wow, make up 97% of all animal species.

That's interesting.

Okay.

My gut is telling me there's something wrong with that too.

Invertebrate number three, invertebrates not only lack a backbone, they completely lack any bone or cartilage at all, either internal or external.

I don't know, man.

That doesn't sound right to me either.

I think that one sounds the most shifty.

I'm going to say that one's the fiction.

Okay, Bob.

Yeah, these first two here, you know, the phylum, you know, with sea cucumbers and sea urchins and jellyfish.

Sure, it could be at the phylum level.

I just don't have a good enough sense of, you know, how,

you know, how distant the phylums, you know, the different the phylums could be in terms of all these species.

What's underneath phylum anyway, besides species?

It's, I don't have a good sense of how high up.

But yeah, so that's that should be totally fine, or it could be wildly ridiculous.

I just can't tell the difference.

Um, so that kind of sucks for me.

97% invertebrates making up 90%, 97%.

That sounds like a lot.

And that's, I,

but, but then again, it's like, uh, could be, it could be true.

But I actually, I agree with Jay.

Three rubbed me the wrong way viscerally.

So when you, as you read, I'm like, no, I bet you there's something in there

that some cartilage,

I would think,

even in, at least in some invertebrates.

So I'll just go with Jay here, GWJ, and go with the invertebrates number three and the backbone cartilage dinghy.

Okay, and Kara.

So I used to teach intro bio.

And I feel like you guys are wrong, like very wrong.

And I could be wrong too.

I don't have absolute knowledge about this, but I'm trying to remember back.

I have no idea if the 97% number is correct, but if 97% of all animals are invertebrate, and remember, that's every like insect on the planet, then like, yeah, that makes sense to me.

And that there are 30 different phyla compared to one.

Okay, if that is true, it makes my suspicions about the first item stronger, which are if there are 30 different phyla, and I remember jellyfish or sea jellies being cnidarians, but I don't remember if that's a phylum or if that's a class or an order.

Right?

Who remembers that?

I think they're different, exactly.

But I think they're different enough from like sea urchins, like spiny sea urchins and sea cucumbers that are really structural.

I think they're different enough that they're not the same phylum.

But I definitely buy the final one that they don't have bones or cartilage.

Yeah, like fish, fish have bones, sharks have cartilage.

Right.

These aren't not invertebrates.

These are all vertebrate species.

Invertebrates have have exoskeletons.

They don't have bones.

I think it's that jellyfish are not echinoderms.

I think that's the fiction.

Okay, so you guys all agree on number two, so we'll start there.

Invertebrates make up 97% of all animal species with 30 phyla compared to only one phylum for vertebrates.

You guys all think that one is science, and that one is

science.

That is very interesting.

So yeah, 97% sounds high.

It sounds dramatic.

When you realize that includes insects, including beetles, you know, and all that stuff, that's, yeah, that it makes sense.

There's a lot of, you know, most animals are invertebrates.

And phyla is right below kingdom, right?

So it's kingdom, animal, phyla, there's 30 phyla of invertebrates, one.

So basically, we take all of the animals and we divide them into two groups.

One phylum and then every other phylum, which we define as being not not the first phylum, right?

You have vertebrates and then invertebrates, which is most life, most animal life on earth.

It's very human-centric, right?

Think about that.

Think about dividing up all of animals into those two groups.

Things like us and everything else.

And everything else.

And the rest.

So, yeah.

To first approximation, most animals are invertebrates.

And there's one little tiny little branch of vertebrates that we happen happen to belong to.

Okay, let's go back to number one.

Jellyfish are in the same invertebrate phylum echinoderms as sea cucumbers and sea urchins.

Kara, you think this one is fiction.

The boys think this one is science.

And this one

is

the fiction.

Because Kara is correct.

This is why Kara went last.

I figured you were going to get this one.

So, yeah, jellyfish or sea jellies or jellies are cnidarians.

What other animals are cnidarians?

Ooh,

okay, so they're the floaty.

They're just as crazy, which is why I thought

corals?

Corals.

And corals.

And sea anemones.

Sea anemones.

Anemones, yes.

Okay, anemones.

I didn't realize corals were cnidari.

Yeah.

Oh, that's.

So corals and jellyfish are in the same phylum.

Right, because corals aren't actually, I mean, they are corals.

They're colony creatures.

Yeah, they're colonies.

Yeah, exactly.

And then with fronds like these, who needs anemones?

Ah, stop.

That's out of Nemo, finding Nemo.

Okay, anyway, Evan would be proud, regardless.

Thank you.

And then echinoderms do include sea cucumbers and sea urchins, but one other group?

Sea cucumbers, sea urchins, and sea stars.

Stars.

Yes, this is, this is.

What about sand dollars?

I think that's also echinoderm.

Okay, yeah.

Then you have mollusks, which are snails, slugs, squids, octopuses, segmented worms like earthworms and leeches, sponges, another group.

Arthropods, insects, spiders, crustaceans.

Yep, those are the invertebrates.

Which means that invertebrates not only lack a backbone, they completely lack any bone or cartilage at all, either incarnate or external.

Yeah, that whole developmental pathway

did not exist in the invertebrates.

That was a unique innovation of the vertebrate line.

Yeah, they're made of a lot of different types of

types of.

Yeah, they could have exoskeletons, right?

Yeah, but that's not

bone.

Bone is a specific type of tissue.

It's not anything hard, right?

It's a specific type of tissue.

They do not have it.

Okay, we believe you.

Yeah, I'm just.

I'm trying to remember.

Yeah, I think they're chitin.

They're made out of chitin.

Yeah, yeah, the actual cells are made out of chitin.

There may be other ones, too.

Like, I don't know if

clam shells are also chitin.

I don't know.

I think they're just calcified.

They're calcified.

No, yeah, that's probably calcium carbonate or something.

So calcium carbonates, yeah, right.

But not but not cartilage/slash bone.

Roy.

Roy.

Yeah, that was fun.

Good job, Kara.

Thank you.

Evan's not here, so I came up with a quote for this week.

Of course you do.

You ready?

No.

Nothing in life is to be feared.

It is only to be understood.

Oh, I love it.

Yeah, right?

Love it.

That's who said that.

Nice.

If I say this was by a female scientist,

who said it?

That sounds so familiar.

There are a lot of female scientists.

I know that,

but most people can only know.

Oh, they'll say Marie Curie.

Marie Curie.

Seriously, the vast majority of people who name a female scientist, it's literally the only one they could name.

Marie Curie.

Yeah, very, I love that.

I I love that sentiment.

Yeah.

Because I do think, you know, the flip side of that is not understanding things leads to fear.

Yeah, it is.

And that's a huge thing that I talk about with patients, you know, around death, right?

And around like end-of-life things.

But I will say, a little tongue-in-cheek, maybe Marie Curie should have been a little afraid of reacting to the fear.

Yeah, I was going to say that too.

Yeah.

Her attempts to understand it as well as she did caused her demise.

But, you know, there's risks to life.

You know what I mean?

You can't live in a bubble.

Who could have thought?

Would you do that?

Take that trade-off ahead of time.

You're going to win two Nobel Prizes, be world famous, be the most famous female scientist to ever live, but you're going to die a little early of radiation poisoning.

Right.

And we didn't know.

Like, she's why we knew.

Yeah,

there was that small outside chance she could have become a superhero, you know, that's always,

you know.

Yeah.

Maybe that was her angle, you know?

Yeah.

She was rolling the dice.

Come on, superhero.

Damn.

Didn't she have like the radium, like in the little thing glowing, and like keep it, you know, by her bedside?

Yeah.

Yeah.

It's probably a bad idea.

A lot of exposure.

A lot of expression.

But how cool would that have been before we knew?

Like, look, it's like a perpetual light.

Yeah.

Pretty neat.

Yeah, I would have loved that.

I would be like, yeah, put it by my bed, put it in my pocket.

Remember, this is at a time where they were making pictures out of radium.

Oh, and people were taking it.

Drinking radioactive tonics for health.

Yeah, for health.

Yeah.

Give you radioactive energy.

All right.

Well, thank you all for joining me.

The show.

Sure, man.

Thanks, Steve.

And until next week, this is your Skeptic's Guide to the Universe.

Skeptic's Guide to the Universe is produced by SGU Productions, dedicated to promoting science and critical thinking.

For more information, visit us at the skepticsguide.org.

Send your questions to info at the skepticsguide.org.

And if you would like to support the show and all the work that we do, go to patreon.com/slash skepticsguide and consider becoming a patron and becoming part of the SGU community.

Our listeners and supporters are what make SGU possible.

At Larson, we've perfected storm doors, like the Larsen 60 Maximum View Wisher Latch.

It's a guardian, keeping your little escape artists securely inside.

The defender, protecting against what you don't want with the most secure, first-ever magnetic latching technology.

When you hear, you know your 60 maximum view is secure with Surelatch.

Larson, it's not just a storm door.

Find us in aisle or learn more at larsondoors.com slash Surelatch.