The Skeptics Guide #1103 - Aug 29 2026

What's the word: Encephalopathy; News Items: AI Assisted Research Training, Robot Outruns Humans, Latest CTE Study, Bacteria Share Survival Genes; Who's That Noisy; Your Questions and E-mails: Peptides vs Proteins, Stellarator; Science or Fiction

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:02Speaker 00

  You're listening to the Skeptic's Guide to the Universe, your escape to reality.

• 0:09Speaker 01

  Hello and welcome to the Skeptic's Guide to the Universe. Today is Thursday, August 27th, 2026, and this is your host, Steven Novella. Joining me this week are Bob Novella.

• 0:18Bob Novella?

  Hey, everybody.

• 0:19Speaker 01

  Kara Santa Maria.

• 0:20Cara Santa Maria?

  Howdy.

• 0:21Speaker 01

  And Jay Novella.

• 0:22Bob Novella?

  Hey, guys.

• 0:22Speaker 01

  So, we had some recent celebrity deaths this past week.

• 0:27Cara Santa Maria?

  Oh, my God.

• 0:27Speaker 01

  Dolly Parton.

• 0:28Cara Santa Maria?

  So sad.

• 0:29Speaker 01

  I know.

• 0:30Bob Novella?

  Two bad ones, man. Tim Curry, ah.

• 0:34Speaker 01

  Tim Curry. Both were 80, which you know, is 80 is not like tragically young, but that's not that old these days, you know?

• 0:42Cara Santa Maria?

  No, and for women, I think, well, actually, what is the average age?

• 0:46Bob Novella?

  In our country, I mean, it's not as good as it.

• 0:49Cara Santa Maria?

  It's not as good as it used to be.

• 0:50Speaker 01

  No, not true. It's higher, higher this year than it's ever been. That is not true.

• 0:53Cara Santa Maria?

  Oh, yay. Okay, so for women, it's 81.4.

• 0:56Bob Novella?

  Not as high as it should be.

• 0:58Speaker 01

  That's a different question.

• 0:59Cara Santa Maria?

  Yeah, that's what, yeah.

• 1:00Speaker 01

  It took a dip in COVID and then recovered, and in the last year that we have the data, it was the highest it's ever been. So it's resumed its upward course.

• 1:08Cara Santa Maria?

  So, so he.

• 1:09Bob Novella?

  But still, not in the top 10, I don't think.

• 1:11Cara Santa Maria?

  He lived longer.

• 1:13Speaker 01

  Women have a life expectancy that's five years longer than men.

• 1:16Cara Santa Maria?

  Yeah, so he lived longer than expected.

• 1:18Speaker 01

  Yeah.

• 1:19Cara Santa Maria?

  But she, she just missed it by 1.4 years.

• 1:22Speaker 01

  And yeah, that's at birth. You have to also clarify for that. Like, when you're 70, your life expectancy isn't 80, right?

• 1:28Cara Santa Maria?

  Oh, right.

• 1:29Bob Novella?

  It goes up, yeah.

• 1:29Speaker 01

  It's a lot, it keeps going up, yeah.

• 1:31Cara Santa Maria?

  I'm gonna look that up.

• 1:32Speaker 01

  Because a lot of that is.

• 1:32Bob Novella?

  You could live forever, though.

• 1:35Speaker 01

  A lot of that is, you know, infant mortality, for example.

• 1:38Cara Santa Maria?

  Wow, if you are 80, or sorry, if you are 70 years old in the US, your average total life expectancy is 83 to 87 years.

• 1:46Speaker 01

  Yeah.

• 1:47Bob Novella?

  That's awesome.

• 1:47Cara Santa Maria?

  So for women, it's 86 to 87. Yeah, so, but, you know, she, from what I read at least, she had, she had cancer. I don't know many of the details, because I don't know if many of the details have been released, but she had cancer and it was a relatively kind of quick, difficult thing.

• 2:02Speaker 01

  Yeah, that's the thing. When you're that age, you are susceptible to acute illnesses, which can kill you.

• 2:06Cara Santa Maria?

  Yep, yep.

• 2:07Speaker 01

  And then also, Hayden Panettiere died.

• 2:10Cara Santa Maria?

  Yeah, yeah.

• 2:11Speaker 01

  Recently, very tragically, of an overdose. She's, you know, she was young, she was 36.

• 2:17Bob Novella?

  Who was she, Steve?

• 2:18Cara Santa Maria?

  Actress.

• 2:18Speaker 01

  She's an actress, she was the, she was the main actress in Heroes, which I, just by coincidence, was rewatching.

• 2:24Cara Santa Maria?

  Hmm.

• 2:25Speaker 01

  Yeah, you know, just a childhood star tragic story, you know, just her life was a mess.

• 2:30Cara Santa Maria?

  Well, dealt with a lot of mental health issues, but was very public about them, and I think helped a lot of people through her own sort of openness about her struggles.

• 2:38Speaker 01

  Yeah, very unfortunate.

• 2:39Cara Santa Maria?

  Mm-hmm.

• 2:40Speaker 01

  All right, well, let's get right into the show, guys. Kara, you have a what's the word?

• 2:44Cara Santa Maria?

  I do. Along this depressing note, no, I am going to cover the word mostly because it will be relevant later in the show. I'm going to cover the word encephalopathy.

• 2:56Bob Novella?

  Ooh.

• 2:57Cara Santa Maria?

  I want, I should have, Jay, I should have typed it in the chat and said, Jay, what is the word today?

• 3:02Speaker 01

  Oh, God, thanks, Kara.

• 3:04Cara Santa Maria?

  So, encephalopathy is one of those words that, you know, Steve obviously has had to use all the time. You can immediately break it down into its component parts. It's one of those beautiful, clean biomedical terms that is exactly what it says it is. And sometimes I think it's good to go back to those roots. So, it comes from the Greek root for the first half of the word, encephalo, which we could also break down into two other components, where there's the N part and the cephalo part. So, taking the N off of the front of the word, the EN, what does cephalo mean? Anybody but Steve. Cephalo, like cephalopod.

• 3:42Speaker 01

  Like encephalitis.

• 3:44Cara Santa Maria?

  Like encephalitis, yeah.

• 3:44Bob Novella?

  Basically, the brain.

• 3:45Cara Santa Maria?

  Take the N off, now just cephalo. Like a cephalopod.

• 3:50Speaker 01

  Yeah.

• 3:50Cara Santa Maria?

  An octopus is a cephalo.

• 3:53Speaker 01

  Has more than one brain.

• 3:53Cara Santa Maria?

  Nope, it's a cephalo and a pod. It's a head, foot. That's cephalopod. Head, foot. It's a head, foot animal, because it has a head and feet. So, encephalo now means we're inside of the head, which refers to brain. So cephalo is head, encephalo is brain.

• 4:08Bob Novella?

  Interesting, okay, cool.

• 4:09Cara Santa Maria?

  So that's a good.

• 4:10Bob Novella?

  That makes total sense now, all right.

• 4:11Cara Santa Maria?

  And then pathy, right, encephalopathy.

• 4:14Bob Novella?

  Disease, you know, disease.

• 4:16Cara Santa Maria?

  Yeah, and what do you think that originally comes from, the, the kind of root word for that? It's the same root for pathos, so.

• 4:24Bob Novella?

  Oh, pathos.

• 4:25Cara Santa Maria?

  Yeah, it comes from, yeah, this, this kind of core meaning of like suffering, of passion, and eventually it came to mean disease. So, encephalopathy is a disease of the inside of the head or the brain, but it actually means something a little bit more specific than that. So, it's, it's real definition is any disorder or disease of the brain, but often we use it to refer to chronic degenerative conditions. So, encephalitis, for example, itis usually refers to some sort of like inflammation, swelling, like, like, you know, the tissues are kind of angry, whereas encephalopathy is a broader definition. And then if we actually

• 5:10Cara Santa Maria?

  get into the different kinds of encephalopathies, there are, there are quite a few. So, there's chronic traumatic encephalopathy, CTE, which we'll be talking about here in a little bit. There's Wernicke's encephalopathy, which comes from a B vitamin deficiency, Hashimoto's encephalopathy, hepatic, hypoxic, acute necrotizing, mitochondrial. And here's one that you guys have probably all heard of, bovine spongiform encephalopathy, otherwise known as mad cow disease.

• 5:41Speaker 01

  Mad cow disease, yeah.

• 5:42Cara Santa Maria?

  Yeah, yeah.

• 5:42Speaker 01

  That's not good.

• 5:43Cara Santa Maria?

  Yeah, so actually, interestingly, prion diseases are a subgroup of encephalopathies, usually spongiform encephalopathies. So, there are a lot of different ways to, to slice and dice it, but if you look up prion diseases, their kind of more technical name is transmissible spongiform encephalopathies. So, they also lead to this pathology in the brain. It's a pretty broad word, and its root has a broad meaning, but in practice, it usually does refer to kind of the larger bucket of brain diseases that tend to be degenerative in nature. And then you have more specific terms sometimes for different things, but like

• 6:26Cara Santa Maria?

  Steve, for example, Alzheimer's disease would be technically a type of encephalopathy, right?

• 6:32Speaker 01

  Yeah, I mean, encephalopathy is a maximally nonspecific term.

• 6:36Cara Santa Maria?

  Right.

• 6:36Speaker 01

  Right? And actually, there's multiple different ways you can modify it. There is chronic encephalopathy, but you can have an acute encephalopathy, a metabolic, like a category is a metabolic encephalopathy. You were giving a lot of specific pathology types.

• 6:49Cara Santa Maria?

  Yeah, I was giving names of actual, like, things that you would see in a chart, for example.

• 6:54Speaker 01

  You could modify it with anything, time course, you know, acute versus chronic, specific pathology, or place in the brain. So, what do you think a leukoencephalopathy is, Kara?

• 7:03Cara Santa Maria?

  It's not in a leukocyte?

• 7:04Speaker 01

  No.

• 7:05Cara Santa Maria?

  Oh, interesting.

• 7:06Speaker 01

  But the root is the same, but.

• 7:08Cara Santa Maria?

  Yeah, what is the, oh, no, I don't know what that prefix means.

• 7:11Speaker 01

  Ooh, what's a leukocyte?

• 7:13Cara Santa Maria?

  Oh, my goodness.

• 7:15Speaker 01

  What is a leukocyte?

• 7:16Cara Santa Maria?

  It's a blood cell.

• 7:17Speaker 01

  What kind of blood cell?

• 7:18Cara Santa Maria?

  A white blood cell.

• 7:19Speaker 01

  Yeah.

• 7:20Cara Santa Maria?

  So, does it mean white? Oh, it does mean white. Okay, so where in the brain is it really, really light? Where is it?

• 7:25Speaker 01

  It's white in the.

• 7:27Cara Santa Maria?

  Oh, white matter. Oh, that, oh, okay, okay.

• 7:29Speaker 01

  So, that's an encephalopathy specific to the white matter of the brain.

• 7:32Cara Santa Maria?

  That's so funny.

• 7:33Speaker 01

  Leukoencephalopathy.

• 7:34Cara Santa Maria?

  When you asked me that, I was thinking, so specific, I was like, well, it's not substantia nigra, because that's where there's pigment. So, where is there no pigment, no white matter? That makes sense. This is what I was just saying to my partner as I was talking about prepping for this. I was like, if you study, if you end up going into medicine or even biology, and you studied Latin or Greek in your undergrad intensively, you have such a leg up.

• 7:58Speaker 01

  Absolutely. At least to the terminology, yeah.

• 8:01Cara Santa Maria?

  Well, that's a big part of it, is just remembering what all these words mean. So, yeah, watch this space, because we're going to talk a little bit more about one specific type of encephalopathy soon.

• 8:10Speaker 01

  All right, thanks, Kara. All right, guys, this, this is a very interesting study that came out, that is, there's so many layers to it that I, we could talk about. So, this is a study looking at AI-assisted research training.

• 8:21Cara Santa Maria?

  Okay.

• 8:22Speaker 01

  There, you know, there's at least two layers here that I think are important. So, one is, what is research, research training? Like, what are you actually trying to teach? And this is at the undergraduate level, you know, students in science, for example, and about how to do, how to conduct research or ask meaningful research questions. And second is, how is AI being used to help this? So, it has a lot to do with the discussion around AI in education. I think this is potentially a very powerful use case. So, let's first talk about research. Jay or Bob, let me ask you guys, because Kara, you're trained in

• 9:04Speaker 01

  how to ask research questions. What do you think are the components of a good research question?

• 9:09Bob Novella?

  Open-ended?

• 9:10Speaker 01

  No.

• 9:11Bob Novella?

  Uh, produces more questions.

• 9:13Speaker 01

  Well, that's sort of the, the feature of the outcome, of like the results of a good research question.

• 9:18Bob Novella?

  Okay, Jay, I shot my wad.

• 9:19Speaker 01

  Now, how do you devise?

• 9:20Bob Novella?

  Your turn.

• 9:20Speaker 01

  How do you devise a good research question? How do you come up with a good question?

• 9:24Bob Novella?

  Ask GPT.

• 9:27Speaker 01

  That is not a good way to do it.

• 9:28Jay Novella?

  I would imagine you're coming up with a situation where you're trying to find information, right? So I guess the end game would be that you have to know, you have to know what the, what, hold on a second. This is a fucking hard.

• 9:40Bob Novella?

  What about very focused, very specific?

• 9:42Speaker 01

  That is critical, Bob. You are, you have hit upon one of the creed.

• 9:45Jay Novella?

  Good job, Bob.

• 9:46Speaker 01

  So one is, you, you start with like a topic, like a broad category, but then the next step is to really focus it down. Is to come up with subtopics, smaller and smaller, until you have a very, very, very focused area. That's one of the, so now you have your really, really super hyper focused area. Then what do you do? It's not obvious, is it? You guys have been talking about science for the last 21 years, but this is like really tricky when you try to break it down operationally.

• 10:13Jay Novella?

  Well, you have to find good sources, no?

• 10:15Speaker 01

  Clear deliverables.

• 10:17Cara Santa Maria?

  How you formalize the question, right? Is this a measurable question?

• 10:21Speaker 01

  Yeah. So that's, that's a little bit towards the later in the process. So yes, but at some point you have to decide, is this an answerable question? Is this a question that it is possible to design a research study to answer? Versus when this is technically difficult or not focused enough or whatever, there's some, you can't operationalize it to an actual research protocol.

• 10:43Bob Novella?

  Okay, yeah.

• 10:43Speaker 01

  Right? So it's got to be answerable practically. You know, you know what I mean?

• 10:47Bob Novella?

  Yeah, that's like, oh, this is going to, you could do this, but it's going to cost a billion dollars. Well, that's not very practical.

• 10:52Speaker 01

  Well, we don't have the, we don't have the technology to look at the thing you want to look at or whatever. But there's some big thing in between focusing your area and then refining an answerable research question.

• 11:05Jay Novella?

  If you're, if you're not conducting tests yourself, you're literally relying on other people's work, then you have to, you have to be able to find credible work to study.

• 11:15Speaker 01

  Yeah, so.

• 11:16Cara Santa Maria?

  So the lit review.

• 11:17Speaker 01

  Yeah, you have to do a lit review. Yeah. So, and what are you trying to, there's like two basic questions you're trying to answer with the lit review.

• 11:23Cara Santa Maria?

  Where are there lacks in knowledge? Like where are there holes in the?

• 11:26Speaker 01

  So what do we know now?

• 11:27Cara Santa Maria?

  Yeah.

• 11:28Speaker 01

  And what's the gap?

• 11:29Cara Santa Maria?

  Yeah.

• 11:29Speaker 01

  So, so first you do like a literature synthesis and then you have to find a gap in our current knowledge.

• 11:37Cara Santa Maria?

  You're describing the entire process of getting a PhD. It's so fun.

• 11:41Speaker 01

  That's exactly correct.

• 11:43Speaker 01

  Because that's, but PhD is geared towards being a researcher, right? And so this is exactly what you have to learn, how to go from, I'm interested in birds to this very specific research question, right?

• 11:55Jay Novella?

  And Steve, would it change much if your audience is different? Like if you're writing this to professors versus doing it for yourself or whatever, like I'd imagine.

• 12:03Cara Santa Maria?

  Well, I think.

• 12:04Speaker 01

  This is actual researchers.

• 12:05Cara Santa Maria?

  Yeah, you're going to assume that you're writing always for an academic audience at the beginning. But Jay, you're right. Depending on how multidisciplinary your field is, there are considerations there too. Like when I wrote my dissertation, I had to narrow my question into a psychological lens because there's a ton of literature from chaplaincy and from palliative physicians and from all these different sites on medical aid and dying.

• 12:35Speaker 01

  Yeah.

• 12:35Cara Santa Maria?

  So I had to say, well, what's the psychological lens that I could talk about this through? So that is a consideration as well. And we're at a point now where you just can't read all the literature on a topic. You can't. There's too much.

• 12:48Speaker 01

  So teaching this process has often been considered a bottleneck in education in terms of training. You know, basically scientists, like getting, getting them to understand, like just coming up with a really good research question that you could actually study is tough. It's difficult. So there's a couple of mistakes that people typically make. One is they don't focus enough. Or they don't, they, they come up more with like a question and then they, they cherry pick and backfill the literature to sort of fit their question. Rather than, you know, doing, going, do what you should go, like really, really synthesizing all the literature and identifying a proper gap in the knowledge.

• 13:30Speaker 01

  You know what I mean? It's easy to say, I want to study this and then just, you know, just look for studies that support your hypothesis or your formulation. This is where the researchers come and say, they want to improve teaching of students, of science students on how to do this process. And they want to know if AI can be used to improve that. They devised a LLM to do that and it's, they call it the Socratic questioner. So this is the process. So first they break down, they operationalize what I was talking about into eight steps. Topic selection, subtopic generation, keyword expansion. So basically you go from your sub, your subtopic,

• 14:13Speaker 01

  your very narrowly focused area and say, now what are all the keywords that would bring up papers in that narrow area? Then you do a literature search and synthesize the literature. That's four. Then you identify gaps in the knowledge. You come up with a specific question and then you refine it and iterate it and refine it until you have a good researchable question. That's six. Then you look at the literature to validate everything that you've already decided you're going to do. And then you seek expert peer feedback. Basically say, all right, you go to somebody who knows what they're actually talking about and say, what do you think, you know, about this?

• 14:50Speaker 01

  That's the eight steps. And then they had for steps two, three, four, five and six. So not one, you know, topic selection and not seven and eight, which is later verification. For the middle steps, they basically had the students in a course teaching this, use the Socratic questioner as a thought partner in developing each step. And basically what that means is the AI was just asking a bunch of questions of the student to, in order to challenge them. Say, well, is that really a good idea? Is that narrow enough? What about this? Did you think of that? And then just keep doing that and forcing the student to iterate and iterate that process until they were done with that step.

• 15:36Speaker 01

  Does that make sense? So in this case, the AI isn't generating ideas or isn't doing the work, right? It's just asking questions and it's forcing the students to think more critically about each step in this process.

• 15:51Cara Santa Maria?

  I love this.

• 15:52Speaker 01

  Yeah, right. I was very impressed by the whole thing.

• 15:55Cara Santa Maria?

  Cause like right now, what we rely on for this, to be honest, is library scientists.

• 15:59Speaker 01

  Yeah.

• 16:00Cara Santa Maria?

  Like most universities are going to have a library science department where they support the, the academic research. And A, I don't think most students take full advantage of them at all.

• 16:12Speaker 01

  And they certainly think that too.

• 16:13Cara Santa Maria?

  Yeah. And B, I think most students in a PhD program, they're sort of, like you mentioned, Steve, there's this assumption that they already know how to do this when they don't. And so it's kind of like, go forth, do well, you know?

• 16:28Speaker 01

  Because a lot of times, this is, this is taught through a mentor, not in any systematic way. And if you had a great mentor, you may be great. And if you didn't, you may be really lacking. So it's very variable. And a lot of people who advance, they have a good intuition about this, right? And you know, like the brilliant researcher people who could just intuit their way to an awesome question.

• 16:48Cara Santa Maria?

  Yeah.

• 16:49Speaker 01

  But, but that you can't teach that kind of intuition. You have to teach a process, a method, right? A technique. And that's what this is doing. It's operationalizing. It's breaking it down, saying each step of the way, you've got to do this thing. This is your goal and this will help you get there. This, again, this iterative process of forcing you to think more deeply about it. So the, the researchers, both quantitative and qualitative, you know, the quantitative study, looking at basically students using this was just like, how did you feel about each step of the way? The average response for each step was they felt that the AI helped a lot, right? That was one of the average response. So it was generally a very positive experience.

• 17:31Speaker 01

  And the outcomes were generally good. So their, their conclusion was this has potential as a teaching tool for this. Now, but in, if I read the whole paper, because it was really fascinating, because again, they, sometimes the most interesting part of these studies is the literature synthesis, right? That's a really great way to have an expert synthesize the literature for you because they do it at the beginning of their paper, right? The introductory section of their paper. I highly recommend that as a, as a place to go. If you can read the jargon for a good synthesis, where, where, what's the current state of the art? They will tell you at the opening of their paper.

• 18:06Cara Santa Maria?

  Yeah. And then it's what you do when you, for example, a dissertation is basically like a published, you know, peer reviewed, published research paper. That's like way too long. And so that's usually what your whole first chapter, maybe it's your second chapter, depending on how it's laid out. And it's, it's a lot. It's, it's, it's supposed to be all of it. Once upon a time, it was all of it, but now that's impossible.

• 18:28Speaker 01

  Yeah, yeah, yeah. Right, right. Because the literature is just so huge. That's why you have to narrow your, your focus quite a bit. So they, so in the discussion part, they were saying like, what are some of the problems with, with students not, you know, having this kind of assistance? And also the, these AI tools are not new, but they just haven't been used in this way. So if you use an AI tool, like an LLM to help you with this process, but not in this stepwise fashion, right? So not with a, a specific structured workflow, right? If you use it in an unstructured way, they said, what happens is the AI tends to suggest a very, what they call weakly justified

• 19:13Speaker 01

  perspectives or conclusions or, or whatever. Because LLMs are good at mimicking the structure of human thought and speech, right? But they're not deep thinkers themselves. So they kind of create this illusion that what they're saying is, is very intelligent, but it's actually very superficial. And that obviously doesn't work when you're trying to develop a reason, a cutting edge research question. So a lot of the output was what they call weakly justified or artificially hallucinated. Right? So in order to control for those weakly justified or artificially hallucinated outputs, that's why they, they came up with this workflow structure and the specific use of the LLMs as

• 19:55Speaker 01

  the Socratic questioner, so that it wouldn't do that. And it seems that that worked. So then my question is, how well is this a template for other areas of education of incorporating AI into education, rather than seeing it just as a problem for educators? Because now students are going to cheat, right? If you say, well, let's just design the whole process around using AI, therefore, using AI is baked in, you can't really cheat, right? What do you guys think about that?

• 20:26Cara Santa Maria?

  Well, I mean, A, that's where we're going.

• 20:29Speaker 01

  I, I think so, yeah.

• 20:30Cara Santa Maria?

  I mean, it's like, you know, I'm a healthcare provider and like more and more AI is just part of the process in the electronic medical record. Like, it's just everything we use. Like, it's baked in so much that we don't even know sometimes.

• 20:41Speaker 01

  Yeah.

• 20:42Cara Santa Maria?

  When an LLM is part of the search function.

• 20:45Speaker 01

  Yeah, I agree.

• 20:46Cara Santa Maria?

  Yeah.

• 20:46Speaker 01

  But there's another layer here too, and that is, I just think the whole AI phenomenon, when we get to this level of thinking about like, how is it used in art, or in education, or in writing, or whatever, kind of forces us to think about thinking, right? To think about the process that we are using creatively or scientifically, because I think AI is getting better and better at mimicking that process without really using that process. And so it kind of forces us to think, well, how am I different than the AI? What am I doing that the AI is not doing? And how can I do better than the AI? Right?

• 21:31Cara Santa Maria?

  That would be a cool, like, like, exercise.

• 21:34Speaker 01

  Totally.

• 21:34Cara Santa Maria?

  Yeah. In, in a class where you're learning research methodology is to say, you know, do this same search using AI prompts over here. Now do it, you know, with this functionality turned off and see if you can find things that the AI couldn't find.

• 21:49Speaker 01

  Yeah. And there are lots of, of lessons like that where you're basically, like, the AI slop, I think, is the baseline.

• 21:56Cara Santa Maria?

  Mm hmm.

• 21:56Speaker 01

  You, you're not doing anything unless you are, you are producing output that is greater than what the AI could do by itself. Because otherwise you are adding no value to the process. But in this case, the, it's kind of built in, like, you're, you're giving the students a structured process to follow, not having them wing it, or just say, you've just, you've got to do better than AI, or don't use AI, or whatever. You're saying, all right, yeah, if you're going to use AI, this is what you're going to do. This is, and the, and the end results are going to be a collaboration, and you're going to be judged on your part of that collaboration, right?

• 22:28Cara Santa Maria?

  Right.

• 22:28Speaker 01

  Not, not whatever the AI produces. So I think something like that is where we're headed. And again, if, best case scenario, the end result could be really good. I mean, we could, you know, in the same way that photography had a huge impact on, on art, on painting, because just doing realistic paintings was not enough anymore. You had to do something else. I think that AI could have this, could have the same positive impact at the, the good end of the spectrum, you know, where it's sort of forcing people to say, oh, okay, you know, a, a, the, the process that I'm adding, the creativity, or the insight, or whatever, has got to be clearly better than what AI can do. And how do we do that?

• 23:13Speaker 01

  You know, how do we accomplish that? So anyway, I thought that was, that was so interesting on so many levels, you know?

• 23:19Cara Santa Maria?

  Absolutely.

• 23:20Speaker 01

  And, and it is something we absolutely have to think about very, very carefully.

• 23:24Cara Santa Maria?

  And it's inevitable.

• 23:26Speaker 01

  Yeah, it's inevitable. It's just, what's not inevitable is if it will be a good thing or a bad thing. And, or the degree to which it will, it'll degrade education and turn out a generation of people who are intellectually crippled by dependence on AI versus a generation of people who are able to, to achieve much more, because they've learned how to optimally collaborate with AI. And it makes them think better, more clearly, again, operationalizing things. Their education is actually better. The pot, there's definitely a huge potential upside. And I think figuring out how to get that upside also simultaneously mitigates

• 24:06Speaker 01

  the downside. You know what I mean? So I think there's a huge motivation to do that. So anyway, I really like the approach that this Socratic questioner takes.

• 24:16Cara Santa Maria?

  Yeah.

• 24:17Speaker 01

  All right, Jay, kind of related to topic, not really, but tell us about robots who can now outrun humans.

• 24:24Jay Novella?

  Well, first off, what do you guys think about that?

• 24:26Bob Novella?

  Well, Jay, I've seen the videos, and of course, my first reaction is, is this even real? You know, and I didn't dig too much deeper. So you're, what you're saying then is that those videos of the robots running into walls and sparking and, and stuff, those are real? That, that, those are real running robots?

• 24:42Jay Novella?

  Well, I didn't see exactly what you saw, Bob, but I can confidently say that the, uh, the, the speed run that happened was real and that, you know, this was done at an event in China where they, they show robots doing tons of different tasks.

• 24:57Bob Novella?

  Okay.

• 24:58Jay Novella?

  Yeah. So it, it, it happened. But you know, Bob, you can't trust anything you see now.

• 25:02Speaker 01

  I know, right?

• 25:03Jay Novella?

  You have to really verify it.

• 25:04Speaker 01

  My guess would be that, sure, it could happen, but probably not every time. You know, I imagine the robot probably trips a lot, and it's probably only for a short distance because of energy.

• 25:15Jay Novella?

  Okay. Well, the, uh, there's a, there's a question here that is like, kind of thematically running through what I'm about to discuss with you guys. And that's like, you know, sure, a robot can now run faster than the fastest person on the planet, but can it plug in a, uh, extension cord?

• 25:31Speaker 01

  Right.

• 25:32Jay Novella?

  Right? And that, that shows you like, you know, the brittle nature that's, that's uh, interwoven here with this whole thing, right? So Bob, the, uh, the place where this happened was the, this was the Second World Humanoid Robot Games, which happened in Beijing. I wish I could have gone to see that. There was over 2,000 robots that competed in like, you know, running, kickboxing, factory work, restaurant services. It was all over the place. So this, this robot you're talking about, Bob, um, it is called the Tiangong, Tiangong Ultra. It ran the 100 meter dash in 8.64 seconds. And if you see the video, it looks so goofy, the way that this robot moves, but it friggin motors, man. It, it,

• 26:15Jay Novella?

  it beat Usain Bolt's world record of 9.58 seconds.

• 26:19Bob Novella?

  Oh, really? They already beat the world record?

• 26:21Jay Novella?

  Yeah, yeah.

• 26:22Bob Novella?

  There we go.

• 26:23Jay Novella?

  They're on top.

• 26:24Bob Novella?

  It's happening.

• 26:25Jay Novella?

  And, you know, this is, uh, one of the things I think is very important to take note on is that how quickly this technology is improving. So a year ago, the same robot ran the same distance in 21.5 seconds, right? So it went from 21.5 seconds to 8.64 in one year. Yeah, I'm not saying that they didn't make changes to the robot. Of course they did. But, you know, I think a huge change that they made was to the software that's running it. You know, there's a lot of people that don't like to hear things like that, but it's, it's happening. You know, it is, it is absolutely something that we have to understand. And, and on some level, we have to embrace it because it's coming and there's nothing that can really stop it. And this isn't even the most interesting part of the

• 27:09Jay Novella?

  whole thing, because the games were also testing robots, like I said, in ordinary physical tasks, you know? And in order for robots to enter the human world, they have to do things that we can do, right? For when they truly become useful, right? So we want mundane tasks like being able to open up the port on an EV and plug it in. All different types of things in a restaurant, say, or, you know, emergency response, industrial work, industrial assembly, you know, whatever, you know, like, just think about it. Anything, anything with hand dexterity would be very, very helpful in lots of cases. And a lot of cases, like, this is work people, some people don't want to do. Um, and it can be,

• 27:50Jay Novella?

  it could be difficult even for people to do. But let's, let's reduce this down to something that most of us can do without even thinking about it, right? We want to plug in an extension cord into the wall, say, right? So think about all the things that we do without really consciously thinking about it. Your brain is processing it, but we, it doesn't even hit your consciousness because it's so easy for our brain to do this for most people, right? You see the cord, you recognize it, and you understand its function. Then you know that the connector on the cord is the correct one for where you want to plug it in. Then you grab the connector. Then you figure out the correct orientation it needs to be in. So you rotate,

• 28:32Jay Novella?

  you know, you move your arm and your hand to move it around till you get to the point where you think it's in the correct position. Then you visually see the socket, or if, God forbid, you're doing this in the dark and you're using your finger to find where the holes are, anybody?

• 28:44Speaker 01

  Yeah, I've done that. I hate that.

• 28:45Jay Novella?

  Yeah, happens all the time. Then there's that guy with the black glove that's slapping your hand away. So you find the socket, you line it up, and then you insert it, and you're inserting it with the right amount of force, and you also know when to stop. Now, here's the problem, right? Because that all is complicated. You know, I wouldn't say for humans, but it is sometimes it's mostly impossible for robots to be able to do consistently. But it's also common that you plug it in and it doesn't work. It doesn't actually go in, right? Because sometimes the connectors can be bigger. In the US, some manufacturers have connectors that can only go into a very specific orientation. It can't be like turning it around backwards. You know, it needs to be the right way.

• 29:27Jay Novella?

  And then, you know, everyone knows the USB A problem, right? You plug it three times. Every time you try to plug one in, you try it, doesn't work, flip it, try it, doesn't work, flip it around, it works. I don't know what the hell that is, and whoever came up with that design should be slapped. But that, we've all gone through that. So what does a robot do? Well, then it has to troubleshoot, and man, the different pathways that it could take there could be profound, just to troubleshoot whatever the problem is. You know, you're just assuming this plug is going to fit in there, and it just has to get the orientation correct for it to go. But that is not the human world. And if it's slightly off, you know, we can feel it in our hand, we can, we feel the resistance,

• 30:09Jay Novella?

  and we can make a miniscule adjustment to jiggle it in there to get it to do that. It's another thing that the robot would have to do. So, like I said, all this, we do it unconsciously. We do all these things unconsciously, like walking, incredibly hard thing for robots to do. They can do it now, but it was, took a very long time for them to get it. We don't have to think about it, because our brain can process it effortlessly. So any step along the way of what I just described stops robots in their tracks. And it is one of the biggest challenges that roboticists have right now. And what do they do? Well, they're using AI, right? We talked to Christian Hubicki about this, and he's saying that AI is, not his words, but it's basically an accelerant to train robots.

• 30:53Jay Novella?

  You know, and the better AI gets, the better the software that can drive these robots will get, and it will speed things up. And it's become, AI is very good at doing lots of things, right? It can manipulate information, you write computer code, it can recognize images, it can create images, it can do mathematical problems, it, you know, it can beat humans at really complicated games. But when you strap an AI on top of a physical robot body and put it in the real world, then, you know, it's not just an order of magnitude more difficult, it's much, much greater difficulty for us to be able to make these physical things happen. And that's why, the reason is, is because the world is messy. The human world is completely messy.

• 31:35Jay Novella?

  It isn't organized, it isn't predictable. There's all of these challenges that roboticists are trying to overcome, and the problem is, is you can't program it to handle every single case and edge case, yet it has to have a general intelligence in order to handle situations like the human brain can do. Because we, you know, the troubleshooting that our brain can do, not only is it very quick, but it's largely successful.

• 32:02Speaker 01

  I mean, I'd push back on your, I don't know if you intended to say general intelligence. I don't think you need an artificial general intelligence for that.

• 32:08Jay Novella?

  I don't mean that, I'm sorry, Steve. It needs a programming layer that can handle common occurrences that would come up, but it can't be programmed specifically.

• 32:19Speaker 01

  It needs to be able to troubleshoot in real time, based upon its environment.

• 32:22Jay Novella?

  Right. It has to read the environment and be able to troubleshoot, and have enough processing to come up with solutions to things, and not have it all be pre-programmed. Because that, you know, ultimately, robots are going to have to handle things that aren't predictable, right? That's the messiness of our world. So then, there's a Chinese robotics company that is giving an estimate that's saying existing real-world robotic training data sets, you know, they contain, say, 100,000 hours of data. And what they're trying to do now is they're trying to collect tens of millions of hours, right? This kind of goes back, this goes back to Tesla, Steve, right? With all their driving hours that they're collecting from all the people out there,

• 33:04Jay Novella?

  and Tesla does have the best autonomous driving right now, and it's for very specific reasons. One, they got in first, but two, they are collecting massive amounts of driving data that they fold back in to their system all the time. So today's robots, you know, they're not learning how to perform all these different tasks. You know, they're generating trade, you know, this training data that can, you know, that will get them to the point where they can handle unpredictable things, right? And I think that this is kind of like a holy grail in the programming that they need to be able to do to put robots into our environment. And so these demonstrations are cool, right, Bob? Like, you know, Steve, you said it. Could it do it repeatedly? Can that robot run that race and win every time?

• 33:48Jay Novella?

  Like, I don't honestly know the answer to that, but it's a good example of the fact that these are demonstrations. You know, these machines that were at this competition, and during the games, you know, the robots are still falling over. One of them, like, came apart in pieces, and listen to this, another one caught on fire after crossing the finish line. It literally finished in flames.

• 34:09Bob Novella?

  Did he win, though? Did he win?

• 34:12Jay Novella?

  I don't think he did, Bob, but at least he finished. Have you guys ever heard of Moravec's paradox?

• 34:19Bob Novella?

  Moravec, V E C? Yeah, Hans Moravec, loved the dude.

• 34:22Jay Novella?

  So what is it? It's things that we think of as intellectually difficult can sometimes be really easy for computers to do. And then there are abilities that humans have that we don't even have to be consciously thinking about, perception and balance, manipulating objects, and this can be extraordinarily difficult for machines to do. And that is in full effect here. We're seeing that with all, all roboticists know that that is what they're fighting against. So the robot games were cool, there were a bunch of demonstrations and everything, but, you know, we're not there yet. But I think it's important to acknowledge, guys, that it is really just a matter of time. Like, it's going, it's going to happen, multiple companies are going to get there.

• 35:06Jay Novella?

  Working in the home and working in, you know, small businesses, those style robots, of course, in the beginning will be expensive, or, you know, whatever, whatever the circumstances are, but they're going to get there. You know, there's companies, like, we've talked about this not too long ago, they're selling robots for the home. You know, one of them is like, you know, has like a, remember that one? It's really soft, it's like, you know, it's basically like a plushy robot?

• 35:28Bob Novella?

  Yeah.

• 35:29Jay Novella?

  Just, you know, and it's weak, it can't lift heavy things and all that, just to make it safe around humans. So it's really not the, the physical part. The robots can run, they can move, they can do backflips and all that. I think the biggest problem they have is the programming, and that's just a matter of time.

• 35:44Bob Novella?

  And if you like what Hans Moravec or Moravec says, check out his book that I read about, God, 25 years ago, Mind Children. I haven't read it in decades, but I think it probably holds up. Interesting, very interesting book, Mind Children by Hans Moravec or Vet.

• 36:03Jay Novella?

  And before we move on, I'm curious now, I'm very curious to hear what Kara has to say. So Kara, in the home robot, it's, you can get one, you know, I'm giving you a situation, I want you to tell me what you think. You can buy one, you could pay it off over 20 years, and it's not crazy expensive, you could, you could easily afford it, and it can do physical tasks around your house. Would you get one?

• 36:27Cara Santa Maria?

  I mean, it just depends on how, it's, the risk benefit analysis for me is always, does it reduce my labor, or does it end up being more hassle than it's worth?

• 36:37Jay Novella?

  Of course.

• 36:38Cara Santa Maria?

  And that's a really hard question, you know? Sometimes the most technologically advanced, the most amazing things we have, we've talked about this on the show 100 times, the more complex it is, the more frustrating it is when it breaks.

• 36:49Jay Novella?

  I mean, look at your cell phone, right? I have an iPhone. My iPhone requires, you know, a good chunk of my time to figure things out. Like, yeah, there's, there's things I do every day I don't have to think about, but as soon as you go off the beaten path, I have no idea what to do. And I'm looking online, I'm trying to figure things out. I don't want to be troubleshooting a robot in my home.

• 37:08Cara Santa Maria?

  No. And like, something as simple, Jay, is like, the windshield on my truck has a crack in it, and I'm just living with it, because if I replace my windshield, there are computer components that have to be replaced. My windshield has a freaking computer. Why?

• 37:25Speaker 01

  I know, I had that same experience, where I had a little tiny crack in my windshield, and like, that'll cost $1,000 to replace, because the windshield is integrated with all of this stuff, most of which I didn't even want.

• 37:36Cara Santa Maria?

  Exactly.

• 37:37Speaker 01

  I don't need my windshield to detect raindrops, and, you know what I mean? And trigger the windshield wiper. I could just set it manually to some intermittent setting, and it's fine.

• 37:47Cara Santa Maria?

  Yes.

• 37:48Speaker 01

  It's like 99%. You have this tiny incremental bent, you know, bells and whistles, that then massively complicate the ability to maintain and repair your car. I hate that.

• 37:57Bob Novella?

  It's infuriating. I wonder if you can get a windshield, can you get a windshield that doesn't have the computer in it, and just say, just give me?

• 38:02Jay Novella?

  Turn it off, yeah.

• 38:03Cara Santa Maria?

  I think you have to just like disconnect it. But see, this is the same thing that my parents said about automatic windows. They were like, ooh, the crank, you know, those manual windows, they just don't break. You can always like reset it, but the automatic ones, once they break, what do we do? The window's stuck down. So we may sound a bit like Luddites right now, but I also, you know, this is our generational kind of line in the sand.

• 38:30Speaker 01

  But it's all relative. I do, I agree, you get accommodated to the luxury, and then that becomes your normal, and then they have to push it to something else, which you think, which then you think is unnecessary. But the next generation might find like, oh my God, you had to manually set your wiper speed? I mean, what the hell? Things just happen automatically in the background all the time, you know what I mean? So, yeah, I know, I know, it's all, it's all relative.

• 38:53Bob Novella?

  Well, how many times has your automatic window been a problem? It has happened to me, but like, once, once in my life.

• 39:01Speaker 01

  Literally, never.

• 39:02Bob Novella?

  Where it's like, oh my God, I can't, I literally can't close my window.

• 39:04Speaker 01

  Not that I can remember ever. Certainly not recently.

• 39:07Bob Novella?

  How many times have you broken your windshield, though? That's, you know, that's a, that's not very often either, for me, anyway.

• 39:13Speaker 01

  It's not infrequent.

• 39:14Bob Novella?

  A couple of times, but $1,000, that's a deal breaker, man. Screw that.

• 39:19Jay Novella?

  But would you have a robot in your house, the same circumstances, if, if, you know, it was very reliable, like it could actually do your laundry and you don't have to worry about it, it could, you know, does all these things?

• 39:29Bob Novella?

  I would absolutely consider that.

• 39:30Speaker 01

  It's all those variables, Jay. It's like, how much does it cost? How reliable is it? How easy and expensive is it to maintain? What is the net benefit to my, running my household? What are the downsides? Like, you know, does it drive the pet crazy or whatever? And you have to factor all those things in. It's probably the kind of thing you can't really know until you do it.

• 39:50Bob Novella?

  Yeah, right. And but what I want to see though, now, now, this, this robot ran faster than Usain Bolt. That to me is one of the, it's one of those really interesting milestones because I want to see robots now that are far more dexterous and agile than people. I want to see a robot like gymnastics, you know, like doing a floor routine. I want to see that. That would be fascinating to see something more agile than humans could ever be without some weird augmentations. Uh, that would be, I'm looking forward to something like that.

• 40:21Jay Novella?

  But look, I'm gonna just say it very plainly. I want a robot.

• 40:25Bob Novella?

  Yeah.

• 40:26Jay Novella?

  I want it, I want to name it. I want it to have a voice that I pick. I want to, you know, I want it to do stuff around my house and, uh, and when I look sad, I want it to like, ask me if I'm okay.

• 40:37Speaker 01

  So Jay, I was, a couple of things. One is, uh, I suspect you like the idea of a robot and that not, wouldn't necessarily translate to the reality of owning a tr- a robot. But my second thing is, I probably would want first to have a robotic pet. Cause I think that's a much lower bar to be able to function as a pet being robotic than as like a humanoid servant. Um, and, and there's a ton of functionality that you could build into a robotic dog-like creature, a ton. That would be extremely useful. I think that would be a much better way to go. Plus, you know, first, because it'd be four-legged, it wouldn't have to do things like cook or fold your laundry, but the things that it could do.

• 41:22Bob Novella?

  Security?

• 41:23Speaker 01

  Could, yeah, security.

• 41:24Bob Novella?

  Rap killing?

• 41:25Speaker 01

  Monitoring, you know, smoke detection, all kinds of stuff, communications, uh, yeah, and companionship. Would, there'd be tons of stuff, you know, lots of safety benefits.

• 41:35Bob Novella?

  Yeah. We'll probably see that first, but I, if I, there was a robot that could do my laundry and the dishes, O M F G, I would so be there.

• 41:45Cara Santa Maria?

  That's true.

• 41:46Bob Novella?

  If it was reasonable. That would be a game, that would be like, oh my God, I mean, I got a pile of dirty clothes upstairs. Like, I'm just like procrastinating like crazy, like, oh.

• 41:54Cara Santa Maria?

  Think about like-

• 41:55Bob Novella?

  No underwear, I guess I gotta do my laundry now.

• 41:57Cara Santa Maria?

  Yeah. Think about like, we have robot vacuums, right? But they're obviously very limited. They can't go down the stairs. They can't actually, they're like using sensors and they're only in like one specific place. But think about if you had a robot that could vacuum your floors and dust everything once a day.

• 42:11Bob Novella?

  At night, quietly, don't even hear it.

• 42:15Cara Santa Maria?

  Yeah. Your house would always be clean. It would be amazing.

• 42:18Bob Novella?

  Always.

• 42:19Cara Santa Maria?

  And like scrub the, like vacuum the carpets, clean the hard floors and dust everything.

• 42:24Jay Novella?

  And Kara, it could yell at you when you don't get out of bed fast enough.

• 42:27Cara Santa Maria?

  Oh, and clean all the windows.

• 42:29Jay Novella?

  Oh, yeah.

• 42:30Bob Novella?

  How about this? Yard work, yard work. Don't even get me started on that. That would just be incredible. And then when I'm 80, I'm gonna need a helper decorating for Halloween that, you know.

• 42:42Jay Novella?

  I thought you were gonna say wiping your butt, Bob. Like, your priority is like Halloween.

• 42:46Bob Novella?

  I totally got that. I mean, I've got my minions in, in preparation, like Steve and Jay's daughter, you know, kids are like, I've been prepping them to be my Halloween minions for when I, I'm in a wheelchair. So, but having a robot also would be cool.

• 42:59Cara Santa Maria?

  But you know what? That actually, that, the point that you just made, Jay, like, is, is a really important one. I have so many patients who would benefit from having a home healthcare robot. They'd still have nursing care, but they could have nursing care a few hours a day instead of 24/7, because most of what they need is, can you help me up? Can you help me sit back down? Can you bring me some water?

• 43:21Bob Novella?

  Assistant living would take a hit. I mean, my mom could be alone if she had a, a capable robot that could make sure she takes her medicine. If she falls or anything like that, immediately notify somebody, immediately.

• 43:32Speaker 01

  Well, it certainly would extend the, the amount of time that people could live safely and functionally at home without going into a more higher level of assistance. Absolutely.

• 43:42Cara Santa Maria?

  Yeah, and reduce the amount of time that the skilled nursing staff needs to be present in the home.

• 43:47Speaker 01

  Right. Which we are underserved right now.

• 43:49Cara Santa Maria?

  Absolutely. Yeah, it's not like we'd be taking jobs. We, we need more gaps filled in this area.

• 43:56Speaker 01

  Absolutely. All right, Kara, tell us about this latest CTE study.

• 44:00Cara Santa Maria?

  Yeah, so this is, um, it's an interesting one. And Steve, I know that you wrote about this recently. Obviously you have a lot of expertise in this area, so there will be plenty of times where I may defer to you, or you can obviously-

• 44:12Bob Novella?

  Chime in.

• 44:13Cara Santa Maria?

  Chime in and, and let us know what's going on. But my take in reading this new study, which was published in, uh, the British Medical Journal, uh, just, you know, a few days ago, is that the numbers feel very scary. It feels very like, oh my gosh. But also I am not surprised by this at all. It also doesn't feel like it's, it's new in that we, we, we looked at more brains and we used them to, some statistics on more brains. But when you look at the literature over the course of CTE being studied in the NFL, the National Football League, it's just bad.

• 44:51Speaker 01

  Yeah. It wasn't surprising. As I said in my paper, it's perhaps alarming, but it's not surprising.

• 44:57Cara Santa Maria?

  Right. And I mean, I found a study that was published in 2017 that had numbers that were just as scary. Like in this one, it said, there was a convenient sample, this is the old study now, I'm just gonna say this so we have it, you know, in our minds. There's a convenient sample of 202 players who had died, who played football throughout their, uh, lives. Some of them were NFL players, so about half of them were NFL players, and the rest played at different levels, like amateur ball or even high school ball. And they found on this one, that in this convenient sample, which we'll get into, 99% of the donated NFL brains had CTE and 87% of the non-NFL brains had CTE.

• 45:41Cara Santa Maria?

  Now, let's talk about what a convenient sample is, right? A lot of times, a football player will show signs and symptoms of some form of cognitive decline, or some form of dementia, or some form of neuropsychiatric issue, like personality change, or difficulty with impulse control. And there will be, at this point in our time in 2026, or even going back to when this study started in 2008, a concern about chronic traumatic encephalopathy. So, at the end of their life, they will decide, or their family will decide, to donate their brain to research.

• 46:22Cara Santa Maria?

  Already, there is an idea that there is CTE in this brain, or that there's some sort of concern around it, often when it is donated. So we can't really read too into, this percentage of these brains had CTE. It's like, yeah, that's why they donated them to you.

• 46:40Speaker 01

  Yeah.

• 46:41Cara Santa Maria?

  To figure out what was going on with these brains.

• 46:42Speaker 01

  To buy a sample.

• 46:43Cara Santa Maria?

  Yeah, we have to remember that. Um, so in this study, uh, it, the new one, it's a retrospective population-based cohort study, spanning from 2008 until 2021. So in this study, they took 1,000, uh, or they looked at 1,712 former NFL players who died between 2008 and 2021. 338 of them donated their brains for full eval, meaning that they were able to, you know, diagnose them, take microscope samples and do a lot of testing on them. They also looked at a bunch of information around cause of death and, um, some detailed

• 47:28Cara Santa Maria?

  information in some situations, sometimes we were lacking detailed information. So in those 338 brains, they were able to, um, ensure that the clinicians didn't know what level of CTE there was a concern about, you know, how that player had played. There was some amount of masking, but again, this is, you know, it's, it's a convenient sample. And what they looked at was basically the number of donors with CTE over the number of NFL deaths, which is kind of a weird way to measure this, because they didn't look at most of their brains. Does that make sense?

• 48:08Bob Novella?

  Hmm.

• 48:09Cara Santa Maria?

  Like, that was one of the biggest, I, I don't want to say it was like one of the biggest limitations per se. I mean, yes, I guess we can all say it's an obvious limitation, but it's not like the authors of the study aren't aware of that. Like, that was a big part of the conversation. So, one way that like, Steve, you talked about it, is that if 338 players donate their brains for examination and make them available, those 338, like I said, are very likely already concerned about CTE. And if we say, of the 338, how many had CTE? That's one way to measure.

• 48:49Cara Santa Maria?

  If we say, of the 1,712, 338 were studied and 315 of those had CTE, well, what's now the range of possible CTE prevalence? Well, on the low end, we assume that all the brains that weren't studied didn't have CTE. And on the high end, we assume that all the brains that weren't studied did have CTE. So now you have a range of 18.5% to 98.7%. That's a big range. That doesn't really tell us anything, does it?

• 49:21Speaker 01

  Yeah, it just sets a floor, really.

• 49:23Cara Santa Maria?

  Exactly. And so, what we do know is that of the 338 brains that were donated, 315 of them were able to be diagnosed with CTE. That's huge. That's 93.2% of that sample. But again, there was already a concern that they had CTE. And the reason I'm saying all of this, like, did they or did they not get diagnosed? Which I should have caveated at the beginning, is that you cannot diagnose CTE until death. It's an autopsy-based diagnosis.

• 49:52Speaker 01

  Yeah, it's a pathological diagnosis.

• 49:54Cara Santa Maria?

  Yeah, we are increasingly pointing, there's a lot of really interesting literature on the horizon, pointing to different, like, markers. And so hopefully, we're going to get to a point where individuals can be tested for, you know, certain things that are released into the, let's say, cerebrospinal fluid, or even into the blood, due to the type of injury that leads to CTE. Or they've already got this pathology, maybe there's a way. But this is a similar concern that we have to Alzheimer's, right, Steve? Like, we do diagnose Alzheimer's in life, but really, the diagnosis is confirmed on autopsy.

• 50:39Speaker 01

  That used to be true.

• 50:40Cara Santa Maria?

  Oh, okay, so that has since changed?

• 50:42Speaker 01

  It has changed, like, in the last, really, just five to 10 years, where we're getting a better criteria. And it depends on the, it depends on the Alzheimer's subtype as well. But now with spinal fluid, blood, and MRI scan, we could actually make a pathological diagnosis of Alzheimer's short of an autopsy. Although autopsy is still the best way to do it.

• 51:02Cara Santa Maria?

  Yeah, beautiful. So that kind of goes right in line with what I was saying. They're working towards figuring this out with CTE.

• 51:07Speaker 01

  Yeah.

• 51:08Cara Santa Maria?

  We've got to remember, though, that, just in sheer numbers, Alzheimer's affects so many more people than CTE. CTE does have a background level in the population, but it's exceedingly low compared to the level that we're increasingly seeing among contact sport athletes. Anybody who has repeated head trauma over a long period of time, and it could be subconcussive, we have to remember that. You don't have to have all the telltale signs of a concussion over and over to have a risk of CTE. These can be blows to the head that occur below a measurable baseline, and we still see the pathology that's involved in CTE.

• 51:53Cara Santa Maria?

  I think it's important to point that out. I think it's also important to point out what some of the major risk factors are here. One of the really big ones is length of time experiencing those types of hits. So like, yes, if you have one really, really, really, really, really, really bad concussion, and we have seen this especially in exposure to blasts from explosives in military veterans, you could be diagnosed with CTE, but it's actually much more rare to get CTE from a single injury

• 52:32Cara Santa Maria?

  than it is to get CTE from a consistent, like, repetitive concussive or subconcussive blows. I did look up a couple-

• 52:40Speaker 01

  I don't know if that's true, though, that you could ever get it from a single injury.

• 52:44Cara Santa Maria?

  No, you can. It's been, I mean, studies have shown that exposure to blast from explosives can produce the symptoms. But again, you're right, like, the symptoms.

• 52:54Speaker 01

  I thought the pathology was specific to, like, the repetitive injury. It's chronic.

• 52:58Cara Santa Maria?

  That's probably, yeah, that's probably a really, really important-

• 53:01Speaker 01

  Otherwise, it's just TBI.

• 53:02Cara Santa Maria?

  Distinction, exactly.

• 53:03Speaker 01

  We just call it traumatic brain injury.

• 53:04Cara Santa Maria?

  Which, and the interesting thing is, a lot of the symptoms of CTE mimic the symptoms of TBI, but they also have some unique kind of components to them as well.

• 53:16Speaker 01

  A lot of overlap, yeah.

• 53:17Cara Santa Maria?

  And the main reason that CTE is unique is because, well, there's a few main reasons. One is that it's sort of diffuse trauma to the brain, right? This can be from direct blows, it can be from head-in-motion injuries. So you can see axonal shearing across the whole brain. You can see lots of different kind of point sources of injury over and over and over. Obviously, you're going to have all of the downstream effects that happen then, like swelling, potential bleeding. But it's also within the subclass, right? It's a neurodegenerative disease that's within the subclass of tau pathies.

• 53:56Cara Santa Maria?

  So they do aggregate abnormal tau protein, similar to Alzheimer's, and you do find these neurofibrillary tangles in CTE brains. So it's partially the physical trauma and, like, blood loss and swelling and all the things that come with that. But a big part of it is all the pathology that comes downstream when these cells release different contents and when different kind of post-injury molecular processes occur. So I think it's important to remember that as well. They're diagnosing this based on what the brain looks like as a whole, but also based on what these microscope slides look like.

• 54:33Speaker 01

  Yeah.

• 54:33Cara Santa Maria?

  But you can see it when you look at the whole brain in a slice, like, with the naked eye.

• 54:39Speaker 01

  Hmm.

• 54:39Cara Santa Maria?

  You may have heard other names for CTE. Chronic traumatic encephalopathy, CTE, is sort of the main, you know, the accepted term that a lot of people use now. But you may also hear it called, it was first diagnosed, when they first developed it, as dementia pugilistica. That was its first name. And then, of course, punch drunk syndrome, right? So, yeah, and there do seem to be specific subtypes, there seem to be specific levels. And it does appear to be the case that boxing or martial arts, but especially where you can strike in the face, does have a tendency to produce more head-in-motion injuries,

• 55:22Cara Santa Maria?

  like, the twisting that causes axon shearing. So it's interesting to see that, like, Parkinsonism-type syndromes, or type symptoms, can be more common, for example, in that group. So they are starting to be able to say, okay, which groups of people, based on their sports, are developing which kind of constellation of symptoms? But none of this categorization is perfect, because we're still learning about this. There are also different stages of the disease. And as I mentioned before, dementia is and can be a big part of it. So that includes a lot of different kind of cognitive categories,

• 56:05Cara Santa Maria?

  like frontal lobe disinhibition, difficulties with executive control, obviously, memory problems, confabulation, perseveration. But there are big neuropsychiatric components as well. And one of the very, you know, I don't want to say very small, I'm just going to say one of the subtypes, because this does not happen in all people with CTE, even though this has kind of become the popular, in the popular discourse. Some people with CTE will experience things like apathy, loss of empathy, aggression, anger, and it may even be associated with violent behavior. That doesn't always happen. It is not definitely going to happen.

• 56:50Cara Santa Maria?

  And the risk of that actually increases with a history of violent behavior. So, you know, yes, it can be a major personality change that happens in some of these athletes, but some people already have, you know, that history, and then it becomes exacerbated in CTE. So, anyway, that was a lot about CTE and less about this study. The big takeaway on the study is that range number, which is a huge range. So I don't really think we're learning too much new. And I do think that that number that most people are citing, which is at least 1 in 4 players, is based on accumulative data. So it's, yes, at least 215 of the total former players that were looked at

• 57:35Cara Santa Maria?

  in this, in this statistical analysis, had CTE. So at least 24.5% had CTE. It's probably much, much higher than that.

• 57:44Speaker 01

  Yeah.

• 57:44Cara Santa Maria?

  It's probably higher than 1 in 4. And that just goes in line with all the other literature out there.

• 57:49Speaker 01

  Hmm. So, did you mention that, you know, you can get CTE without ever having had a concussion, right?

• 57:56Cara Santa Maria?

  Yeah, yeah.

• 57:56Speaker 01

  It can happen with cumulative hits that are subconcussive.

• 58:01Cara Santa Maria?

  Yep, very important. So it's not just about that, that, you know, kid is losing balance or loses consciousness, or is having kind of these concussion symptoms, and then you pull them off and you allow them. I think that was, or has been, the guidance for a long time. But if anybody hits their head, or even just takes a big hit, and you don't know if their head was involved, it's probably better safe than sorry. You need to pull them out of the game.

• 58:27Speaker 01

  That's still the recommendation, but that's more to prevent TBI. There's no evidence really that it prevents CTE. And also, like, the helmets do reduce the risk of concussion. It's not clear if they reduce the risk of CTE.

• 58:41Cara Santa Maria?

  Yeah.

• 58:41Speaker 01

  It kind of makes sense that they would, if it's reducing concussive risk, but because you can get it with subconcussive hits, it certainly wouldn't eliminate the risk.

• 58:51Cara Santa Maria?

  No, soccer players who had the ball get CTE.

• 58:54Speaker 01

  Yeah.

• 58:55Cara Santa Maria?

  Like, there are plenty of documented examples of soccer players who had the ball multiple times over their career, ending up with a diagnosis of CTE on pathology, who never experienced any contact injury to another, you know, player. So it's this repeated, and the risk factors are, do seem to be like, kind of, I don't know if they're perfectly dose-responsive, but it does seem to be the case that the longer you are playing the sport, so your risk increases the younger you are when you start, and the longer you go until you retire.

• 59:30Speaker 01

  Yeah.

• 59:30Cara Santa Maria?

  And the reason that we see a higher risk in professional players versus amateur players, or even people who play for fun, is, okay, sure, they may be playing at a higher level of, like, aggression or a higher level of contact, but really, it's that they just play more.

• 59:44Speaker 01

  Yeah.

• 59:45Cara Santa Maria?

  They're practicing more, they're playing more games. So they're getting more playtime, which translates to more concussive and subconcussive injuries.

• 59:53Speaker 01

  All right. Thanks, Cara. Well, everyone, we're going to take a quick break from our show to talk about our sponsor this week, Eleven Labs.

• 59:59Jay Novella?

  Steve, you ever call a call center and you get someone that can't help you?

• 1:00:04Speaker 01

  All the time.

• 1:00:05Jay Novella?

  And you're frustrated and you're like, I wish I could just have someone else that knows what they're talking about. Check out Eleven Labs. Kara, what do you think?

• 1:00:12Cara Santa Maria?

  Yeah, I mean, they're amazing. They make something called Eleven Agents. This is not a chatbot. It's not a phone tree. It's not a demo. This is like, working. And it's a voice AI agent, but it listens, it understands, and it can escalate to give you all of the information that you need. It can access your account, process that request, and here's the best part, it sounds like a human being. It runs 24/7, and it's in over 30 languages with very low latency performance.

• 1:00:41Bob Novella?

  So if you run a business with any kind of customer operation, support, sales, or onboarding, you can start with a demo at elevenlabs.io/sgu. See what Eleven Agents can do for your specific workflows. Guided rollout capacity is limited each quarter, so if you want to be one of the first companies with a customer experience that people actually talk about in a good way, now is the time. Elevenlabs.io/sgu.

• 1:01:09Speaker 01

  All right, guys, let's get back to the show. Bob, tell us about how bacteria share survival genes.

• 1:01:15Bob Novella?

  If you insist, Steve.

• 1:01:17Speaker 01

  I do.

• 1:01:18Bob Novella?

  Guys, researchers have drilled down how microbes can thrive in a place as inhospitable as Antarctica, and it has to do with horizontal gene transfer and something wonderfully weird called aerotrophy.

• 1:01:30Jay Novella?

  I don't like it.

• 1:01:32Bob Novella?

  You will. This is from Nature Communications. The study's name is "Widespread Horizontal Transfer and Strong Selection Enhance Microbial Adaption in Antarctic Soils." Fairly straightforward title there. Okay, so let's talk about Antarctica, guys. You're familiar with this. Not a fun, happy place, I mean, in terms of just like, being hospitable. The coldest, windiest, and driest continent on the planet. It's actually a desert, specifically a polar desert, and that's because it gets less rain than 10 inches a year, which is the kind of the cutoff, the arbitrary cutoff for what's a desert.

• 1:02:13Bob Novella?

  In fact, it gets 1/5 of that, gets only 2 inches of rain down there. The available liquid water is basically anathema down there. It just is hard to find and doesn't last long if it is. And that's because the air is just so cold, it cannot hold much moisture. And when it does snow, it just basically stays as snow. So liquid water is just something that's not very common. So how can anything live there in such an inhospitable place? These researchers studied 676 microbe genomes. Many of these showed evidence that they swapped genes with other microbes. Now, they suggest that this is from horizontal gene transfer, which is something we, I think

• 1:02:57Bob Novella?

  we've talked about here and there. It's well-known, and they contend that this is probably the key to survival in Antarctica for these microbes. Now, humans overwhelmingly engage in, not horizontal gene transfer, but what do they call it?

• 1:03:12Jay Novella?

  Sex?

• 1:03:13Bob Novella?

  Sex, right, Jay. They engage in sex. It's vertical gene transfer from parents, obviously, but some microbes can engage in horizontal gene transfer, basically getting genes from non-ancestors, which is kind of interesting if you think about it. So there's three types of horizontal gene transfer. Do you guys have any ideas of what some of them, or one of them, might be, or all three even? What are other ways to get genes into your genome, other than inheriting them?

• 1:03:41Jay Novella?

  A virus?

• 1:03:42Bob Novella?

  From your parents? Good, Jay. That's actually, that's absolutely correct. That's called transduction, when you're infected by a virus that accidentally transfers genes from a recent host's DNA. I didn't know that that was actually a thing until I did this. Yeah, I mean, I thought it was just like, this is its own DNA, you know, its own genome from the virus, and not like, taking some from the DNA. So that's transduction.

• 1:04:10Speaker 01

  That's happened historically, so we, like, I think there's a potato that has a horizontal gene transfer from a virus, from a bacteria or something. So that, yeah, that kind of trans-kingdom gene transfer can happen via viruses. But there's also, Bob, you can eat critters and have their genes incorporated into their genomes, like, fish eat bacteria or plankton or whatever who end up incorporating the genes.

• 1:04:36Bob Novella?

  Sure, that's called transformation, horizontal gene transfer. And yeah, basically scavenging DNA. Oh, look, here's some dead cells, I'm going to eat that, and somehow, that DNA gets into your genome. And there's one more, one more type of horizontal gene transfer, and that is...

• 1:04:51Speaker 01

  Genetic engineering.

• 1:04:52Bob Novella?

  Yeah, this is like, it's called conjugation, and that's where the DNA is transferred through these microscopic bridges, right, that are briefly created to facilitate that between microbial cells. So those are three different ways. So the researchers took a closer look at these, and when they looked at which genes were being transferred, they found that the genes that were involved in energy metabolism were essentially transferred disproportionately often. It was like, more than they would have expected. And the standouts among them were genes for aerotrophy, and that was one of the, one of the big surprising results from this study. Like, why, you know, why is so much aerotrophy in there? So let's actually define this. Aerotrophy is when bacteria, say microbes, actually pull gases like hydrogen, carbon

• 1:05:37Bob Novella?

  dioxide, and then some microbes at methane, and they gain energy from them, combining them with oxygen. So because they're combining them with oxygen, it's a chemical process that, in a, it's kind of analogous to a very slow kind of burning, right? You can kind of think of it that way. Steve, you say, you say very often that our old ideas of a hydrogen economy won't work. Well, this is essentially a hydrogen economy that clearly works for these microbes in Antarctica. Now, here's an, here's a really interesting angle here. I wasn't aware of this. If this energy is then used to fix atmospheric CO2 into biomass, then the aerotrophy then essentially becomes what?

• 1:06:20Bob Novella?

  Now, you're taking energy from the atmosphere, you know, you're taking the gas, combining it with oxygen, and you're taking that energy to convert CO2 into biomass. What you then become is a chemosynthetic organism. Specifically, it's called atmospheric chemosynthesis, and I've always been fascinated by chemosynthesis, because, I mean, no photosynthesis, you essentially do not need a sun to live your lives, which is, which is so interesting. Now, I've talked about it on the show a few times, and this is traditional chemosynthesis, and this is, this involves dissolved mineral compounds in water, right? You've got hydrogen sulfide near deep-sea hydrothermal vents, and you've got many, many organisms that can, that can use that energy source and then create an entire ecosystem

• 1:07:05Bob Novella?

  down there. So really fascinating. So this is atmospheric chemosynthesis for the microbes that turn this energy from the atmosphere into biomass. So other experiments that have been done with this have shown that even at minus 20 degrees Celsius, some microbes in some scenarios with some Antarctic soils can cover basically all of their basic energy requirements can be taken from these gases from the atmosphere. So that's pretty amazing. It's thought, though, it's also thought that this can even support whole communities, even when some of the microbes don't, you know, engage primarily in this chemosynthesis or aerotrophism, I guess, if that's even a word, because other bacteria can get carbon

• 1:07:49Bob Novella?

  potentially from eating their aerotropic friends, right? And they could also eat their carbon-rich ooze, which I, some of them exude. So that's interesting. So you can feed your buddies, if not everyone is aerotropic, you can still, your buddies can still get, can still create biomass based on the ones that are aerotropic. Now, aerotrophy is not uncommon, though. It's not like just these microbes in Antarctica have it. There are microbes throughout the world that have it, but it's only a backup energy source, right? It's a backup metabolism. If, if, if they're going through some very, like, rare, lean times, or it's just pure survival mode, they can kind of use that as a backup.

• 1:08:33Bob Novella?

  So it's not uncommon as a secondary thing. But for many of these Antarctic microbes, this is primary, and it's not just these microbes in the Antarctic desert soils, there's also other nutrient-starved places throughout the world, specifically like organisms that exist in volcanic deposits and deep-sea caves also can have aerotrophy as their primary, you know, their primary metabolism. So that's actually interesting. Wasn't aware of that either. So, all right, here's a question. What is a really, really beneficial byproduct of aerotrophy, besides creating biomass? Liquid water.

• 1:09:10Jay Novella?

  Water.

• 1:09:11Bob Novella?

  So when microbes oxidize hydrogen aerobically, water is created. Now, how cool is that? You're in one of the driest places in the world, and you could basically pluck liquid water out of the atmosphere all around you. An amazing, amazing, extremely beneficial byproduct of aerotrophy. So that was fascinating. Beyond this generally interesting aspect, so how can this information be helpful? I mean, okay, this is really interesting, and I loved researching it, but where, you know, where can they go from there? Where could, what could this potentially lead to? And I think that if understanding this process could help us understand how these microbes and Antarctica itself can respond to global warming, right? Because Antarctica is changing.

• 1:09:55Bob Novella?

  And global warming is going to have, could potentially have a dramatic impact, especially on these microbes, which is such an important part of this environment. So what's going to happen to Antarctica and the microbes, you know, this hydrogen economy down there as Antarctica gets warm? So we could learn about that. This could also help define what the actual limits of life are and how could it survive, not only in other earth-based extreme environments, but also other cold worlds that are in our solar system and beyond, which of course brings me to the idea of exoplanets and extraterrestrial life. Hydrogen, we know hydrogen is extremely common in the universe, right? Quite common.

• 1:10:35Bob Novella?

  There's many ice worlds out there, even ice moons that we've in our solar system. So potentially, maybe, perhaps, burning the hydrogen could potentially be a common strategy for life out there. So that's, you know, usually you don't think of that as something that could be really, really helpful for life unless you're looking at it through this lens. Now, we've heard, Steve, I'm sure you've heard of this, Jay, as well, for decades, astrobiologists have said something, they've had this aphorism, "Follow the water," right? Follow the water, find exoplanets, find places where there's water, because we know water is such an amazing solvent for life, that if you find water, then you're hitting the

• 1:11:18Bob Novella?

  jackpot, and who knows what else you could find from that. But this work suggests another potentially fruitful aphorism, and that's "Follow the hydrogen." So that's one of the takeaways, I think, from this study. So there we are, and I just love this, I love chemosynthesis and what these microbes are doing down there. Fascinating study. Check it out online for even more details. And one other interesting point from the study is that they found, I forgot about this, they found that these traits are heavily, heavily conserved evolutionarily, as you might imagine, because when they examined these microbes, they found microbes that were distantly, very distant from each other, separated by many, many hundreds of miles, thousands of miles.

• 1:12:01Bob Novella?

  They also have this, like, basically the same type, the same genes, and because they've probably been separated for many millennia. But evolution, you know, conserves these so well, they're so important, obviously. I mean, it's a source of energy. That evolution really, once a microbe has this, it wants it to keep them because it's just so amazingly helpful. So that was another big takeaway from this study. So that's all I got.

• 1:12:22Speaker 01

  All right. Thanks, Bob.

• 1:12:24Bob Novella?

  Sure.

• 1:12:25Speaker 01

  Jay, it's Who's That Noisy Time?

• 1:12:26Jay Novella?

  All right, guys. Last week, I played this noisy, and watch out, because it could be annoying. Lower the volume.

• 1:13:04Speaker 01

  What? What is it?

• 1:13:06Cara Santa Maria?

  I mean, it sounds enginey or like cutter, like it's cutting something, like a grinding, cutting kind of sound.

• 1:13:12Bob Novella?

  Can't disagree with that, that idea. I mean, sure, it's some engine powering up and doing something.

• 1:13:17Speaker 01

  Yeah, that's what it sounds like.

• 1:13:18Bob Novella?

  I mean, what else? I mean, I can't imagine that's biological.

• 1:13:20Speaker 01

  It could be a seal.

• 1:13:21Bob Novella?

  A cyborg seal, perhaps.

• 1:13:25Jay Novella?

  All right. Well, a listener named PJV, or Peter, first off, I have a correction from last week. Peter was the one that said, "A detector that is metal." I said some guy named George, it was not George, it was Peter who guessed that correctly. Now that that is cleared up, Gordy Swam wrote in and said, "Hello, this weekly noisy sounds like a bench grinder. My father has one, and I've used it many times. The initial sound is the spin-up sound, and then the sound of metal grinding, and the end is the sound of it winding down. This is not a bench grinder, but that is not a bad guess, because you are right, a lot of these sounds are kind of similar." Another listener named Casey Duro wrote in, said, "My guess for the noisy is an old-style high-speed camera.

• 1:14:09Jay Novella?

  The noise sounds like the mechanism inside the camera coming up to speed." This was a very good but incorrect guess, right? Because there are cameras, Bob, you know this. I think the IMAX cameras, or at least they used to do this before they were digital, where they were using film, that they'd have to push the film so fast through the camera that towards the end, the film would just start to rip apart. The unused film would disintegrate while it was in the camera.

• 1:14:37Bob Novella?

  Well, not only that, Jay, it was so noisy, I guess it's kind of obvious, it was so noisy that any sound actually from on set would be completely unusual, because all you could hear was the damn, you know, the damn motor, and everything, every sound you hear would have to be post-production.

• 1:14:53Jay Novella?

  Wow. I have another guest, Patrick Carr wrote in, and Patrick said, "Hi, Jay, and all the rogues, great show as always. My guess for this noise is that it's a large saw used to cut stone, and something goes wrong at the end, like the blade breaks," which sounds very deadly to me. That is not correct, but again, you know, the wind-up, the thing, the thing about this guess that I like is what is the story there, that the thing winds up and then something breaks? Because the correct guess was sent in by Brandon Whited? And I must be right with one of those two, Brandon? And Brandon said, "The noisy sounds like someone spinning up an object attached to a motor, increasing the RPMs until the force rips the object apart, perhaps a piece of

• 1:15:38Jay Novella?

  wood or a vinyl record." Now, this was close enough, because that's exactly what you're listening to, except it is a CD. So the wind-up is the motor just spinning it up, and it's a really crazy loud motor, but my God, the thing is spinning so fast. All right, so what force, though? What is doing this?

• 1:15:54Speaker 01

  That is centripetal force.

• 1:15:56Jay Novella?

  Correct, Steve, as usual. So thank you, Steve and Glaub, for sending in the correct guess this week. I will only play the end again, if you, I don't even think you want to hear it again, but here's the end part.

• 1:16:17Jay Novella?

  That's the guy who broke the DVD going, "Rah!" I think that his noise could have been a lot more interesting, but whatever. So I got a new noise for you guys this week. This one was sent in by a listener named Lacy No Wiki.

• 1:16:32Jay Novella?

  That's it. That is the noise that you get to guess this week. If you think you know what it is, or you heard something else, that's cool. Email me at [email protected]. Steve, this show will be out this weekend. It's not too late to go get tickets to see us at Dragon Con. You can go to theskepticsguy.org. We have one private show over Dragon Con weekend. It's on the 8th, it's on Saturday, the what, 5th, Steve?

• 1:16:59Speaker 01

  Yeah, Saturday, the 5th.

• 1:17:00Jay Novella?

  At 8:00 PM, you can go to theskepticsguy.org and check out the details on that. And then you'll also see a list of all the other things that we're doing, and we're talking about pretty much everything that's being talked about at Dragon Con. We're on every panel, we're talking about everything. We even have a panel on Kara and the shoes that she's not going to buy.

• 1:17:17Cara Santa Maria?

  Wow, that's weird.

• 1:17:19Jay Novella?

  Because you like that, don't you? We're doing a bunch of panels, robots, you know, robots in the apocalypse, robots that can run faster than Usain Bolt. It's like, it's crazy robot talk we're doing with Christian Hubicki, which I mentioned twice in this show already, and we'll be with him the whole weekend talking about robots, basically. Go check out all the different things.

• 1:17:43Bob Novella?

  Can't wait.

• 1:17:44Jay Novella?

  Yeah, it's going to be a lot of fun. Dragon Con's always a blast. It's kind of sensory overload for geekdom, but it's worth it.

• 1:17:48Speaker 01

  All right. Thanks, Jay. Couple of emails. The first one's just a simple correction. So last week, we were talking about peptides versus proteins, and I think we landed in the right place-ish, but just to clarify, because there was a third category that I was thinking of, that is why I was confused, that we didn't mention specifically. You guys know what that is? We're talking about peptides?

• 1:18:10Cara Santa Maria?

  Polypeptides?

• 1:18:11Speaker 01

  Yeah, polypeptides. So the difference between a peptide and a protein is entirely size.

• 1:18:17Cara Santa Maria?

  Yeah, they're short.

• 1:18:18Speaker 01

  Yeah, peptides are small, proteins are bigger.

• 1:18:21Bob Novella?

  And what's the cutoff?

• 1:18:24Speaker 01

  So, like, up to 50 amino acids is a peptide?

• 1:18:29Speaker 01

  And more than that is proteins are typically more than that. They they're usually in the thousands. But yeah, up to 50 amino acids is considered a peptide, two, two to 50. And then proteins are much bigger.

• 1:18:43Bob Novella?

  And and the peptides don't fold?

• 1:18:45Speaker 01

  No, they can't.

• 1:18:46Cara Santa Maria?

  No, they don't fold. They can't. They're too small.

• 1:18:47Speaker 01

  They may or they may or may not fold to some extent, but they don't fold up fully like the protein does because they're not big enough.

• 1:18:54Cara Santa Maria?

  They don't form like the sheets and the twists and all the different shapes.

• 1:18:56Speaker 01

  They don't have enough to do some of the structures that we typically see in proteins. But I was thinking of polypeptides. The difference between a polypeptide and a protein is entirely the folding.

• 1:19:07Cara Santa Maria?

  Yeah, because a polypeptide is a bunch of peptides.

• 1:19:11Speaker 01

  Yeah, so it is.

• 1:19:12Cara Santa Maria?

  It's more than 50 acids, amino acids.

• 1:19:14Speaker 01

  Yeah, so that's you. It could be it's bigger. It could be bigger than a peptide, right? So it's bigger than 50 amino acids, but it's not folded like a like a protein. As proteins require folding. All right. So that was just a a technical jargon thing. Okay. Next email comes from Gott in Norway, who writes.

• 1:19:31Bob Novella?

  Mein Gott.

• 1:19:33Speaker 01

  I saw this and found it compelling. They need links to a YouTube video of science communicator half hour about Max Planck Institute working decades on stellarator fusion design and 2023 spinning off company to try to make viable reactor. Okay, has anyone had a chance to look at this video other than me?

• 1:19:50Jay Novella?

  I watched the whole thing.

• 1:19:52Speaker 01

  Yeah, it's great.

• 1:19:54Bob Novella?

  Oh, my God. This is this is basically exciting the crap out of me because I I was not aware that eat, you know, Takamak designs like eater are these are intrinsically pulsed devices, which means that it needs to restart frequently, which is going to not only not only increase or decrease the lifetime of the device of this device, because it's, you know, heating it and then cooling it and then heating it and cooling it, you know, decreases your working lifetime because it's because you're, you know, you're you're shrinking and expanding and streaking and expanding. That limits the lifetime of of the device, but also you'll have much less the actual creating energy time because you gotta oh, gotta stop now.

• 1:20:36Speaker 01

  Let's back up a bit to explain what we're talking about. So you're talking about the Tokamak design, which is basically a donut, right? So you have, right, so yeah, donut like magnetically confined plasma inside a torus, you know, donut shape. The stellarator design is like that, but different in that that that donut shape, that torus, has a lot of fine detail to it. It's it's this curvy twisty shape. It's not just a straight torus.

• 1:21:04Bob Novella?

  Yeah, it's a helic, a helic twist.

• 1:21:07Speaker 01

  And the idea is that that.

• 1:21:08Bob Novella?

  Magnetic fields.

• 1:21:09Speaker 01

  It's following the flow patterns of the plasma and the magnetic fields down to like the millimeter resolution exactly, right? And this design has been tried out 50 years ago, and it was basically abandoned because it was too complicated. So what do you think is now making it possible, Kara, to to make such a complicated design that we weren't able to do 50 years ago?

• 1:21:37Cara Santa Maria?

  AI.

• 1:21:38Speaker 01

  AI, of course.

• 1:21:39Bob Novella?

  Because essentially. It's also, Steve, it's also modern manufacturing.

• 1:21:42Speaker 01

  Absolutely.

• 1:21:43Bob Novella?

  The final nail was was not a nail, but the final big advance was the computer controlled, you know, algorithms designing the magnetic fields shapes.

• 1:21:52Speaker 01

  Yes, so because and you think about it this way, that given this basic idea of a helical twisty torus, the design space in within that is massive. And AI is great at looking in a huge potential space and figuring out an optimal design, right? That's so we talked about this with protein designer with medication designer, material science, all these things, battery design, AI is helping with that. And so, yeah, so computer aided design has allowed the the modern now stellarators to have this optimized three dimensional structure. The other limiting factor, as Bob said, is material science, is the magnetic field itself.

• 1:22:33Speaker 01

  So here's one thing I learned that I maybe I forgot I heard before. I don't know, Bob, if you knew this. So if you had, so the big factor with magnetic confinement is how strong a magnetic field you can get.

• 1:22:47Bob Novella?

  Oh, yeah.

• 1:22:47Speaker 01

  So you have to use super strong magnetic fields. And they have to be superconductors because if any resistance in these powerful magnetic fields is going to heat up and it's not the whole thing's going to fail. So the question is, how powerful a superconducting magnet can you make to surround this donut, right? This this torus?

• 1:23:11Bob Novella?

  And Steve, that's, as you know, this is critically important because because fusion power scales to the fourth power of the magnetic field. That means that if you double the magnetic, the magnetic power, you get 16 times the output fusion power from the same volume of plasma fusion. And then if you triple it, that's 81 times. That's a that is such a gargantuan help that that you absolutely need to make that magnetic field as powerful as you ever like you possibly can.

• 1:23:39Speaker 01

  So so the the specific thing I learned is that the ultimate limiting factor on the power of the magnets is the point at which the superconducting breaks down.

• 1:23:52Bob Novella?

  Because the Cooper pairs.

• 1:23:53Speaker 01

  Yes, you get the Cooper. You need the Cooper pairs, right? These two electrons paired up together to to flow resistant list through the material.

• 1:24:01Bob Novella?

  And therefore no heat.

• 1:24:02Speaker 01

  Right, and therefore no heat generated. And that those Cooper pairs are ripped apart by really powerful magnetic fields. And so the the critical feature of a of a superconductor, so it's two things really. One is the temperature, the higher the better. But also the maximum strength of a magnetic field where it will retain its superconductivity. And so that is one of the innovations that they're talking about with this company that the the email that was linking to, where they they are using a new material, a rare earth, you know, kind of material that is a ceramic. It has a higher superconductivity temperature. So you go, you're increasing from four Kelvin to to what was it 60 Kelvin?

• 1:24:48Bob Novella?

  They're a little bit higher, but yeah, definitely liquid nitrogen level.

• 1:24:51Speaker 01

  But the the big thing is that the maximum superconducting magnetic field strength was 20 Tesla.

• 1:24:58Bob Novella?

  20 Tesla. It stays super. Yeah, stay superconducting at 20 Tesla. The previous type of current, the current magnets they used were 12 Tesla. The Cooper pairs would basically disintegrate, you know, if they break up.

• 1:25:09Speaker 01

  So this is almost double.

• 1:25:10Bob Novella?

  Basically, your device would would probably just melt and explode because you you'd it would all of a sudden be converted to this tremendous amount of heat. And then you're you're basically a billion dollars out.

• 1:25:21Speaker 01

  All right, so we have a AI designed stellarator. We have better superconducting ceramic material for the for the magnets. So they they feel, plus there's something else in there that they're secret, they're not telling, right? They're saying, yeah, the.

• 1:25:36Bob Novella?

  Secret sauce.

• 1:25:38Speaker 01

  The conducting wires in the magnets are like, yeah, we're not telling anybody what's inside there. That's our proprietary, whatever, information.

• 1:25:45Bob Novella?

  So they said reasonable, but we got to trust them on that, you know? This it reminds us of so many other things we've talked about in the past, especially that that battery, that that stupid battery technology that basically was the ultimate battery. And we're like, no way.

• 1:25:58Speaker 01

  But this is different, though, because that they said they built it and they they they wouldn't show it. This is like they're saying, we have the design. We have the material. We've done the proof of concept. So now what we need to do is build a test reactor. And then if everything works out, then build an actual commercial reactor. So they're saying.

• 1:26:19Bob Novella?

  My take, Steve, though, was that they were going to take that that was it the W seven X?

• 1:26:24Speaker 01

  Yeah.

• 1:26:24Bob Novella?

  Reactor, and they were going to essentially convert that into an economically viable react, working reactor. So that's actually if that's if I understood that correctly, that's much better than having to create it from scratch because they're talking like a million man hours to put this thing together. So if they just need to change the, you know, the, you know, the the magnets, you know, well, they're saying the 20, the 2030s.

• 1:26:49Speaker 01

  They're saying 2030s.

• 1:26:51Bob Novella?

  Early 2030s. So right now they're at the point where the design and simulation is largely largely completed. And there are portions of their research that are peer reviewed, which is always good, but I don't know what portions means. I don't know how much is peer reviewed, but peer review is good. And the company is apparently in full manufacturing processes. So so 2030s is around the corner. I mean, you know, I really was excited by this one. And then remember, this is one video we saw, but by the, you know, essentially, you know, assessment of the the big players involved. And these guys know what they're talking about. They want the main guy worked on a Takamak at MIT for 10 years. So he's he's in it. He's not just like an investor.

• 1:27:33Speaker 01

  But we've been here before.

• 1:27:36Bob Novella?

  We've been here, yeah.

• 1:27:37Speaker 01

  And and they're they're looking for 1.2 billion dollars of additional funding. When anyone's looking for funding, I take everything they say with a grain of salt, right?

• 1:27:46Bob Novella?

  Everything.

• 1:27:47Speaker 01

  And until the, you know, I'm skeptical until they freaking do it is the bottom line. Until they do it, they can generate commercial electricity, more electricity than you need to feed into the thing. And so it's, you know, and I think probably the 2030s is optimistic. Usually these things go over budget and long and whatever. And so, but we'll say, I mean, I I am hoping to be pleasantly surprised. But, you know, I I would be surprised if like before 2035, they have a commercial stellarator.

• 1:28:19Bob Novella?

  Absolutely.

• 1:28:19Speaker 01

  But let me say this one thing, Bob, just by coincidence, because I was researching news items for for this show, I came across a very relevant news item looking at high magnetic fields in a new material, not really new, but in a material called nickelates. And I don't know if you've heard about this. So they made a superconducting material known as samarium out of it's a nickel out of nickel based materials.

• 1:28:45Bob Novella?

  Right.

• 1:28:45Speaker 01

  And they discovered this new phenomenon. And it's really interesting.

• 1:28:51Bob Novella?

  A new type of superconductivity?

• 1:28:53Speaker 01

  What they found was that as you increase the strength of the magnetic field, the superconductivity goes away, right? The same way we talked about it breaks, it tears apart those Cooper pairs. But then if you keep going up on the strength of the magnetic field. It comes back.

• 1:29:11Bob Novella?

  Woof.

• 1:29:12Speaker 01

  Because of some secondary field that's being generated that counteracts the primary field. And then you get your, you get your super conductivity back.

• 1:29:22Bob Novella?

  Yeah, but what happens to all that heat that's generated in between?

• 1:29:25Speaker 01

  Well, I guess you gotta power through that, I guess, to get to the higher magnetic field.

• 1:29:28Bob Novella?

  That's not good. That might be untenable.

• 1:29:30Speaker 01

  But they said, but they did it, Bob, right? They, so they, they were able to, they reappear and they persist, guess how high of the Tesla they persist? So remember 20 Tesla is where we are with the current new, you know, batteries.

• 1:29:44Bob Novella?

  Okay, so say they can get to what, 50 Tesla, 75?

• 1:29:47Speaker 01

  60, 60.

• 1:29:48Bob Novella?

  60, alright.

• 1:29:49Speaker 01

  So that's

• 1:29:50Bob Novella?

  Nice.

• 1:29:50Speaker 01

  Triple.

• 1:29:51Bob Novella?

  That's triple. Do you know how much more fusion power you can get out of that?

• 1:29:55Speaker 01

  To the power of four, right? So what's three to the power of four?

• 1:29:58Bob Novella?

  Oh my god, man.

• 1:30:00Speaker 01

  So that is, that's 81 times the power, right?

• 1:30:04Bob Novella?

  They, they estimate that they can increase output 81 times by going from three Tesla to nine Tesla.

• 1:30:10Speaker 01

  Yeah, you triple it, you get 81 times more power. But if you go from 20 to 60, that's also a tripling, or 81 times.

• 1:30:17Bob Novella?

  Right, right.

• 1:30:18Speaker 01

  The, the power. So I don't know if these will be usable in a stellarator, but best case scenario, this might replace what they're using now. And we could have 60 Tesla fusion reactors rather than just the 20 Tesla ones. And that's a game changer, you know, in terms of the energy output.

• 1:30:37Bob Novella?

  Do they even have the 20 Tesla ones? I mean, essentially though-

• 1:30:40Speaker 01

  That's what they're saying, the material they're going to be building out of it is sort of the upper limit of their strength.

• 1:30:46Bob Novella?

  They, they were saying that the solution that they're working on would boost on access field strength from three Tesla to nine.

• 1:30:54Speaker 01

  So maybe the difference is sort of the maximum super conductivity field strength versus the on access field strength.

• 1:31:00Bob Novella?

  So, so compared to what they're claiming, Steve, this material can go from three to 60. So it's even, it's much more than 81 times.

• 1:31:08Speaker 01

  Well, again, I don't know if we're comparing apples to apples, because they said the maximum of that new material, the maximum super conductivity was 20 Tesla. And this, the, the, the nickelates, the maximum is 60. How that translates to whatever the on field, magnetic field, whatever they're talking about, I'm not sure. But if, if we assume it, it's a tripling, that's massive. That's an 81 times the energy output. So there's a lot of exciting things happening is the bottom line. There's advances in AI and material science and the theoretical control of the plasma. And we're getting into that zone where fusion could be possible, but I'm trying not to get excited. I'm remaining skeptical and I basically, I'll believe it when I see it is still where I am.

• 1:31:52Bob Novella?

  Yeah.

• 1:31:53Speaker 01

  Uh, but early 30s, early 2030s is like the earliest anything could possibly happen in terms of commercial fusion.

• 1:32:00Bob Novella?

  So I like having so many irons in the fire, though, you know?

• 1:32:04Speaker 01

  Yeah, yeah, yeah, yeah.

• 1:32:05Bob Novella?

  But I'm also very disappointed that, that Takamak, Takamak Fusion Designs have that Achilles heel that I wasn't aware of. I mean, they, they could still get it to work, but it's very limiting. I'm surprised I wasn't aware of it, but oh well, this, this, this isn't limited by that. So hopefully-

• 1:32:19Speaker 01

  But I wonder if it does have a limitation they're not telling us because they're looking for funding. You know what I mean? I'm just always worried about that.

• 1:32:25Bob Novella?

  Right, true. And, and it's amazing how they, as you go through the, um, the video, Steve, they're, all the roadblocks that they hit were just like, oh my god. They solve it and then there's another but. Then they solve it, then there's another but. And but, but they keep, they, they kept solving them. And now they, they may, at the very least, Steve, um, with their, their latest stellarator design solved the biggest problem that stellarators had for decades, and that was leaking, you know, leaking, you know, leaking the plasma. That was something they could not get past. And their latest design got rid of that leak. That was the biggest, biggest problem. And that solved. So that, that was a huge gain unto itself. And, and based on their performance, their latest design performance, the stellarators

• 1:33:08Bob Novella?

  are definitely, you know, need to be taken seriously now.

• 1:33:11Speaker 01

  Absolutely.

• 1:33:12Bob Novella?

  And they're, they're really, they're really joined the big boys on the block to, to a large extent.

• 1:33:16Speaker 01

  Alright, thanks, Gao. Let's move on. Guys, it's time for science or fiction?

• 1:33:24Speaker 00

  It's time for science or fiction?

• 1:33:33Speaker 01

  Each week I come up with three science news items or facts, two real and one fake that I challenge my panel. Let's get this, tell me which one is the fake. Just three regular news items this week. You guys ready?

• 1:33:44Jay Novella?

  Yep.

• 1:33:45Speaker 01

  Here we go. Item number one, researchers have developed an AI model that can diagnose hypertension with more than 90% accuracy and diabetes with more than 80% accuracy from a five second facial video.

• 1:34:00Speaker 01

  Item number two, material scientists have demonstrated a living textile, a fungus-based fabric that can be used to make clothing. And item number three, new research shows how simply leaving reviews at common sites like Yelp can lead directly to identity theft. Jay, go first.

• 1:34:18Jay Novella?

  Researchers have developed an AI model that can diagnose hypertension with more than 90% accuracy and diabetes with more than 80% accuracy from a five second facial video. What? Right out of the gate, I, I can't imagine how it can, it can, can model, developed an AI model, it can diagnose hypertension. Alright, so 90% accuracy with hypertension, 80% with diabetes. There's no blood test. Doesn't, it says a facial video. It doesn't even say like a, a five second facial scan or, you know, it's like, is it reading the person's temperature? You know, what could the video find in the face that would be telling of these two diseases? I, oh my god, I can't imagine. Other than like, other than like

• 1:35:01Jay Novella?

  temperature, what else could it possibly do? And what would that even tell you? Well, I'm missing something.

• 1:35:06Bob Novella?

  Well, Steve, I'll, I'll throw out a question for the first one here. Is it, is this purely visual?

• 1:35:12Speaker 01

  Yes.

• 1:35:13Bob Novella?

  I was thinking it could be, it could use smell a vision.

• 1:35:15Jay Novella?

  Yeah, but it's, I know, I thought about that too, but it, it does, it's just video. So the AI is what, is seeing the video of, of people and it's able to determine what, what are they doing? I mean, is, it can't be the quality of their skin, the color of their skin, something to do with their eyes? Oh god, that one is crazy. I mean, my gut's telling me fiction, but I gotta keep going here. Second one, material scientists have demonstrated a living textile, a fungus-based fabric that can be used to make clothing. A living textile. It's fungus, guys. They want us to wear mushrooms. Can you like eat it while you're walking, Steve?

• 1:35:52Speaker 01

  Uh, they didn't say anything about how edible it might be.

• 1:35:56Bob Novella?

  Maybe you could soak it with nutrients and suck on your sleeve.

• 1:35:59Jay Novella?

  Okay, so is it fungus? It's not living when it's on you. I mean, I will, I have an anecdote about mushrooms.

• 1:36:06Bob Novella?

  Ooh.

• 1:36:07Jay Novella?

  That could talk about their durability. Um, my daughter left one of her blankies out when she was a kid and it was outside for a long time and I ended up finding it. And I brought it in and I put it in the washing machine and mushrooms had grown through a part of it, right? So I washed it. Mushrooms were still there. I washed it again. Mushrooms were still there. I washed it again on sanitary. The mushrooms were still there. So there's, there is a durability thing here about mushrooms. Um, well, you know, what, what else have I tested mushrooms for durability, Bob? You know what I mean?

• 1:36:39Speaker 01

  Yeah, right. I mean, that's a good anecdote.

• 1:36:40Jay Novella?

  After seeing that, you know, I couldn't scrape the mushrooms out. I would think that they, they do have some type of protein durability or whatever that, that could probably be selectively manipulated to, uh, to have more properties. Steve didn't name a specific fungus or a mushroom, right? He's just saying in general.

• 1:36:59Speaker 01

  Oh, do you want me to tell you what specific fungus they use?

• 1:37:02Jay Novella?

  If it exists, go ahead. Yeah.

• 1:37:04Bob Novella?

  Cordyceps?

• 1:37:05Speaker 01

  It is a cordyceps.

• 1:37:06Jay Novella?

  What?

• 1:37:07Cara Santa Maria?

  It's always a cordyceps.

• 1:37:09Bob Novella?

  Zombie cordyceps.

• 1:37:11Speaker 01

  It is cordyceps, cordyceps militaris.

• 1:37:16Jay Novella?

  Yep, there it is.

• 1:37:19Jay Novella?

  Yeah, those are bastards, Steve. We know, you know, we fought them in the war. They're durable. Alright, so I mean, I think this one's plausible. Doesn't mean that they have the clothes yet or whatever. They could just be speculating. Hey, this is durable. We could, it could be made into, you know, underpants, whatever. Um-

• 1:37:34Speaker 01

  No, they, I'll say they made a, an article of clothing out of it.

• 1:37:38Jay Novella?

  Was it a bandana?

• 1:37:39Speaker 01

  It was a dress.

• 1:37:40Jay Novella?

  It was a dress. Okay. I think that one is science. Third one, new research shows how simply leaving reviews at common sites like Yelp can lead directly to identity theft. Alright, so out of the gate, this is sciency to me. One, because, you know, you're identifying a couple of things here. You're saying, sometimes your name is there, right? I don't know if you, if, you know, if you would expose your email, but maybe someone exposes their email on Yelp. And then you're also talking about a review, which means like, you know, you're, you, you purchased a service or you, you interacted with a brand or whatever. Is there enough information there for that person to then be able to get you? And I think the answer is yes, because then you could fashion an email that could pretend to be from that

• 1:38:25Jay Novella?

  brand. And they could, uh, they could spoof you. And you might think you got a coupon or something and they might capture more of your data that way. So I think that one is science. I think the second one is science. I think the first one is, is fiction, Steve.

• 1:38:39Speaker 01

  Okay, Kara?

• 1:38:40Cara Santa Maria?

  I don't know if I agree, and I think that I'm seeing, trying to always read these for specific words. I feel like lately it's been, well, the operative word here was, um, so the AI model that can diagnose hypertension with 90%, diabetes 80%. Okay, so neither of those are perfect. Um, a five second facial video. Okay, so can you see facial components that are diagnostic of high blood pressure and diabetes? Maybe. I mean, yeah. I wouldn't be surprised if this was science. I think just like increasingly we're seeing that

• 1:39:21Cara Santa Maria?

  from a radiological standpoint, basically like pattern, uh, detection, image pattern detection. AI is just like getting really good at that. So that one doesn't surprise me.

• 1:39:32Cara Santa Maria?

  The living textile, I don't see anything wrong with this either. So it's fungus based and it can be used to make clothing and you didn't give any other details. So you didn't make any claims about it being like good at it or having any sort of like cool powers or anything. So that one doesn't really bother me. The one that bothers me is the Yelp one because it says, new research shows how simply leaving reviews. Okay. So I go online, I leave a Yelp review. It leads directly to identity theft. So somebody could hack that information and steal my identity as opposed to Jay just like told a meandering way that you would get identity theft. Like somebody would figure out who you are based on certain details and

• 1:40:16Cara Santa Maria?

  blah, blah, blah, blah, blah. So I have a feeling it can lead to identity theft, but I don't think that simply leaving the review does. And I don't think it leads to it directly. So I'm going to say that's the fiction.

• 1:40:27Bob Novella?

  Okay, Bob. Well, I mean, I don't know how you could say that, that this is, that three is false because leaving a review absolutely cannot lead directly to identity theft. I don't think you could, it'd be hard to make a blanket statement like that because who knows what subtle, you know, information might be usable.

• 1:40:47Speaker 01

  All right. I'll, I'll give you that just because this is a generic thing about how I formulate science or fiction. This is like the, in order for this to be fiction, you don't have to believe that it's impossible for this to happen. You just have to believe that there's no research showing that it can happen. Right?

• 1:41:04Cara Santa Maria?

  Exactly.

• 1:41:05Speaker 01

  I'm not making the opposite claim.

• 1:41:07Bob Novella?

  That it's, that, right?

• 1:41:09Cara Santa Maria?

  That it's impossible. Yeah. It's not worded like research shows that, yeah, that under no circumstances can a direct, you know.

• 1:41:17Bob Novella?

  All right. That's fair, I guess.

• 1:41:18Speaker 01

  Don't overread. That's such a generic thing about science.

• 1:41:21Bob Novella?

  Okay, that's fine. Reasonable. Um, yeah, living, living textile sounds totally reasonable. I mean, the other ones, yeah, I mean, I could also see that, um, that the hypertension one, there could be some, I think it's something about the eyes that can be diagnostic or, you know, correct 80% of the time anyway, uh, something very subtle, some interesting pattern. Given what I learned about the third one, then, then I'm, I'll do, I'll say the Yelp one is fiction as well. Go along with care of the hair and see what happens. So this was another tough one. Good job.

• 1:41:55Speaker 01

  All right. So you both agree on number two. So we'll start there. Material scientists have demonstrated a living textile, a fungus based fabric that can be used to make clothing. You all think this one is science and this one is science. So yeah, this is from the Chinese Academy of Sciences. They developed a flexible living textile based upon caterpillar fungus or Cordyceps militaris. Uh, and they said it's, um, it's fairly, they, supposedly they have to, they dry the sheets of it, but then it's brittle. So they have to treat it with, they treat it with glycerol, which then makes it more pliable. You can twist it into, into ropes or you could use it as a sheet. And then with the sheets, they could actually sew it into clothing. It has, you know, it's sufficiently strong to do that.

• 1:42:38Speaker 01

  And you could also incorporate other organisms, um, in order to color it. It's like, so it could be naturally colored and you can incorporate organisms to make it resistant to UV light. You know, so.

• 1:42:51Bob Novella?

  Organisms?

• 1:42:52Speaker 01

  Yeah. Like Aspergillus niger, for example, could provide ultraviolet protection by, um, because they're melanin rich, for example.

• 1:42:59Bob Novella?

  Interesting.

• 1:43:00Speaker 01

  And yeah, so they were able to make, make functional sheets, make it into a dress. They demonstrated the material could be shaped and sewn into larger structures.

• 1:43:08Bob Novella?

  How about washing it?

• 1:43:10Speaker 01

  So that's, that's the next step. They need to now see how, you know, how could it be manufactured in a consistent way? How washable is it? How abrasion resistant is it? Uh, they also say that, so it, because it's dried, uh, it's like in a dormant state, but you could like, you have to say you have a tear in it. You could add liquid to it. And they said theoretically, it could start to grow and repair itself.

• 1:43:31Bob Novella?

  That's cool.

• 1:43:33Speaker 01

  Yeah. So again, that, we need to see if that's actually functional, but like theoretically they said that was a possibility. All right. All right. Let's go back to number one. Researchers have developed an AI model that can diagnose hypertension with more than 90% accuracy and diabetes with more than 80% accuracy from a five second facial video. Uh, Jay, you think this one is fiction. Bob and Kara, you think this one is science and this one is science. This is very cool. Sorry, Jay.

• 1:44:01Bob Novella?

  Nice.

• 1:44:02Speaker 01

  So.

• 1:44:03Cara Santa Maria?

  Huh.

• 1:44:04Speaker 01

  Think about this though, guys. It's a five second video, not a picture. So it's not information you can get from a still frame.

• 1:44:12Cara Santa Maria?

  Right. So you can see like vessels, like pulsing and stuff.

• 1:44:16Speaker 01

  Exactly. And how do they see vessels pulsing?

• 1:44:18Cara Santa Maria?

  Uh, with like filters?

• 1:44:20Bob Novella?

  Uh, yeah. What, how they, what, what wavelength of light are they looking at?

• 1:44:23Speaker 01

  Yeah. That's the key. So Bob, when you asked that question, you like, let me ask you a question. I thought you were going to ask something relevant to this, but you just said, is it video? I was like, no, it is video. Um, but it's not just regular video. Uh, what kind of video do you think it might be?

• 1:44:37Bob Novella?

  Is it infrared?

• 1:44:39Speaker 01

  No.

• 1:44:40Bob Novella?

  Ultraviolet?

• 1:44:41Speaker 01

  No.

• 1:44:42Cara Santa Maria?

  Are we polarized? I don't know.

• 1:44:43Speaker 01

  Not polarized. Spectroscopic. Um, so it's essentially, instead of having like each individual pixels, either red, green or blue, each individual pixel could be any color of the entire spectrum. Right? And so you need that data. And with that, you can look like just under the skin and you could see the blood vessels pulsating under the skin. And that could give you information about the pulse wave, pulse wave dynamics. And that tells you basically how stiff, you know, the, the blood vessels are. And with that information, they could accurately diagnose diabetes and hypertension. If you go to 30 seconds rather than just five seconds, then it gets even better. So it, the, it increases

• 1:45:28Speaker 01

  to 95% for hypertension, 89% for diabetes. Uh, but also, you know, you could break it down to like, how sensitive was it at detecting normal blood pressure versus hypertensive blood pressure? It was a hundred percent at detecting normal blood pressure. Every time it said it was normal, it was normal. It was 89.2% with hypertension with a 30 second video.

• 1:45:51Bob Novella?

  Could it spit out, uh, diastolic and systolic?

• 1:45:54Speaker 01

  No.

• 1:45:55Bob Novella?

  Can it keep?

• 1:45:56Speaker 01

  No, I was just telling you how stiff, how stiff the arteries are. Yeah. So pretty cool. I mean, this could be used as a screening test. You know, a lot of people out there with hypertension who don't know it and they're, because it could be asymptomatic and they're not being treated for it. And it's a huge risk factor.

• 1:46:12Cara Santa Maria?

  Yeah. Like, you know, the cuffs at the grocery store. What if you just like do a, you know, you stand in front of a camera for five seconds?

• 1:46:18Speaker 01

  Exactly.

• 1:46:19Cara Santa Maria?

  That's cool.

• 1:46:20Bob Novella?

  Yeah, Steve, that'd be great for mom. Because of the, the medication that mom's on, uh, taking her blood pressure, it's painful for her. Um, so if they could use that for her, that'd be wonderful.

• 1:46:30Speaker 01

  All right. Which means that new research shows how simply leaving reviews at a, at common sites like Yelp can lead directly to identity theft is the fiction and Kara, you're right. The operative word there was directly. Uh, that's not what it does. So what the study does show is that if you leave reviews on Yelp, that you can use AI, of course, AI is everything now, right? You can use AI to see patterns of, of reviews, including like the length of your review, for example. And you can use that to infer, uh, what people you have online relationships with. And then that information could be used it for what's called a spear phishing campaign where you call somebody pretending

• 1:47:15Speaker 01

  to be somebody they know in order to trick them into giving you information or giving you money or whatever. So it's, it's indirect, right? It's like you're using it really just to infer who their network of people are. And then be like with 50% accuracy. And then so you have a 50, 50 chance of guessing right and of telling somebody, hey, I'm this person from your network and I need you to give me your credit card information. And it's highly effective. They, they show.

• 1:47:43Bob Novella?

  Really?

• 1:47:44Speaker 01

  Yeah. I mean, it's very, I mean, it's always a numbers game, right? The more number of people. So they said with 500 attempts, with 500 attempts, they were able to get a hundred and nine percent return on their investment. Basically whatever money they spend doing this, sending emails and doing everything, they make, you know, but you know, they make that, double that money back with 500 attempts. With 10,000 attempts, their profit goes up to 1,098%. And that 10, 10,000 attempts, that's not that much for.

• 1:48:13Cara Santa Maria?

  A computer.

• 1:48:14Speaker 01

  Yeah. Or not just a, this, this is people, right? So it would be, unless of course using AI voice, you know, duplication. But if you have a, what do you call those? You play a room with a bunch of people, you know, scamming.

• 1:48:25Cara Santa Maria?

  Oh, like a, not a bot farm, but like, kind of like a bot farm.

• 1:48:28Speaker 01

  Yeah. It's like a scam farm, you know? Uh, yeah. So they, they, you could make basically a lot of money just using, you know, basically reviews to infer people's networks of connections. Uh, but yeah, there's not, it's not like directly giving identity theft information from leaving the review.

• 1:48:44Bob Novella?

  Mm-hmm.

• 1:48:45Speaker 01

  All right. Bob, give us a quote.

• 1:48:49Bob Novella?

  All right, guys. So I was, uh, bemoaning the, the sorry state of science research in, in our damn country and, uh, came across these remarks by President Obama in 2009 at the National Academy of Sciences annual meeting. He said, Lincoln refused to accept that our nation's sole purpose was mere survival. He created.

• 1:49:12Bob Novella?

  This academy, meaning National Academy of Sciences, but believing that we must add, and I quote, "The fuel of interest to the fire of genius in the discovery of new and useful things," and that's directly from Lincoln there, that "fuel of interest to the fire of genius in the discovery of new and useful things," that's directly from Lincoln. And then he finished a little bit later in his speech, saying, "At such a difficult moment, there are those who say we cannot afford to invest in science. That support for research is somehow a luxury of the moment defined by necessities. I fundamentally disagree. Science is more essential for our prosperity, our security, our health, our environment, and our quality of life than it has ever been before."

• 1:49:55Speaker 01

  Absolutely.

• 1:49:56Bob Novella?

  Yes, we said that in 2009, and before he announced a huge, I think, a historic level of support for science research. And it's just like, ugh, you know, it's just so frustrating.

• 1:50:08Speaker 01

  And now that infrastructure is being hollowed out.

• 1:50:10Bob Novella?

  Hollowed out, you know, just like, not just, not just, Obama was bemoaning, you know, the decrease in support. And what we're seeing here is beyond a decrease.

• 1:50:22Speaker 01

  It's a walking away of support.

• 1:50:24Bob Novella?

  Hollowing out, it could take years, decades, generations to get it back to where it was. It's just like, it's just like, you know, like a crime of the century. It's just like, how is this happening? How could we allow it to happen? It's just like, I can't even think about it for too long.

• 1:50:41Speaker 01

  It's so sad.

• 1:50:42Bob Novella?

  It's just going to a bad place.

• 1:50:43Speaker 01

  Our investment in science and the amazing infrastructure of research that we built in this country is our most competitive strength. It is the best thing we have going. And just giving away that advantage, just destroying it for no reason other than just pure, unadulterated ignorance, is, it is a crime. It is an absolute shame. It's hard not to get furious every time I think about it.

• 1:51:06Bob Novella?

  Yeah.

• 1:51:07Speaker 01

  And, you know, all of the research findings that are not going to happen over the next 20 to 30 years because of this, it's just hard to, hard to even wrap your head around. It's hard to estimate. Scientific research is an investment. It's not an expense. It is an investment, and it's the best investment we have going, and we have an infrastructure for it. Just destroying it this way is just terrible.

• 1:51:28Bob Novella?

  The envy, the envy of the world, and now it's just like, now it's the opposite.

• 1:51:31Speaker 01

  And the thing is, you can't just flip the switch back on, because we have scientists who are moving their research to other countries, who are, who are now capitalizing on our folly, right? There's now a reverse brain drain away from the US because of this. And they're not just going to snap right back. Why would they? Now they're going to have a lab and investments and everything in other locations that are happy to lure them away, you know, from the US. Yeah, that's, that's generational damage. We may never get to the place where we were before. You know what I mean? Like, the world will never be the same.

• 1:52:07Bob Novella?

  And the only little silver lining that I see is that, that if these researchers go to other countries to do their research, at least the research is being done, you know, and some of these discoveries will still be made, just not in my own country. And that's, that makes me feel better, because at least the research, you know, some of it, some of it will get done.

• 1:52:28Speaker 01

  Yeah, the fact that researchers are going elsewhere and they're doing that research elsewhere is, yeah, it's still, as again, as an American, that's, that's bad. But in terms of just as a science enthusiast and as a human being, that's not that big a deal.

• 1:52:43Bob Novella?

  Right, exactly.

• 1:52:44Speaker 01

  But it's not like, like, this is just a movement of the research. The amount of investment in this research is going, is reducing, is going down, because, you know, no one has the resources that the US has. And again, they don't have the same infrastructure that we have, so it's not necessarily going to be as effective or efficient or whatever. They will build, they will build their infrastructure, they will build their research, but it's not going to exactly replace what we are losing. There's going to be a massive net negative to scientific research.

• 1:53:13Bob Novella?

  Absolutely.

• 1:53:14Speaker 01

  That's the bottom line. Terrible. Okay. Sorry to end on that sour note, but it's something that we need to be aware of, because this is like happening in the background. I don't know how, how aware people are of it. Read about it. Look, you know, we've talked about it, but just do some research about it as well. Just, you know, summaries of like, what's been happening to science research in the United States, you know, since Trump 2.0.

• 1:53:34Bob Novella?

  Right.

• 1:53:35Speaker 01

  It is a scandal. It's a scandal. We can't talk about it enough, actually. It is an absolute scandal.

• 1:53:40Bob Novella?

  Tragic.

• 1:53:41Speaker 01

  Tragic. All right. Thank you all for joining me this week.

• 1:53:44Bob Novella?

  You got it, brother. Anytime.

• 1:53:45Cara Santa Maria?

  Thanks, Steve.

• 1:53:46Speaker 01

  Next week, we're using, I think, the show that we recorded in LA, because we'll be away at Dragon Con. And the week after that, I think, we'll, I don't know if we're going to use one of the episodes from Dragon Con or not. We'll see. But we will be back with a regular show in two weeks, probably. 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 theskepticsguide.org. Send your questions to [email protected]. And if you would like to support the show and

• 1:54:25Speaker 01

  all the work that we do, go to patreon.com/skepticsguide and consider becoming a patron and becoming part of the SGU community. Our listeners and supporters are what make SGU possible.