SGU Episode 984
|
This episode needs: proofreading, formatting, links, 'Today I Learned' list, categories, segment redirects. Please help out by contributing! |
How to Contribute |
| SGU Episode 984 |
|---|
| May 18th 2024 |
 |
"Researchers utilize Nvidia’s Eureka platform, a human-level reward design algorithm, to train a quadruped robot to balance and walk on top of a yoga ball." [1] |
| Skeptical Rogues |
| S: Steven Novella |
B: Bob Novella |
C: Cara Santa Maria |
J: Jay Novella |
E: Evan Bernstein |
| Quote of the Week |
Science is the only self-correcting human institution, but it also is a process that progresses only by falsification. |
Allan Sandage, American astronomer |
| Links |
| Download Podcast |
| Show Notes |
| Forum Discussion |
Introduction, aurora viewing from solar CME
Voice-over: You're listening to the Skeptics' Guide to the Universe, your escape to reality.
S: Hello and welcome to the Skeptics' Guide to the Universe. Today is Wednesday, May 15th, 2024, and this is your host, Steven Novella. Joining me this week are Bob Novella...
B: Hey, everybody!
S: Cara Santa Maria...
C: Howdy.
S: Jay Novella...
J: Hey guys.
S: ...and Evan Bernstein.
E: Good evening everyone.
S: How's everyone doing this fine evening?
C: I am great, but I'm so sad that I could not see the Aurora Borealis because I live so far south.
B: I don't want to talk about it.
S: We tried really hard. We literally set our alarms for like one in the morning and got up like three nights to try to see if we could see it. So really, the auroras existed only that first night, that Friday night, and from our perspective, it was cloudy, and then it was a little bit too much light pollution, and we never saw anything. Now, people took pictures, but the thing is, the camera will show the aurora way better than your eyes did. Right?
E: Yeah, that's right. It makes all kinds of difference.
C: And it lets in so much more light.
S: Yeah. You saw these amazing pictures, but that's not what people were seeing with their eyes.
C: But there are places in the world where like a friend of mine got a picture from a friend in Scotland, and they were like, it looked like this. Like there were places in the world where it was really intense and bright, but I think probably at very, very high latitudes.
E: Yeah, 50 or above.
S: Or dark sky locations probably had a good view. If you had a clear night and a relatively dark sky, it probably would have been beautiful. We had neither, unfortunately.
S: I mean, we're in Connecticut. It's a long shot, but this was a reasonable opportunity.
C: I mean, people saw aurora just north of where I am. Lots of my friends got in the car, drove up north and had a light show. It's just in the city. In LA, I couldn't see anything.And I don't even know if it reached this far south.
E: Well, wait a minute. I thought the curse was over.
B: So did I.
E: We had a whole segment about that.
B: This would have been such a great capper to the year of seeing an eclipse and seeing aurora borealis would have been an amazing one-two punch. Oh, well. I mean, I still don't think it would beat. It still wouldn't beat the eclipse, but that's on my bucket list. I want to see some wicked auroras, whether it's aurora borealis or australis, I don't care. Whatever side of the planet, I don't care. I just want to see it really, really good. And one of these days, I mean, we still have more time.
S: You're going to have to plan a trip with a high probability.
C: Yes. And statistically, I think it's going to be easier for you to see that. Much easier.
B: Maybe, man. You wouldn't have to seriously travel.
C: You would. You would have to seriously travel, but you don't have to time your travel so perfectly as for an eclipse.
B: True. True.
E: Oh, yeah. Correct.
B: It's tough. I mean, maybe my daughter may have a little bit of the curse. She's been in Alaska for a past year and she hasn't seen. She has not seen.
C: What?
B: In Alaska. So it's kind of crazy.
S: How is that possible? Is she overcast all the time?
C: Is she colorblind?
B: You know, there just hasn't been. Maybe she could have seen one if she traveled a little bit here and there, or maybe, you know, I'm sure it's been cloudy or whatever. She hasn't seen them. She's still waiting. So she's frustrated as well. But yeah, I mean, in Alaska, I have not talked with her in the past week, though. So I don't know if she had a chance this past week. I'm hoping she did.
C: She had like a religious experience and now she hasn't called you.
B: Good luck, Ashley.
E: That's right.
S: All right. We're going to get right into it this week.
Special Report: Steorn Free Energy Update (3:42)
- [url_from_show_notes _article_title_] [2]
S: So, Evan, tell us, give us an update on the Steorn Free Energy Company.
E: The Steorn Free Energy Company, yeah, yeah. So we have a Lister Philip from Ireland. He was the inspiration behind this report. He emailed us the other day to share with us the latest news about the Steorn Company. But I'm going to take you through the timeline about when we first had Steorn on our radar right up to today. So this is effectively a five to 10 years later segment, even though it's longer than that. But I'll explain. So it was December of 2006. Oh, my gosh. We were only podcasting for barely a year at that point. Gee whiz. The company, based out of Dublin, Ireland, called Steorn. That's spelled S-T-E-O-R-N. The O is silent, at least in our language. They promised to demonstrate their Orbo device. Oh, OK. What was the Orbo device? A device that converts time into energy. No, that's not a catchphrase. And that's not marketing jargon. That is the actual claim, that they have a device that can generate energy from the flow of time. I kid you not. Right from the words of Steorn's founder and CEO, Sean McCarthy. How exactly did this device convert time into energy? Here was his description of the device. This is straight from him and the company. Orbo is based upon the principle of time-variant magneto-mechanical interactions. The core output from our Orbo technology is mechanical. This mechanical energy can be converted into electrical energy using standard generator technology either by integrating such technology directly with Orbo or by connecting the mechanical output from Orbo to the generation technology. Here are some relevant quotes from Sean McCarthy. This was at the time. These are direct quotes. Technically, it isn't a battery at all. You'd call it a battery substitute technology. It's something that replaces the function of the battery. It's really a generator rather than a storage device. What we found is that we could speed up and slow down electromagnetic fields, which traditionally should travel at the speed of light. When we did that, we got these energy anomalies. So I suppose you could take a view which we are not qualified to make scientifically. But if you ask our guys what we really think we are doing, we think that we are converting time into energy. Which sounds very grandiose when you see three strips of metal and two wires. As to whether it is really contravening any laws of physics, there are bigger questions there. All we can say is that it's not degradating or drawing on any known energy source, but there are vast swaths of energy that we call dark energy in scientific terms. Yeah, that gobbledygook really can be condensed down to two words. Perpetual motion or free energy. So pick your poison as to which of those two-word phrases you prefer. Free energy machines over unity machines, perhaps you've heard them referred to as. They produce more, they claim, they produce more energy than they consume. Greater than 100% efficiency, right? So there you go. Unity means greater than 100%. But then you have this little something called, what, the laws of thermodynamics. First law of thermodynamics, you cannot get more energy out of a process than you put into it. Second law of thermodynamics, you cannot even get the same amount of energy out. There will always be some loss of the energy because of the work that's being done. That's entropy. And you can think of it in terms of, what, the conservation of mass and energy. Energy cannot simply come from nowhere. There has to be a source of the energy. So in July of 2007, that big day came in which Orbo was to have its first public demonstration. They turned it on, and guess what, it made that sound that the Millennium Falcon made when its hyperdrive system failed. Not even a punch from Han Solo could start it back up.C Kara, for your benefit, those are Star Wars references.
C: Thank you.
E: And when they couldn't get the thing to work, the folks at Steorn resorted to, yep, special pleading. For those of us paying attention at home for these last 19 plus years, special pleading occurs 100% of the time as an explanation given by pseudoscientists as to why their pseudoscience failed. In fact, we should probably establish a new physical law of nature called the law of special pleading because it's predictable that 100% of the time it will occur. So yeah, here's what they said, technical problems, this was their plea, technical problems arose during the installation of the demonstration unit in the display case. These problems were primarily due to the excessive heat from the lighting in the main display area. Yep. Attempts to replace those parts affected by the heat led to further failures, and as a result, we have to postpone the demonstration until a future date. Okay, so the Orbo was not 100% ready. Fine. All they had to do was make a few slight adjustments. They would come back soon with a new public demonstration that was absolutely guaranteed to work. And those slight adjustments would take three years to complete, basically. While this was all going on during the same timeframe, Steorn, they appointed a jury of 22 scientists and experts to evaluate the technology and the claims. This was an independent group of experts, none of them invested in the company, basically a third party people, and they had the expertise to evaluate the Orbo technology to see if it actually works. So, while Orbo is getting their tweaks done over the course of three years, simultaneously it's being evaluated by non-affiliated experts, and they also took about three years to evaluate the device and came up with their results in 2009. Here's what they said. 22 independent scientists and engineers were selected by Steorn to form a jury. It has, for the past couple of years, examined evidence presented by the company. The unanimous verdict of the jury is that Steorn's attempts to demonstrate the claim have not shown the production of energy, and the jury is therefore ceasing its work. Oh, well, that was unfortunate for Steorn. I guess you can't really get energy from nothing. Who would have guessed that? However, this result was not enough of a dissuasion to stop Sean McCarthy and Steorn from marching forward with a new and improved demonstration. After all, when your company raises, ultimately, $23 million of capital from investors, you might say that they were, yeah, invested in yielding some kind, any kind of positive result. What do you do in this case? Well, what you do is you go full Fyre Fest. You're in way too deep to stop, both financially and reputationally, despite the inevitable outcome of a purely negative result. So bang, they did it again in 2009. They resolved their key technical problems, so they said, related to the implementation of the Orbo, and they're focused on the commercial launch towards the end of this year. With a demonstration, and a demonstration they did. The second demonstration, it ran from December 2009 to February 2010, and it involved a motor powered by a battery, which, upon observation, provided no independent evidence that excess energy was being generated. It was dismissed by the press as an attempt to build a perpetual motion machine, and overall called a publicity stunt. Yep. Here's the best headline I read about this demonstration when it came out in 2009. It was the headline, Steorn's Perpetual Motion Machine, Battery Included, which is like summed it up so perfectly. From that article, CEO Sean McCarthy claims that the battery is recharged from the device and that there's a three to one power ratio. Okay, but without any form of meter on the device, the simplest possible thing to include, and with no details of the actual power requirement, it is impossible to say what is going on here. Okay, so that demonstration from a publicity standard, at least, totally failed. From there, it was pretty much downhill for Steorn. Investment money dried up year after year, but the company made one last burst of energy out to the public before going supernova, and this was in 2015. You guessed it, another display or fundraising event, if that's what you want to call it. Arranged around a new and improved device called the Orbo Power Cube. This was in May of 2015. And by the way, we covered all this as it was coming out with the blog and through the podcast and everything else. May 2015, Steorn put out the Orbo Power Cube behind a bar in a pub in Dublin. I'm sure you guys remember us talking about that. This power cube is a small box, which on the pub website claimed to contain a perpetual motion motor, which required no external power source. The cube was shown charging a mobile phone. Steorn claimed to be performing some basic field trials in undisclosed location, and this was one of them. How clever. And then by December of 2015, Steorn also announced they have not one, but two devices available for pre-order. The O-Cube, which supposedly is a sleeker name for the Orbo Power Cube back in May of that year. And it's a USB charging device that they're selling for 1200 euro. Or you can get the O-Phone, a cell phone that never has to be recharged. So this was the final attempt to bilk money out of gullible investors. But it was for naught. Soon afterwards, 2016, Steorn announced it was ceasing operations. They sacked the entire staff. They began the process of liquidating the assets. But today, 2024, there's an update on Steorn and their physics-breaking technology. Their liquidation process has finally been completed. A final creditors meeting was held last week. Statement shows the creditors had been due just under 344,000 euro, with 157,000 euro owed to potential creditors. It basically had no assets at the time, but they were able to release about 80,000 dollars of euros worth of assets. McCarthy's final quote, we didn't set out to do this. We stumbled upon it. I wouldn't believe a word of it if I wasn't working here. Basically McCarthy's words saying that I totally believe in what was going on here. So that was then. This is now. In a way, it's a bit of a sad day for the SGU because this is going to be the final update about this era in time with Steorn in a history where the equation E equals not MC squared, but E equals T, time, energy equals time. For a few hundred invested humans on a tiny insignificant rock in a world, their world is effectively over 2007 to 2024. Thanks for the great time, Steorn. You gave us a lot to talk about.
S: Yeah, basically spanned almost our entire career as the SGU.
E: I know. I know. It ran parallel with the skeptics.
S: And how predictable is the whole arc, right? Right from the get-go. This is what happens all the time. The demonstration's going to fail. They'll blame some technical thing that they have to fix, or they'll say it has to just scale up. There's some tiny little extra bit of energy in there that we cannot positively identify. But trust us, it's there. And once we scale it up, you'll see. And then, of course, never. It never happens because magic isn't real. Because the laws of physics are the laws of freaking physics. As I've said many times, I will bet on the laws of thermodynamics every freaking time.
E: 100% of the time.
S: So far, I've been right every time.
E: Place your bets.
S: All right. Thanks, Evan.
E: You got it.
News Items
ChatGPT 4o (15:42)
- Hello GPT-4o [4]
(Note: this article is not from the SGU show notes page)
S: Jay, you're going to try really hard not to overhype this, right? But tell us about the new ChatGPT.
J: Steve, it's probably the best thing that's ever happened to me in my life.
E: At least you personalized it and didn't extend it to the rest of the world.
J: So Steve and I did a little demo on the TikTok today. So because I'm a paid member of the ChatGPT-4 service, I gained early access to the 4.0, like the full treatment, pretty much. I mean, there's still lots of things that they're going to roll out over the next few weeks. But I was able to use one of the new functions today, which is essentially just talking to it, right? You can talk to it verbally, and it will talk to you back. Now, overall, I mean, I have a lot of things to tell you, but overall, my experience with it today was I thought it was good. I didn't really see the nuts and bolts underneath the hood because I didn't try to push it to any limits. I wasn't asking it to write any code for me or any of the more complicated tasks that it can do. So all I really do is have information from the website and from a couple of news articles about what the new functionality is.
B: Jay, this is called ChatGPT-4.0, and the O is for Omni, right?
S: Yeah, the letter-
B: The letter O, not the number O.
S: Not the number O.
B: Yeah, it's the letter O.
S: Not 4.0. It's 4.0.
J: O for Omni. Yeah, it's ridiculous. It's ridiculous. Horrible marketing. Nobody cares about the word Omni. 4.0 is like a very software-driven concept of iteration, and it sounds like it's 4.0, but there is already a 4.0 because there's already ChatGPT-3.5 and then 4. So I think they failed on the marketing. Stupid. I don't know. They could have called it something-
E: 4.00.
J: Evan, try to follow me on this. Instead of calling it GPT-4.0, they should have said 4 Omni or 4 whatever word they wanted. The word is irrelevant to most of us.
E: 4 plus.
C: 4.1.
J: Yeah, 4 point something, whatever.
B: 4 Turbo. There's so many iterations now. I mean, 4 Omni would have been fine.
J: It's confusing. The thing I don't like about using a word instead of a number is that you don't know what order they come in. In a year or two, Omni, Beta, Zooch, whatever they want to call it, we won't know. We just won't remember the order. So whatever. Let me get to some of the interesting stuff here. So the number one thing that Steve and I experienced today was the voice interaction. So I was talking to it, and it was talking back to me. There was a little bit of a delay that's supposedly going to get better as they ramp up over the next few weeks. It was also able, I uploaded a picture of the studio, and it was able to tell me with very strong precision what it was seeing in the picture. It was and me, Steve, and Ian all were like, whoa like, wow, okay. It identified that it was looking at a monitor of chat, of chat messages. It identified equipment in the studio. It saw a picture of a guy on the screen, and it thought it was one of us just because contextually it was trying to figure it out. It would make sense typically that one of us would be on the monitors, but in that instance Ian was showing us a different video.
C: But would it recognize a cat if it was partially obscured by a tree?
J: I don't know. I didn't test any of that.
C: But wasn't that used to be, why am I remembering that that used to be like a gold standard of computer vision?
J: Yeah, I mean, Cara, at this point, I would think absolutely it can do that from the demonstrations that I've seen. But look, we haven't really tested it. You know, my goal here is to essentially communicate what OpenAI is saying, what some of these expectations are. You know, I really don't have any strong reason to think everything is BS that they're writing because they have blown me away in the past. I mean, when ChatGPT came out, it was mind blowing to all of us. And again we keep saying these large language models, they're going to iterate very fast with significant improvements between each iteration. So this is the latest. So one major thing that it can do is that it can support something called function calling. This gives the developers who use it the ability to build applications that require ChatGPT to perform specific actions. Now these are questions that Steve and I were asking today that I found some answers to. Like through the developer side of things, like an outside company wants to use ChatGPT for Omni, they can actually build software where it does send emails and it does query databases and it can get things like online weather information and stuff and mash that information up in different ways, all depending on what the developer wants to do. Yeah, absolutely. So it essentially can connect with external applications. That's what they're saying here. Now here's a big improvement that they made.
S: Wait, wait, wait. Jay said it does not currently, however, connect to external apps.
J: I asked it if it can connect to my Gmail and it said no. But this functionality here is like backend functionality that a developer would be able to access.
S: Yeah, would be able to, but I don't think it currently, like, yeah, if we wanted it to currently connect to my calendar and update my calendar, it can't do that.
J: No, it can't do anything like that in the regular person's interface.
C: So you're not saying that it can't, you're saying that it won't.
J: It can do it.
C: It potentially can.
J: It only can do it through an API, which is something that a developer would use. It's like software talking to software.
S: Yeah, this functionality can be happened, can happen in the future. It doesn't currently exist.
J: Correct. So here's a big update. ChatGPT-4, the previous version, could handle 4,096 tokens. And just think of tokens as like discrete pieces of information. It could be a word, it could be a letter, whatever. The new version can do 128,000.
B: Nice.
J: So significant jump up. Now, what does this mean to us? This means that the new GPT is much, much, much more capable of handling extensive and detailed interactions with the users, right? So for example, you could have a really lengthy conversation with it and upload tons of documents that it's reading, digesting, and then you're talking to it and making modifications to. So whatever it could do in the past, it can handle much more volume, way, way more volume, like 4,000 compared to 128,000. You can't even compare those two numbers. The enhanced safety features. So they're saying the model includes improved safety mechanisms that are supposed to reduce harmful outputs. I did do some reading on this, and of course, there's always examples where people are allowed to kind of coax a large language model to do and say things that wasn't intended. But there's a whole module here that is supposed to be able to handle this and deal with these types of issues much better. That's great. God forbid we don't want a ChatGPT to tell someone to do something to hurt themselves or anything like that. So like, yeah, they have to have these safety mechanisms in place. It allows for the creation of custom GPTs that would be specific to users. Like for example, we have an SGU ChatGPT right now that we're working on. So you could search up keywords throughout our entire catalog and that type of thing. Now, this customizable capability that it has very specifically would allow you to like, you could develop this into an assistant, like a personal assistant. Someone's going to have to do it. It's probably going to come from a third party company that's using the ChatGPT backend in order to get there. But it's possible. And I think, of course, they'll be building on these features as well. I know that OpenAI understands that all of us want artificial intelligent personal assistants to do stuff for us. And that's a goal that many, many companies are working towards. Another cool thing is that they're making this available to everybody. You know, that's surprising. I'm a paid member right now. So what will happen is if you're not a paid member, you'll have a decreased operational window, meaning you can't, you won't be able to talk to it for very long. It'll limit the amount of basically pings that you have to the server. It's not clear how it's going to work yet, what the limiting factors are going to be. It's not clear, like as a paid member, how much more usage I'll have than non-paid members. But they want everyone to use it. And I think it's very, very smart. From a research perspective and a marketing perspective, they want millions and millions of people banging on this because the more it's used, the better the tool will become.
B: The demo I saw, there was a verbal conversation happening back and forth between a person and this Omni.
J: Yeah.
B: It was a big improvement. I mean, they were throwing lots of different like inflections and pauses and even little giggles into the conversation that really made it sound much more realistic. And the delay was basically negligible.
J: Yeah. Our delay today was a little too long. Again it's just rolled out on Monday they were using it in a pristine environment during the demo to show like what it will be like when it's fully operational, which is fine. You know, of course things ramp up. So I expect that the delay, they did say that the delay will be like talking to another person. So it won't be like this three second pause and then it'll talk. It'll be very conversational.
S: Yeah. My impression from our, not the demonstration, because the demonstrations are like always best case scenario, but our like real life interaction with it, it was in the Uncanny Valley.
B: Was it really?
S: So I was joking with Jay because he was going through all the different voices and it's like, first of all, it's like the cast of Up With People, right? It's like weirdly chipper, like glandly chipper, you know?
J: Too happy. Just a little too positive. A little too happy.
E: A little Ned Flanders-y.
B: They'll tweak that.
S: Strangely devoid of any personality.
C: But that's a good sign, right, Steve? Because previous to that, we weren't in the Uncanny Valley. It was clearly a bot.
S: I agree. I think it's definitely, it's an incremental advance, I think in terms of that. So the natural language conversation was good. The visual functionality is a good increment. But it's not quite a game changer yet. I could see them like the pieces are coming into place bit by bit. When we get to the point where it can interact with a lot of apps and it could, you know what I mean? When it can actually be our assistant, where we can say, hey, book these airline tickets for me. Or give me an example. Give me some options on airline tickets with these parameters. Or update my calendar with this information. Or go through my email for me and look for anything important or get rid of any spam. Or just look through my photos and pull out any pictures of Bob from 2020, whatever. When we could just give it natural language commands like that and it could do tedious tasks that take them off your plate, that's going to be a game changer. I think we're getting very close to that, but we're not there yet with this. But you could see how they're building towards that.
J: Absolutely. I mean, look, they're expanding it out to more languages. It can do 50 different languages now. It's twice as fast and half the cost of GPT-4. So you know, they're really making it very accessible, which is wonderful. Their claim, now here's some claims that they made, which are untested. I'm not going to like weigh in at all. I'm just telling you what they're saying on the website. They're saying that the new model can perceive emotions. I did see a demonstration and it perceived emotions. You know, it said the guy was making facial gestures and it was able to identify them. They supposedly have completely fixed the math problem. So now it can do math correctly. It can code much better than it normally did and be more extensive in its coding. And of course they're opening it off with enterprise tools and targeted business applications and all that. All that's coming. Steve and I had a deeper talk today as well, just about artificial intelligence and like what's happening like what's to come and all that. And my big takeaway here is I'm excited about this kind of technology. Of course, I want it to be safe. I want not expecting this, but I would really appreciate it if governments around the world would align and get into a place where they're helping put the development on rails to make it so we don't develop some really crazy evil software that will do bad things. I mean, I'm sure there's going to be people doing it anyway outside of the big companies, but we do need to figure out how to deal with what it's going to bring. And the one thing I'm sure about is that we are in the beginning of a steeper and steeper curve up in advancement here. We're going to see rollout of AI from different companies happening very frequently with big improvements. And I don't think we're ready for this. And I think us not being ready it can manifest in lots of different ways. I just think it's a good thing for all of us to learn about, get familiar with it, use it, and understand it better than you do right now so we can have more intelligent conversations about how to use it safely.
B: Yeah. It's amazing to think that, I mean, when did these large language models even enter into our ken? When did we first hit our perception, like with ChatGPT, like three, right? Or 3.5 is when we started talking about it. And that wasn't that long ago. I mean, what was that, two years?
S: Yeah, a couple of years.
B: And look where we are now.
S: In a steep curve.
B: That's a huge leap. Now, of course, there's no guarantee that we're, oh, we're now reaching the steeper part of the curve and it's going to accelerate faster and faster. Maybe, maybe not. Who really knows the pace, the acceleration is going to continue or increase? We really don't know now. You know, we'll know when we can look back on it and be like, oh, yeah, yeah, it happened. But now it's hard to say. But man, what I just think 10 years from now, even five years from now, thinking what could potentially be available, it's like, holy crap, this is going to be a ride.
J: Yeah, Bob a good indicator is just how much money are companies pouring into it and how hot is the competition. And both of those, the answers to both of those are extreme. There's an incredible amount of money being put into this. There's an incredible amount of competition happening globally. And you know, you know what happens when that happens. Like I can think of other things where countries are racing to achieve the same thing. You know what I mean?
B: Yep. But the other side of that coin, Jay, and we've seen in the past with AI, is that you've got all this this hype. And of course, it's more than just hype at this point. But you do have this hype that spurs investment. And then the results don't come as you expect. And then there's a chill where they call it the AI winter back in the 60s. And it caused a chill, which eventually thawed and started up again. But it's hard to say for sure. But man, it is looking pretty amazing at this point. And I just can't imagine a huge slackening, although it's not impossible. But man, I try to even just...
J: It's possible, of course, Bob. But we have to look at this. The right way, I think, to handle this is just assume that it's going to move very quickly because I think that is likely. And we should be we should try to get it try to get as much even or ahead of this as we possibly can.
S: All right. Thanks, Jay.
J: You got it.
2023 hottest summer in 2000 years (31:06)
S: Cara. How hot is it getting?
B: Oh, boy.
J: It's hot, baby. Yeah.
B: Hot, hot, hot, hot, hot.
E: In Kelvin, please.
C: I don't have that data for you. So there's a new study that was published yesterday, yesterday as of this recording, in Nature. And the headline gives away a lot, but actually it doesn't give away all of the pretty kind of intense and somewhat controversial claims that they're making in the study. But the headline is, 2023 Summer Warmth Unparalleled Over the Past 2,000 Years.
B: Yep.
C: So the claim here is that 2023, the summer of 2023, which was unseasonably warm, is potentially the warmest recorded average temperature that we've seen in the last 2,000 years. So in order to see what these researchers did to get to this claim, we have to understand a little bit about what we already knew and their sort of new approach to looking at the data. What we probably started to see last summer were news articles, and I'm sure we covered it on the show, news articles saying that 2023 was the hottest summer in recent memory. It was the hottest summer since we started having accurate recordings. It was the hottest summer since pre-industrial levels. But a lot of kind of outlets were actually reporting this was the hottest summer in the last 120 or 140,000 years. And these researchers were like, how would anybody know that? And so they realized that when looking at like core, I think ice core data or sedimentary data, the resolution isn't good. Like your error bars are around 300 years. And so it's a pretty big claim to say this year was the hottest compared to all of these records when we can't really dial into the year, definitely can't dial into the months of some of those kind of less specific forms of looking at climate that long ago. So they said, we want to kind of split the difference. We know that starting in 1850, we have records, and those records are based on instruments. Now, they found that there were some problems with that, and we will come back to that. But what these researchers did is they said, we're going to look at global temperature and actually Northern Hemisphere temperatures, what they settled on, using, what do you guys think?
S: Tree cores.
C: Tree cores. Yeah. 2,000 years, tree cores. So sometimes they were able...
B: Dendocrinology.
C: Dendocrinology. Yeah. And so sometimes they were looking at individual cores that were taking from living trees. Sometimes they were looking at felled trees and were able to count the rings based on that. But I think it was something like 10,000 trees across North America that basically put together a pre-existing, well-supported analysis. So the individuals who published this study, which is three different researchers, they are not the ones who analyzed 10,000 different tree cores. They used an analysis that had already been published that was like a multi-university, multi-lab, multi-institution study. And they looked at that data and they started to compare, well, okay, what were things looking like in 2023 between I think like June and September or June and August? And how do we compare that to that same period across what we consider to be generally the industrial period is usually between 1850 and 1900. That's sort of the beginning. That's what we're often comparing. Like the Paris Climate Accord of 1.5 degrees Celsius is based on original observations between 1850 forward. And so they looked at both that, this sort of recent history, and then they said, okay, we want to go way back to the year one. And we want to compare back to the year one until now. So they found some kind of interesting things. The first one is the claim made by the headline of the article and by most of the write-ups that I found, which is that no other year has had a summer this hot. Some of them have come close, but no other year has been quite this hot. We're talking by a difference of 2.2 degrees Celsius. So the average temperature from June through August in 2023 was 2.2 degrees Celsius warmer than the average summer temperature between the years one and 1890.
B: That doesn't sound like much.
C: That's a lot. And it was also 2.07 degrees Celsius warmer than the average summer temperature between 1850 and 1900. But here's where things get a little bit interesting. That is probably not the number that you have heard before, because 2.07 Celsius degrees is well past the 1.5 degrees Celsius that the Paris Climate Accord calls for us to reduce warming to, right? It's always compared to, quote, pre-industrial levels. And that's based on these instrument readings around that time. But turns out there really weren't that many functional instruments. They were all mostly in Europe. And comparing those records, those actual measurement records, to the data pulled from these tree cores, these researchers make the claim that it was actually colder then. We had a warm bias in our data. And so they believe that last summer was 2.70 degrees Celsius warmer than pre-industrial levels, than that period around 1850. So we already blew past that number, which is disconcerting. So those were the two big takeaways that I saw of the study well beyond over the past 2,000 years. They go into why things were sometimes cooler and sometimes warmer historically. But there were a lot of cold periods over the last 2,000 years.
B: Yeah, mini ice age.
C: Yeah, mini ice age eruptions, a lot of big volcanic eruptions. And when a volcano, a rarely large volcano erupts on land, sulphur dioxide aerosols go into the atmosphere, and it actually cools the Earth. But 2023 was weird for a number of reasons. Not only is it trending warmer because of climate change, and it's following a very kind of appreciable trend, but also it was an unseasonable El Nino year, and apparently there was a below-the-ocean volcanic eruption. And when that happens, water vapor ends up in the atmosphere, which actually has a greenhouse effect. So it's warming when a volcano erupts below the ocean surface as opposed to above. And all of those things kind of coalesce to make 2023 really, really hot. A lot of researchers, though, who are sort of interviewed who weren't involved in the research across multiple different news outlets that I've seen coverage on, they're saying that even if this is true, it's not really terribly meaningful because we already know that. This was to be expected within the next several years. Maybe it's a little fluky that 2023 was the year where we got there, but all of the trends support this. So this has a very good face validity. And some people are questioning, are the tree core numbers more accurate than the instrumentation numbers? Because we're talking about a difference of about half a degree. Based on the instrumentation numbers, I think that we've gone to 1.48 degrees Celsius above pre-industrial levels in the summer of 2023. And of course, based on the tree core estimates, like I said, 2.07 degrees Celsius above. So blowing past that 1.5. But either way, we were pretty dang close to it. So ultimately, this is an interesting question, how cold or hot was it in the past compared to how hot it is now? But it doesn't change the fact that it's really damn hot right now, and that all of the kind of knock-on effects of climate change are actively being observed and continue to be modeled. But very interesting and worrisome, but maybe not newly worrisome information.
S: Yeah. I mean, it shouldn't come as a surprise. It isn't a game-changer. It's like, yeah, we knew that 2023 was unusually hot. It was kind of a perfect storm because it was obviously the background trend of global warming and the El Nino. And so that's not surprising, but it does reinforce what we already know to be true.
C: Yeah. And it kind of gives you a sense of scale. 2,000 years. Wow.
S: Right. 2,000 years. That is interesting.
C: Yeah. We're talking... What was happening 2,000 years ago? Is that Roman Empire?
S: 2,000 years or something? I don't know.
C: You have that too.
E: Oh, yeah. Literally.
C: Definitely pre-industrial or mostly mechanical things at that point.
E: Yeah. Yeah. It was creeping around at that time. That's right.
S: That sort of thing. All right. Thanks, Cara.
C: Thanks.
Spotting Misinformation (40:33)
S: So I saw a good article on the BBC News on spotting misinformation. I thought I would talk about that. So a recent survey, you guys answered me this question. What percentage of people answered the survey, do you verify information you encounter in the news and on social media?
B: 4 foot 1.
C: Oof.
J: I would say 5%.
E: Half.
C: I'd say 25% to 30%, but that's probably an overestimate because people lie on surveys.
S: The survey found that 90% of people said that they do. I don't believe that for a second.
E: Yeah. Right. I don't believe that for a second.
S: That's what they found. Then if you ask, the same survey asked, should we be limiting the spread of misinformation online? And I'll just tell you, 96% of people said, yes, we should absolutely be limiting the spread of misinformation online.
C: Of course.
S: Which seems like a no-brainer. 74% of people report saying that they've seen news labeled as false online. Only 60% of people say that they regularly check information before sharing it, which doesn't quite fit the other number. So I think maybe 90% of people may think that they have ever verified information online, but not always or regularly. But only 60% of people say that they regularly check information before sharing it.
C: And even that is probably a really fattened up number.
E: It sounds inflated.
C: People love to say that they do pro-social things when they don't actually do them.
E: That's why polls are always-
S: Yeah. That's what I'm saying. This is how people answer the question.
E: Take a grain of salt.
S: And we also know from forensic analysis that a disproportionately small number of people spread a massive amount of the misinformation, like 15% of people share 60% of the misinformation. And we know with specific topics, I remember the Dirty Dozen, these 12 people are spreading more than half of all the anti-vax misinformation online. There's also, I also was looking for information on where people get their news from. This is a complicated question. But if you ask people their party identification and what their trusted news sources are, Democrats and Republicans have their list of five most trusted news sources with zero overlap.
B: Wow.
E: No cross-pollinization.
S: The top five for Democrats are completely different than the top five for Republicans.
J: Oh, I got you. Okay.
C: And is this through all media or just online?
S: This is all media.
E: That is not at all surprising to me.
S: It's not surprising. Other things are fine. But that's-
C That's depressing.
S: But that's more, I think, than what we've been seeing in the past. And yeah, so it's like we are, yes, we are living in different information ecosystems. We are getting completely different sources of news. The article mainly was about the SIFT method, which is an acronym for how to spot misinformation and prevent the spread of information. This was put together by a media expert, a digital literacy expert called Mike Caulfield. But I thought that I could expand upon it.
E: Do you have an acronym?
S: I do. And my acronym is SKEPTIC.
E: Cool.
B: No way.
J: Cool.
E: Clever.
S: So some of it, there's some overlap with SIFT, but SKEPTIC is my approach. Okay. So here they are. S stands for stop. So don't spread that claim. Think about it before spreading it. Don't retweet it. Don't post it. And don't make a TikTok video where you're pointing to the person making the claim and nodding. Please do not do that.
E: Wait, wait, that never happens.
S: Don't do any of that until you do the rest of the steps, right? So stop. First thing you do is just stop. Don't spread that. And then-
E: Don't let your first impulse get the better of you.
B: I would have sworn that that S would have stood for Steve, but okay.
S: Yeah. That's an alternate. That was my second choice. All right. The K is for keep an open mind. So consider all possibilities for plausibility and evidence, even that the claim is not true or that what you want to believe may be wrong. Right?
E: Perfect.
S: Okay. The E is for evaluate the source. Examine the source for authoritativeness, potential bias, and expertise. What other claims does that source make? And what do other sources have to say about it? P, put the claim into context. Is there a controversy? Is this a politically hot topic? Is the person making the claim selling something such as a political perspective? T is for track, and this was the same T in SIFT. Track back all sources to their original source. We've said this many times.
E: That does help.
C: That helps a lot, but it's a lot of work.
S: It can be a lot of work.
C: It's hard to do.
E: Yeah. It's extra effort.
C: It's 100% should be required of all journalists.
S: Yes. Absolutely.
E: At the least.
J: Steve, should we trademark this? What do you want to do?
S: Well, I'm not done yet. Let me finish. And-
C: Jay's like, I don't have any money on this.
E: Jay. Let him finish. Let him finish the pitch.
S: Now, there's a corollary to that, and that is rely on primary sources whenever possible. What is a primary source? A primary source may be, for example, a peer-reviewed published study. The study itself is a primary source. Somebody else talking about that study is a secondary source, right? But a lot of news reporting, mainstream reporting, is secondary tertiary sources. It's somebody-
C: Yeah. Very often, we are tertiary on the SGU.
S: Yeah. We're often secondary and tertiary sources. If we're interviewing somebody, then we're going to the primary source. We become a secondary source. If we're reporting on somebody else's reporting, we're a tertiary source. But I always try, at least again, track back to that primary source. And like I almost always, like unless it's just not available because it hasn't been published yet or whatever, read the original study as much as you can because it's amazing how different they are than what is being reported about them often.
C: Well, even like the study that I just cited, very few people in their write-ups talked about this change in how we're comparing pre-industrial levels. They just kind of reinforced the headline.
S: Right. Yeah.
C: And it's like, oh, but there's a whole other lead you're burying.
S: Don't believe headlines. That's for sure.
C: Yeah.
S: All right. I is for investigate. Do background research on the science involved, what the experts say, and why, and who disagrees with the claim and why. Right? So investigate. The C is consider before concluding. Consider all this information when making a determination if the claim is likely to be true or not. It's okay to conclude that you don't know or to wait for more information. And remember to be humble and charitable. So that's it. Skeptic. Stop. Keep an open mind. Evaluate the source. Put the claim in context. Track back to original sources. Investigate the background science. And consider all of that before reaching a conclusion.
B: Nice. I think you need more letters though.
S: Why?
B: You should add an S and then G-U-I-D-E.
E: And then a T-O.
S: All right. We'll work on that. We'll fluff this out. We'll pad it out. So that's it. That's my quick guide to spotting misinformation, dealing with misinformation online, skeptic.
B: Nice.
S: You like it?
B: I like it.
S: All right.
E: I like that. More than SIFT. I like it.
S: SIFT is fine.
C: SIFT is good too.
E: SIFT is fine. Yeah. I know. I'm biased. No doubt.
S: There's a few other things in there. So the SIFT was basically stop, investigate, find other sources, and trace back to their source, the original one. It's fine. But it wasn't, it was not as, I think, as thorough.
E: Yeah. I mean, that's a little Newtonian. Whereas yours is Einsteinian.
C: Stop.
E: How's that? Is that a fair corollary?
AI Training Robots (48:38)
S: All right, Bob, tell us about AI training robots.
B: I really like this. I love this. NVIDIA in the news again, but not in the way you might think. The new tool is called DrEureka, and it has shown that it can leverage the power of large language models to quickly train robots in simulated environments, and then transition that learning and behavior of the robots into the real world better than humans can. Another study is called DrEureka, Language Model Guided Sim-to-Real Transfer. These are, again, these are by scientists at NVIDIA and the University of Pennsylvania and University of Texas. Okay. Austin. Okay. So now, who has seen that video of the guy walking a robotic dog on a leash while that dog is also balancing on a big blue yoga ball?
E: I'm watching it right now.
B: Inflated yoga ball. The robotic dog, it's one of those smaller ones, kind of like the ones from Boston Dynamics. Now, the robot dog is not only balancing on the ball, it's moving the ball at the same time to walk along with its human master, at least its master for now. It can even do this facing sideways and on different terrains, even when the ball is being kicked by a human, or even if the robot dog is on a partially deflated ball, it could still balance and move in any direction that it wants. It is an amazing feat, no matter who or what is doing it. It's quite a complicated feat. So how did they pull this off? So first off, it's not uncommon to train a new skill for robots virtually, entirely in a simulated environment. It makes sense, right? You could just create a specific environment you need, and then you also upload all the stats on the robot, and you have it interact with the environment and learn these different skills that you can then transition to the real world, right? It's been done for years. Now, one way to teach uses what's called reinforcement learning, which we've talked about on the show before, but reinforcement learning also has a critical component that's called a reward function. This reward function defines a specific goal and essentially gives a number grade on how good the performance was. It's like giving a meatball for a reward. So Jay accomplished a task very fast and very efficiently. So he gets one meatball.
E: He's not the only one who likes meatballs in this group, by the way.
B: I know. For task B, though, Jay only gets a half a meatball because-
J: I already don't like this.
B: -he gets half a meatball because he stepped on one of his kids to accomplish it. The objective is to get a better and better reward function number because that means you are accomplishing your goal better, okay, essentially. So the problem is, though, how do you know whether to give Jay one meatball or two or a tenth of a meatball? It's very difficult for human experts.
E: Ooh, a "deci-mated" meatball.
B: It's hard to determine what is worthy and how do you code that, and it takes some time and real skill for human experts to craft a complex and an effective reward function, a lot of time to do that. So you would say, perhaps, that the humans are the bottleneck in this process. So once the virtual learning for a robotic system is complete, then there's another big hurdle, and that's the transition from virtual to real, right, because you may develop these behaviors and skills for the robot in the simulation, but then that code has to transition to the real world. So this is called the sim-to-real gap, the sim-to-real gap, and it's massively difficult. Humans have to, again, they have to step in and laboriously tweak the code that worked very well in a clean, idealized virtual world, but it's not yet optimized for a real, messy, unpredictable world. So it takes effort to make that transition, to transition the reward functions into a form that will work in the real world. So this is where Eureka and its successor, DrEureka, steps in. Eureka was introduced last October 2023, I believe, and it's essentially NVIDIA's virtual raining environment. It minimizes the need for human input, and it leverages large language models like ChatGPT4 to write the entire reward function with no humans really in that loop at all. So the LLMs, the AI, if you will, write the reward function, which is then run in the sim, and then the results, the codes come back to see how well the reward function number comes back, and the results are kind of mulled over by the LLM in a sense, and then it iteratively improves it so that the reward function gets better and better and better, the number gets higher and higher. The real advantage comes into play, as you might imagine, when it can run, the system can run hundreds of reward functions at the same time and then pick the best to improve those. So just taking advantage of the digital nature of all this entire process. Now Eureka was and is quite successful and very fast at training robots in this simulated world, but it's still virtual though. This was still virtual at this point. The sim-to-real gap, as I said previously, is still there, and this is where DrEureka comes into play. Like Eureka, DrEureka uses large language models again, instead of slow humans, to tweak the reward function code and simulation parameters so that it works in the real world. But not only that, the other addition to DrEureka over and above Eureka itself is that there's safety instructions. There's a safety instructions function that's added as well so that reward functions are generated that are safe for people in the real world. And basically, take that Skynet, which is really, it doesn't matter if it's not terribly safe virtually, but when you're updating the behavior of a robot in the real world, yeah, it's pretty kind of a no-brainer to make sure that the reward function incorporates a safety function for people that are going to be dealing with it. And also, for what it's worth, the initial letters doc-DR in DrEureka doesn't stand for doctor. It really stands for domain randomization. That means that the key innovation here and the key real difficult part of this, of creating DrEureka, is that it randomizes some of the elements of the virtual physical parameters so the system can generalize some of this unpredictable stuff that happens in the real world. So, adjusting these parameters requires not only knowledge of the robot in question, right, because you need to know the details about how the robot behaves and everything it can, but you also need to know the sort of common sense physical reasoning that large language models now have, right? You'd have to know how the physical world works. You have to know about friction. You have to know about gravity and all that kind of stuff. And large language models do have a decent understanding, if you will, of those concepts in order to pull this off. So how good is Eureka and DrEureka? In their paper, the researchers say, Eureka can generate superhuman level reward functions across 29 tasks. Eureka rewards outperform expert human written ones on 83% of them with an average normalized improvement of 52%. So this is significantly better than what humans have been able to do in terms of crafting these reward functions for the simulation and for this sim-to-real gap, transitioning it from the simulation to the real world.
C: How demoralizing, Bob, to be a grad student in that lab where every day you're just trying to play beat the robot and you're like, I lost again.
B: I know, man. I know. And you have to go on to other fields that they can't quite handle yet. But there's still a role. There's absolutely still a role for people in this, as of now, anyway. So how successful was DrEureka itself in creating the real world code? Well, I couldn't find any solid numbers yet, but one way to quantify that is that the dog robot successfully balanced on that ball and moved around successfully the very first time it tried to do it with no additional tweaking.
E: It's encouraging.
B: So basically, the robot dog graduated from the sim. It had the reward functions and other code uploaded from the simulation, and immediately, I could balance. I know Kung Fu, right? It's I know Kung Fu right there. It knew how to do the balancing, and it didn't really need any other additional tweaking, which is kind of extraordinary. So that's pretty cool. What does the future hold? Who the hell knows? One thing, though, over the short term, we're likely going to see two key improvements with DrEureka. Failures happen, right? The dog does fall from that ball from time to time. It's not perfect. So what they want to potentially do in the near future is to take these real world execution failures and feed them back into the system, right? To improve this.
C: Yeah, I mean, you think that's a good thing, that it fails.
B: Right? It seems obvious. To improve the whole sim to real transition by feeding the real world failures, I think it's obvious and it's a great idea. Also, right now, all the data that's flowing from the robot itself is proprioceptive, right? The orientation and position of the robot's limbs. And this comes from the internal sensors. In the future, they are probably going to incorporate vision, right? Data from vision and other sensors as well to have even greater real world performance. Over the longer term, it seems pretty obvious to me that robot training will increasingly take advantage of advances in AI, like large language models and other techniques that maybe haven't even been developed yet. Who knows what the training will be like once GPT-5 is available and incorporated into DrEureka. If they even do it, I would think that the training would be even more successful. Some people even think that robotics itself over the next few years is going to see its own, as they call it, their own ChatGPT moment, right? That's how they describe it. You know, I think as these latest advances in AIs, especially large language models and vision models, more fully integrate into robotics, we're going to see some amazing stuff. I mean, we're already seeing it now, if you follow robotics news, we're already seeing some really extraordinary stuff like robots processing natural language and doing tasks that need some sort of complex reasoning that wouldn't have even been possible even a year ago or less. So we're already seeing some really interesting things happening in the robotics world. And I think over the next few years, in 5 or 10 years, we're going to see some extraordinary stuff happening as robotics and AI kind of gets together.
E: When does Jay get his meatball?
J: Let's just cut to the chase, Bob. All that's nonsense. Where's the meatball?
B: I'd love to assess the reward function.
E: RF.
S: All right. Thanks, Bob.
Reincarnated Son of Buddha (59:59)
S: Evan, tell us about the reincarnated son of Buddha.
E: Oh, yeah. Happened across this one the other day, a source of news that I don't recall having referenced before, the Bangkok Post. The Bangkok Post. Yep. It's from their late edition of the newspaper. We can call it the one night in Bangkok.
S: It's called the Bay Po for short.
E: I like that. Yeah. Headline reads complaint against reincarnated son of Lord Buddha and reincarnated son is in little quote marks, an eight-year-old boy is facing a legal complaint over claims that he is the reincarnated son of the Lord Buddha and possesses magical powers to "connect people's minds". Cool. Hive mind. That's some powers right there. Right. Cara, for your consideration, that's a Star Trek. Please not to be confused with my Star Wars reference earlier. We're good.
C: Yep.
E: Magical power to connect people's minds. That's interesting. This complaint was filed with the Central Investigation Bureau against Nong Nice. That's the boy's name, along with eight other individuals who manage Nong Nice's Mind Connection Office and his website. This is for being in violation of the Computer Crime Act, the Donation Soliciting Act, and Child Welfare Protection Act. Okay. So who is Nong Nice? Who is this eight-year-old boy? Recently and I can't find out, I couldn't find out exactly when this started, but I think it was either last year, perhaps the year before. So rather recently, Nong Nice's parents made a claim asserting that their son is the reincarnation of the Lord Buddha himself. For reference, Lord Buddha, the founder of the world religion of Buddhism over 2,600 years ago, revered by most Buddhist schools as a savior, the enlightened one who rediscovered an ancient path to release clinging and craving and escape the cycle of birth and rebirth. Regarding Nong Nice, their parents had some pretty good influential skills apparently because it did not take long before the word of Nong Nice's extraordinary abilities began to spread throughout Thailand. It was said that this child prodigy could connect telepathically with his followers, guiding in them through the intricacies of meditation with wisdom far beyond his years. Courses and forums purportedly led by Nong Nice became the new pilgrimage sites for those seeking enlightenment, drawing crowds eager to bask in the presence of this young luminary. That was from the article. So yeah, and it's men, women, all ages, they are coming to Nong Nice, his organization basically, offering him money in order to have them blessed and also to seek how to channel their inner psychic powers, which he can do for you via live streaming into a person's head. So basically you have an outreach now using modern technology for this particular purpose. So you remember it's the time when we were talking about homeopathy once with James Randier, he was giving a talk or an explanation on it. And it wasn't so much even that the homeopath now can just mix up the concoction in their lab or something and hand it out. You could hook up your computer to the internet, to a glass of water, and the homeopath would be able to transmit the remedy through the computer, through the internet, and wind up in your glass of water, wherever you are in the world. So that reminded me of this, in a sense, right? You don't need to be in the class with him. You can just broadcast live and there you go, it'll magically happen. Let's see. According to his group of handlers, he's an avatar of the great Naga, which is a mythical serpent who became a key disciple of the Lord Buddha, so the legend goes. And he is back here on earth to preach and spread Buddhism, not just to people in Thailand, but especially apparently to Russians so that they can use their wisdom to stop, help cease the wars that Russia has waged against Ukraine. It's not surprisingly, soon after this kind of took on a viral sensation, questions began to arise about the authenticity of all this, just the paranormal powers and what this organization was really all about, enough to catch the attention of the National Office of Buddhism, the NOB. This is an official government agency, apparently, who alongside the Social Development and Human Security Ministry have been tasked, along with the blessing from the Office of the Prime Minister of Thailand, to investigate the claims and what is going on here. This was back in December of 2023 that those orders came down, however, now it's May 2024. I'm unable to determine if the investigation is still ongoing, but some people are claiming to be victims of a fraud here. One person who says they attended the course on Mind Connection and learned very quickly that this was entirely a sham, and there are apparently several other people who are claiming to be victims of this scam. So essentially, the complaint is coming from critics, all of them Buddhists, who are concerned that the claims surrounding the nature and power of Nong Nice are a distortion of Buddhism. You know, it's not like the complaint came from a Consumer Protection Bureau or child welfare advocates or any scientific or secular academy capacitors in a station like that. It came from professional Buddhists. All right. So yeah, it's all worthy investigation. I would argue primarily for the health and welfare of this eight-year-old child, primarily, but then certainly there needs to be more insight as to whether or not this is truly a monetary scam and are these people who are willingly giving their money unwitting victims of a larger scam here. So that's what's going on in Thailand. Nothing conclusive. This claim was officially logged. So that's why it's getting the news attention.
S: Is there evidence that this eight-year-old child is being exploited by adults or do they really believe their claims?
E: Unknown at this time. There are, but there is one of the agencies is looking into that aspect of this. But again, they started the investigation or they announced they were going to be investigating in December of this past year. And I could not find anything online where there's been any conclusion or an announcement or a bulletin, a brief or anything made by any of the investigatory bodies saying anything. So it apparently is still ongoing.
S: But unfortunately, it seems like, as you said, the investigation and the consumer protection here is motivated by defending a religious belief. And if the scam were something that was not directly related to a mainstream religion, it'd be fine. You know, it seems like, you know what I mean? If you said he's psychic rather than he's the son of Buddha, then the Buddhists wouldn't care and nobody would care.
E: Correct. Yeah. Right. It muddies those waters, unfortunately in this particular case. But let's see. Let's see if there is abuse happening here of a person. And if people are, if an actual scam, an operation is underway to defraud people from their money. So we will have to wait and see.
S: All right, thanks Evan.
E: Yep.
Who's That Noisy? (1:07:56)
S: Jay. It's who's that noisy time.
J: All right, guys. Last week I played this Noisy:
[Baby animal grunting]
It's amazing how much information can be communicated in just a few little squeaks, you know. I find this one to be adorable and very funny. Do you guys want to guess before I start going through this?
S: I mean, it sounds like some kind of baby animal.
E: Well, those are baby squeaks, is what they sound like.
J: All right. That's good. That's a good start.
B: But early in the file, it seemed to be some weird growls. To me, it sounded just like a possessed baby.
C: It sounded like a possessed robot baby to me. There's something like weirdly electronic about it.
J: Well, a listener named William Steele writes in and says, hi, Jay, I'll keep it quick. Baby sloth. Very cute, noisy. Good guess, but it is not a baby sloth.
E: I see a song, though, a popular song, baby sloth, sloth, sloth, sloth, baby sloth, sloth. Yeah.
J: Yeah. You're really up to date. That's great.
E: Thanks.
J: Next listener is Lisa A. And she said, would that be a tortoise basically mating with another tortoise? And lots of people guessed that one. Apparently, there's quite a few videos out there of box turtles doing this type of thing and they make funny noises. I think I actually used this a long time ago on the show. But no, that is not correct. But that is a very good guess because I did listen to a few videos and it's kind of similar. A listener named Scott Jones wrote in. He said, hi, been listening for years. This is my first email to you. I think that this week's sound was from the Army of Darkness movie. Bruce Campbell was being beaten up by smaller and squeaker versions of himself. I think the clip was from that scene. So, of course, I had to go listen. I haven't seen that movie in quite a long time. Yes, of course, it's really funny and there are squeaky noises in there. Not correct. But thank you so much for that funny guess. I had a first-time winner here. And many people guessed correctly. But this person sent it in first. This listener's name is Mary Fielder, F-I-E-D-L-E-R, Fielder, Fiedler, Fiedler, thank you. Cara, you laugh every time.
C: Fiedler.
J: Fiedler.
E: It's Italian.
C: He's like, Fiedler. You're like, oh, yeah.
J: I know, as soon as he says it, I'm like, all right, here we go. Hi, Jay. This is my first-time guessing. I think it is a Brazilian porcupine. I teach elementary art and we play a game where I play an animal sound to the kids and they draw the animal they think made the sound or just make up their own creature if they don't have a guess. One of their favorite sounds we did was a Brazilian porcupine. I'm pretty sure that's who your noisy is. Thank you. You're all awesome. That was a wonderful message to us. Thank you. And correct, so congratulations. Yeah, so this is actually-
S: Have you used a porcupine before?
J: We probably have over the years.
E: Yeah, but not a Brazilian.
S: Yeah, maybe not a Brazilian porcupine.
J: So this guy, this squeaker we got here is a very specific porcupine. This is a porcupine named Kemosabe, who I believe is no longer with us, but had a wonderful life and ate lots of great things. Yeah, very cute. You can look him up online. So here is Kemosabe when he was just a little guy, very small. It's funny when they're really small and you can identify what kind of animal it is, but they don't really look like a porcupine, that's where he is. Very small. Take a listen. [plays Noisy]
B: Wow.
J: So cute. Now, Bob, your daughter, my niece, actually interacts with a porcupine almost every day, if not every single day, because she's working in that animal reserve. She's working in that animal reserve and she sends us video of their porcupine who makes really, really awesome eating noises.
B: Oh, yeah. I was there last month and I fed her, I think his name is Twix.
J: Yeah, Twix. Very cool. Very cool noises. Very cute animal.
New Noisy (1:12:20)
J: All right. I have a new noisy for you guys this week. This next noisy is from a listener named Jared.
[Rapid mechanical whirring]
I love Who's That Noisy because it could be so many different things. I could hear a couple, if not three different things that-
E: It evokes a memory from my childhood is what it does.
J: Yes. I have the same thing that you're thinking of.
S: The robot with the flashing eyes?
E: I was thinking the handgun. Remember that? The space ray gun kind of looked like a little spaceship in your hand.
J: I'm sure a lot of people are going to be sending that in.
S: Same tech. Yeah.
J: All right. If you guys think you know the sound this week or if you heard something cool, you can email me at WTN@theskepticsguide.org.
Announcements (1:13:07)
J: Steve, 1,000 is a wonderfully large and round number.
S: Yes, that is true.
J: We will be interacting with this number very soon on this podcast because we are in our 20th year and our 1,000th SGU episode. What am I talking about? In Chicago, we are putting on three shows. Two of them are available to you at this point. There is a mid-afternoon extravaganza that will be happening just on the outside of Chicago that will be on August 17th. That's Saturday, August 17th. If you're interested, you can go to the SGU website. That's theskepticsguide.org. You can check out tickets that are available for that show. If you're also interested, we are going to have a private 1,000th episode, live episode. All of us will be there. We will have tons of virtual guests. George Hrab will be in the room with us. If you're interested, you could also go to theskepticsguide.org. There's a button there for you to purchase those tickets. Tickets are actually getting close to being sold out. Sales are very nice because I think people realize that this five-hour show is going to be totally worth it. It really is. We're going to have a ton of fun. We have lots of things planned. All the more better is we will have a ton of review of some of the funniest and most interesting moments of the podcast over the last 20 years. Guys, 20 years.
E: Jeez. It went fast. I don't know what else to say. It was a fast 20 years.
J: I still clearly remember us playing City of Heroes and that turned into the podcast. That's what we were doing that night.
E: I know.
J: We had a weeknight, a weekly date to play City of Heroes and Steve's like, we're going to quit the game and we're going to create a podcast.
S: Yeah.
J: That's where it started.
S: Sort of. I mean, the game blew up for other independent reasons and rather than trying to read something, I'm like, screw it. Forget the game. Let's do a podcast.
J: I would like to go back and play the game though. Anyway, so-
E: We still can.
J: If you'd like to join us, all you got to do is go to theskepticsguide.org and check out a place where you could buy some tickets to come to our 1000th episode.
S: Thank you, Jay.
Questions/Emails/Corrections/Follow-ups (1:15:18)
Email #1: Protecting Bigfoot
S: One somewhat quick email. This comes from Kristen in Kirkland, Washington, and she writes, "Hi, long time listener/patron. I just finished reading this article. She gets a link. Olympic Peninsula fifth graders get lessened in lobbying with saving Sasquatch and I am appalled on so many fronts. We have let these students down. The teacher supposedly teaching critical thinking, the lawmakers accepting the kid's proposal of nonsense, and the Seattle Times for publishing it. We'd love to hear your thoughts." So basically for several years apparently this teacher has had a project for her class where they come up with a cause and then they lobby the local government about that cause, right? So yeah, it sounds like a good lesson. If the topic that they have picked multiple years is saving Sasquatch, basically lobbying for laws making it illegal to hunt, trap, or kill Bigfoot slash Sasquatch. Yeah, so then I agree that this is disappointing. If the point was to teach critical thinking, it doesn't seem to have been a wild success.
E: Right.
C: Was that the point?
S: Well, that's part that's the buzzword, right? That's what they say.
C: Oh.
E: Well, critical thinking combined with civic action, according to this play out.
C: Yeah, that's what I thought. It was to learn how to lobby Congress.
S: Yeah. I mean, the local government.
E: Which means it's a lost opportunity, right? Because you could have targeted something real.
S: The critical thinking part is that is the debate phase where they debate whether or not they should do this, you know? And some of the students, to their credit, a minority, but they said they voted against it because they said Bigfoot is not real, which is a solid argument.
E: Yes.
C: It's good. I love it.
S: So here is a quote from the class talking to the commissioners. "What if Bigfoot is harmless? Bigfoot could just be another animal wanting to survive like everybody else. Sure, if we attack it, it would probably attack back. That's just self-defense in general. Any animal would do that, even a bunny."
E: Especially, yeah, especially the Holy Grail. Yeah.
S: So not really relevant to the question at hand.
C: How old are these kids?
S: This is...
E: 10 and 11 year olds.
S: They are fifth graders.
C: That's cute, though.
S: Yeah, in elementary school. I mean, yeah, but I think it would be better to get them on a project that's actually real.
C: No, I agree. But I do like that kid's insight that you don't want to poke wild animals because they will bite your face off.
S: Yeah, yeah.
C: That's a good lesson to learn as a child.
S: Talking about mountain lions, solid. I would agree with that.
E: Yeah. Right. But those unicorns, careful. They'll bite your face.
C: But you got to remember that some fifth graders still believe in Santa. This is confusing.
S: I know. I know.
C: When where's that line? I know we always talk about this.
S: But the imprimatur of some kind of official sanctioning of it is not, I think, productive.
C: I agree. But I think it's...
S: Potentially counterproductive.
C: I agree and, not I agree but. Where did you say they were? In Seattle?
S: I didn't say. I don't think.
E: Yeah.
C: Somewhere in the Pacific Northwest.
S: She's from Washington.
C: Yeah, from Washington.
E: Olympic Peninsula.
C: So, yeah. So, I think that the thing that's frustrating here is that there is such a cultural celebration. That's like asking kids who live near the Loch Ness to write to Scottish government and protect Nessie.
S: Yeah.
E: Yes, I agree with that.
C: It's everywhere in the Pacific Northwest. And it's, I think it's a wink and a nod. But maybe that part wasn't.
S: I'm sure, I'm sure on the part of the lawmakers, it's probably.
C: Yeah, and the lawmakers, of course.
S: Like, yeah, yeah, okay, for these cute kids, yeah, sure, we'll protect Sasquatch, sure. You know, but it doesn't necessarily mean that they believe in it. But-
C: Right. But the problem is that's not explicit to the children.
S: Yeah. You know, and there have been efforts to pass, like, no-kill laws for Bigfoot by adults. Like, seriously. Like, not as an elementary school project.
C: It's so embarrassing.
E: Yeah, and Chupacabra preserves too areas of the wilderness preserved.
S: Right.
E: I made that up.
S: So I think, I mean, it's okay to lead debate among students about whether or not something is real. I remember when I was in elementary school, we had a debate about whether UFOs were aliens.
C: I think that's actually a great critical thinking exercise.
E: Okay, but that's productive.
S: It could be. It could be, except that the teacher didn't know what they were doing.
E: Oh. Fail.
S: It just increased everyone's belief in it, you know?
C: Right.
E: Yeah, you have to be careful. Know what you're doing if you're going to go down that road.
S: Exactly. So it's all, the devil's in the details, obviously. Like, were they really given accurate information to base their debate on or their decision on or not? Or was it very superficial? And as you say, putting the stamp of approval on an adorable local legend, you know?
C: Right, exactly.
S: It's like being in New Mexico and voting not to shoot aliens, you know?
C: Yeah.
S: It's the same kind of thing. At the end of the day, it's probably mostly harmless, but it's not a win for skepticism, obviously.
C: Yeah.
S: All right, guys, let's go on with science or fiction.
Science or Fiction (1:20:43)
Theme: Battery materials
Item #1: By weight the most prevalent element in current production lithium ion batteries is nickel, followed by cobalt, and then lithium.[8]
Item #2: A recent analysis concludes that the world will not be able to mine enough copper to make the green transition between now and 2050.[9]
Item #3: A new study finds that extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual US demand.[10]
| Answer | Item |
|---|---|
| Fiction | Lithium third-most prevalent |
| Science | Not enough copper to mine |
| Science | Lithium from shale wastewater |
| Host | Result |
|---|---|
| Steve | clever |
| Rogue | Guess |
|---|---|
Evan | Not enough copper to mine |
Bob | Lithium from shale wastewater |
Jay | Lithium third-most prevalent |
Cara | Lithium third-most prevalent |
Voice-over: It's time for Science or Fiction.
S: Each week, I come up with three science news items or facts, two real and one fake. Then I challenge my panelists to tell me which one is the fake. There is a theme, and the theme is battery material.
C: Oh, God.
S: Battery material.
J: Battery material?
S: Yeah. Stuff that batteries are made from.
E: Oh, stuff you beat people with, like clubs.
S: No.
E: That's battery material, too.
S: Lithium ion batteries, specifically.
C: I'm not going to do well on this one.
S: All right. Here we go. Item number one, by weight, the most prevalent element in current production lithium ion batteries is nickel, followed by cobalt, and then lithium. Item number two, a recent analysis concludes that the world will not be able to mine enough copper to make the green transition between now and 2050. Item number three, a new study finds that extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual U.S. demand. Evan, go first.
Evan's Response
E: Cara, when we talk about batteries and things, you just have to go with the flow. Is my mic on? I didn't hear.
C: I got it.
E: Bob, I didn't hear you laugh.
C: I don't think I have enough capacity.
E: Very good. And Steve used the word current also. I'll link that out. All right. Enough of that. Remain positive. By weight, the most prevalent element in current production lithium ion batteries is nickel, followed by cobalt, and then lithium. By weight. Okay. We're measuring this by weight. Oh, crumbs. What? I have to know my periodic table? Is that what this boils down to? Is this a periodic table test?
S: I mean, would this be easier if it were by volume?
E: Well, no. No, I suppose not. But, you know. Okay. Well, lithium is low on the chart, so that's probably the nickel, cobalt, lithium. All right. That one will be true. Then number two, the recent analysis concludes that the world will not be able to mine enough copper to make the green transition between now and 2050. Well, I mean, this one you could go either way with. Okay. Yeah. Or recent analysis. Sure. It doesn't say. But, you know what? Using currently established mining techniques, I mean, is there no allowance for what mining 10 and 20 years from now is going to look like? I guess they don't assume there's going to be any or not a significant advance in that level of development. And, therefore, if you're going to base it on that, then, yeah, you wouldn't have it. But, I don't know. I don't know. Not be able to mine enough copper to make. I don't know. Something about that one. It could go either way. The last one about the study finding that extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual US demand. Okay. Now, here's the point where you read something recently. It sticks in your head. And then it comes up in a science or fiction. And now I'm thinking, okay, is what I read relevant to this because it kind of had to do with this? But am I misremembering because I didn't study the article closely enough at the time? Shame on me. Had I paid closer attention, it might help me in this particular case. So, gee, it's either that one. It's either that one or the copper. I'll say – I'll just say the copper one is the fiction. I have to guess something.
S: Okay, Bob.
Bob's Response
B: I don't have a lot of insight into the science behind a lot of these. The only thing I can really say is that for three, extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual US demand. That's huge and it's actually too good and we're not that lucky. So, that's fiction.
S: Okay, Jay.
Jay's Response
J: Okay. This one about extracting lithium from shale wastewater, that sounds legitimate. I mean I know that there's – that lithium is in lots of different places. Part of the problem is the extraction itself. So, I think that one is probably science. It wouldn't blow me away that they could – they found a new source that might be very accessible hopefully. Going down to number two, this one that the world would not be able to mine enough copper.
C: That's right, copper.
J: That's right, copper. It doesn't surprise me if this is true just because I know that copper is one of those things that we need in order to transition to electric. So, yeah, I mean that one could be true too. I mean I could see that spanning out to 2050 with the dramatic increase in need. Would we be able to keep up with the demand? We don't even know how much copper is left. Yeah, so that seems right. Then the first one – now I have a theory about the first one. They call them lithium ion batteries but Steve is saying here that lithium is the least needed element, which again is like – sounds crazy. Then of course that could be the rub. You know what I mean? Like that might be the tricky part. So, just because I don't – I think this one is – there's a 50-50 shot. This one is definitely one of the two that I would pick. I'm going to go with the first one. I'm going to go with this one as the fiction specifically because Steve knows none of us know the goddamn answer.
E: There you go. And Cara gets no help. We didn't cluster on anything.
C: Are we all spread out?
E: Oh, yeah.
S: Yeah, so Jay, you're picking the lithium being the third most common element as the fiction. Evan, the copper. Bob, the lithium from shale wastewater. Yeah, so they're spread out. Cara, no help.
Cara's Response
C: Okay. So I wouldn't be surprised that nickel is more prevalent. I don't know about the cobalt. I mean I know we need cobalt for like EV batteries. Are you talking about EV or all lithium ion?
S: Yeah, pretty much. Yeah.
C: Which? Which?
S: Yes, EV.
C: EV. Okay. I feel like maybe –
S: Not like the little ones you put in your flashlight.
C: Right. That's what I was wondering. Yeah. I feel like maybe there are more rare earth metals in there than are listed here. I don't know if they're in higher quantities, but I mean isn't that like a massive controversy, right? Is mining enough of these metals? And I don't know if it's just cobalt and then lithium. And we may not be able to mine enough copper to make the green transition between now and 2050. So you're saying by 2050 there might not be enough copper – like there might not be enough stores of copper or there may be some technical problem allowing us to retrieve that copper in order to make all the windmills and solar panels and things we need and batteries. And then extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual US demand. Okay. This one, the reason I'm going to say that the shale wastewater one is science is because a new study finds that it could. To me that reads as one research lab ran a simulation and said this could potentially be the case. And yeah, a study could totally do that. So I'm going to say that one is science. So now between the battery makeup and can we mine enough copper, a recent analysis concludes that we will not be able to. I have a feeling that we're going to cut it close, but maybe there is a still possibility to do it. So I'm going to say there's probably more heavy metals in the EV batteries. Maybe.
S: Which one is the fiction? You confused me.
C: The EV battery metals.
S: The makeup.
C: The lithium, cobalt.
S: Okay. All right. So since the two of you went for that one, we'll take them in reverse order.
Steve Explains Item #3
S: A new study finds that extracting lithium from shale wastewater in Pennsylvania could produce 40% of current annual US demand. Bob, you think this one is the fiction. Everyone else thinks this one is science. And this one is science.
B: Whatever.
S: This one is science.
E: I read something not long ago about a very large source of lithium in Pennsylvania.
C: Oh, interesting.
B: It's actually kind of cool.
E: This could have been what they were referring to.
S: I don't know. They're referring to the Marcellus shale. And the wastewater from mining that shale has a lot of lithium in it. What the study was was saying basically, well, how much lithium is in that wastewater? And if we extracted all of it, how much lithium would that be? And yeah, if they extracted 100% of that lithium, it would be 40% of US's annual use.
C: Jeez.
E: It would be nice if they had a process to get 100% of the lithium out of that wastewater.
S: Well, here's the thing. Our current process extracts 90%.
E: That's good.
C: That's great. And this is literally just water they're dumping into the ocean?
S: Yeah, basically.
C: That's amazing.
S: That would be 1,160 metric tons per year. And it's just sitting there.
E: Damn.
S: We just need to put it through our lithium water extraction process. We can get 90% of it, which is almost probably a little bit less than 40%. But that's a significant amount of lithium, which we need because the US government basically mandated that we are going to be sourcing all of our battery material from the US. We don't want to be bumping into China.
C: Yeah, we don't want to be causing massive conflicts.
S: And we don't want to be dependent on China for the supply chain. We're trying to get the supply chain back to the US for as much as possible.
C: And people are dying at the hands of the...
S: Yeah, absolutely. Cobalt is the worst. Yeah, absolutely. So this would help. Right now, most of the lithium comes from South America via China. But we do have mines in the US, but they don't produce that much. So this could potentially be a significant increase in local production. All right, let's go back to number two.
Steve Explains Item #2
S: A recent analysis concludes that the world will not be able to mine enough copper to make the green transition between now and 2050. Evan, you said this one is the fiction, correct?
E: I did say that.
S: And this one is science. This one is science.
C: Whoa, no way.
S: Yeah, it's actually worse than you think. It's like really bad.
E: Oh, it is?
S: It's basically... So, okay, so...
E: I said I wasn't wrong.
S: Obviously, there could be some disruptive technology, right? They're not accounting for either some new massive find of copper or whatever. Or they're also not accounting for, oh, we're going to make batteries without copper now. You know what I mean? They're not accounting for that. They're saying, if we had to make enough batteries and EVs and solar panels and wind turbines in order to have the green transition that is mandated in the US, there's basically no way we're going to be able to mine enough copper to make that happen given current technology. They analyzed copper mining over the last 120 years to see feasibly how fast can we expand copper mining. And they concluded... They also calculated how much copper will this technology need. Copper right now is currently mined by more than 100 companies operating in six different continents. And they concluded that we would need to mine 115% more copper than has been mined in all of human history up until 2018. This is just to meet, quote unquote, business as usual. But if we're going to actually dramatically increase our electrification of everything, we're going to need to add, I think they said, 40 new mines every year between now and 2050.
E: That's not happening.
S: So it's very unlikely to happen. I don't get the sense that it's impossible. It would just take an unprecedented infrastructure building in order to make it happen. Either that or we need to really significantly reduce the apparent demand on copper that they're projecting into the future. Which leads me back to the first item.
Steve Explains Item #1
S: By weight, the most prevalent element in current production lithium-ion batteries is nickel followed by cobalt and then lithium. That is the fiction because all of those numbers are wrong. I'll give you the actual breakdown. The most common by weight element in lithium-ion batteries is...
J: Nickel?
S: Graphite.
J: Graphite.
C: Graphite?
S: 16%. That's because of the graphite anode. This will change when we go to silicon anodes. That is actually a huge advantage because graphite is hard to source. That's one of the reasons why the silicon anode lithium-ion batteries, which are coming, are going to be a huge advance. Right now it's graphite at 16%, then aluminum at 15%, then copper at 10%, then lithium and cobalt are tied at 7%, then manganese at 5%, and then nickel at 4%. Nickel is actually the lowest of the major elements. Then 36% is all other materials. Individually too low to list, but that's everything else adds up to 36%. That is the makeup of a lithium-ion battery. Lithium is actually pretty far down there.
E: Why is it called lithium-ion?
C: That was probably the innovation at the time.
S: That's the charge holder. It's made out of lithium, but lithium is the third element. It is very, very light, which is also why by weight it's low. Copper is a lot heavier than lithium per molecule, so is nickel and manganese.
E: A little bit of marketing.
C: Kind of on to something there.
S: Yeah. Yeah, so interesting. The copper thing is like we've got to figure that out. Transitioning the entire economy over to an electric-based technology is a huge order. That is not going to be easy. We should not underestimate that task, shifting entire industries over to a new technology, even if it is a superior technology in many ways. The main advantage, of course, is that it will mitigate global warming. But we can't just say let's do it. We need a plan. And sourcing raw materials is a huge issue. It is a huge issue. I didn't even realize that copper was going to be the rate-limiting one there. I know cobalt's a problem. I know lithium is a problem. Nickel is a problem in terms of the sourcing. But, yeah, copper is going to be a problem too. We need to come up with – that's why they're working on like iron-based batteries because it's like we're silly with iron. Or salt-based batteries. We need to come up with batteries, not even necessarily for cars, like if we're doing it for grid storage. We shouldn't be using lithium-ion batteries for grid storage. We need to save that material for our cars. We could use batteries that are heavier, even if they're less efficient or whatever, less dense, as long as they're made out of really common, cheap, abundant elements. That may be the most important thing in terms of like grid storage, right? It almost seems like a waste. Even like in the house, like the power wall, although I would buy one right now if I could, using a lithium-ion battery to store power in the home may not be the best use of those raw materials. But there's no other option right now. But we really need other options. We need to spread out. Like if you had a bunch of different battery technology so that we are not all dependent on the same battery with the same elements, you know? It's tricky. All right. Well, good job, Jay and Cara. We've had a lot of different match-ups this year. You know what I mean?
C: Yeah, Jay.
E: I'll do it for you. (claps) There you go.
S: All right, Evan, give us a quote.
Skeptical Quote of the Week (1:37:55)
Science is the only self-correcting human institution, but it also is a process that progresses only by showing itself to be wrong.
– Allan Sandage (1926-2010), American astronomer
E: "Science is the only self-correcting human institution, but it also is a process that progresses only by showing itself to be wrong." Alan Sandage.
S: Yeah, not many institutions in human society try to prove themselves wrong, but that's kind of central to scientific progress, right?
E: Yeah, absolutely. Absolutely. Alan Sandage is not a name I'd heard of before, but it turns out he's listed as one of the most influential astronomers of the 20th century. Steve, Bob, Jay, or Cara, have you guys heard of him?
S: I didn't recognize the name, actually. That's interesting. What was his area?
E: Cosmology. He was a graduate assistant to Edwin Hubble, who we definitely know, and effectively his work helped hone in on the age of the universe, approaching that 14 billion. It varied. It wobbled. It went over and under and things, but his work apparently helped get it to the point where we understand it best today.
B: Yeah, cool.
E: Yeah.
S: Nice. All right, well, thank you all for joining me this week.
J: You got it man.
C: Thanks, Steve.
B: Thank you.
Signoff
S: —and until next week, this is your Skeptics' Guide to the Universe.
S: Skeptics' 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 info@theskepticsguide.org. And, if you would like to support the show and 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.
Today I Learned
- Fact/Description, possibly with an article reference[11]
- Fact/Description
- Fact/Description
References
- ↑ 1.0 1.1 Interesting Engineering: Nvidia's Eureka helps robot dog perfect yoga ball balance
- ↑ [url_from_show_notes _publication_: _article_title_]
- ↑ AI-PRO.org: AI Chat Using GPT 4o Power
- ↑ OpenAI: Hello GPT-4o
- ↑ Live Science: Tree rings reveal summer 2023 was the hottest in 2 millennia
- ↑ Neurologica: Spotting Misinformation
- ↑ Bangkok Post: Complaint against 'reincarnated son' of Lord Buddha
- ↑ Heliyon: Lithium battery reusing and recycling: A circular economy insight
- ↑ University of Michigan: Copper can't be mined fast enough to electrify the US
- ↑ Nature: Estimates of lithium mass yields from produced water sourced from the Devonian-aged Marcellus Shale
- ↑ [url_for_TIL publication: title]
|
|
|