Myles: That's all for this edition of the Door County Pulse Podcast. Thank you to Professor Baer for a great conversation and thank you for listening. We'll talk to you again in a couple of days. Welcome back to the Door County Pulse podcast. I'm Miles Dienhausen, Jr. writer and editor here at the Peninsula Pulse. today we're going to be talking about artificial intelligence, which ⁓ very long ago seemed entirely far-fetched. Something science fiction movies. And in the last couple of years has taken on a life of its own, literally, ⁓ is now steamrolling through ⁓ into And it seems like every week there's new updates and new capabilities that we learn about. So this summer, Bjorklunden, when their weekly seminar series, will be doing a whole session taking on the topic of artificial intelligence. ⁓ with me today to talk about this and get introductory ⁓ level, conversation or maybe even deeper than that. on artificial intelligence and how it's impacting our lives is Thomas Baer, physicist and Silicon Valley innovator who also happens to be a Lawrence University graduate. Thomas, thanks for joining me today. Tom Baer: Well, I'm happy to be here and very proud to be a member of the class of 74 at Lawrence University. Myles: Lawrence is a great connection for us to get very interesting guests every ⁓ on the podcast and in the pages of The Pulse, thanks the great work that they do right down the road from us here, from our offices in Bailey's Harbor. So, Professor, is artificial intelligence in the AI world, is this overblown? Tom Baer: don't think so. It depends on who you talk to. There's a lot of hype in the ⁓ area, but I think it is a revolutionary technology that we've really never seen before. And the adoption that's taking place in Silicon Valley is unparalleled. ⁓ so just the way it is now impacting engineers and scientists and everyday people who ⁓ learn to use it is ⁓ just astounding remarkable. And it's mirrored in the fact of the growth in capitalization of the companies that are involved with this. So I think it's a ⁓ change and it's here to stay. Myles: What, for those who aren't really familiar, who have been kind of, like a lot of us... A lot of us hear about it, but don't really dig into it. It's just something that's happening that we can ignore until it's like right in our face. What is AI and what is it not? I mean, I think there are some levels of it that people, there's a version of it that people think of as AI that's basically an enhanced search engine, but that's not what people who are saying that it's going to change society and change the job market. They're not thinking of it in terms of it, what it does for enhancing your search results. Tom Baer: Mm-hmm. Myles: What is it? What is it not? Tom Baer: Well, you I think that's a very good question. And people have been interacting with AI tools for ⁓ more than a decade already with Siri and other automated voice activated assistance. ⁓ Those AI based. They're not as sophisticated as what has grown over the last four or five years. But there's many examples, Google Maps, there's Google search, as you've mentioned, all of these involve elements of machine learning. which are aspects of artificial intelligence. So should be ⁓ familiar. What's unfamiliar is the extent to which current generation of AI tools now have mastered language and have mastered multiple languages and then have expanded their knowledge base to include just an enormous amount of domain knowledge ⁓ across many different domains. ⁓ So really the scope of AI that has changed. and also some of the tools that have been developed, which are more generative rather than just responding to queries like Siri or other Alexa. these ⁓ tools can generate their own output ⁓ and are ⁓ remarkably as I've said, in language. So it seems like you're talking to a responsive person with a vast domain knowledge ⁓ and also very articulate and knowledgeable in many, many languages. Myles: My wife and her line of work was using some of these tools far before me at a point where I was still sort of like dismissive of it. And then one day she came home and just played her ⁓ AI bot that's talking back to her, just sounds like another person and just how quickly it had come. So far beyond what Siri was doing and ⁓ things that we've gotten used to. was a kind of wake-up call for me the first time I heard it and thought wow that could be a human being. I wouldn't know the difference. Tom Baer: Yeah, I think most people would have, would struggle to know the difference, but it's really important to recognize how these machines are using language, which is quite different from the way we use it. And that they have tremendous capability and powers that are almost superhuman, but they also very, very fallible in many different ways. So it's always important to interact, accept what the AI says, but always verify because they often don't know when not. telling something that's accurate. And it doesn't happen often, but 90 % of the time, it's very useful, 10 % of the time, you better check your sources. So don't want to be fooled by how the AI is and how convincing it sounds. You want to use it as ⁓ an encyclopedia, useful source of information, but to go beyond that, to really trust it and not verify would be a mistake in most cases. Myles: Well, you're bringing up something that is one of the dangers that we see because anybody who's on social media has those friends on social media who post things, just basic misinformation that's easy to check, honestly, that you could just put a headline into Google and very quickly find out, no, this is not real. And now we have this other level of some of these AI deepfakes and just a simple search. Just being a reporter, I'm prone to double checking anything before I would regurgitate it. But a lot of us don't do that, clearly, in our day-to-day lives. So it brings up some real fears for where this could go. when used by people who just take it as fact, not just in day-to-day conversations over the, know, having your wrong information for an argument about who's a better baseball player doesn't really matter. But having the wrong information that AI might spit out when it comes to medical advice or legal advice or politicians going at it or even small town level politicians taking information from there. I think there, ⁓ I that we're going to trust it so much and not do what you just said is doing that verify step. Tom Baer: Yeah, I think that's important. It's important to recognize how these machines are actually programmed, how they've learned language, which is a remarkable invention by itself. That's Nobel Prize worthy in my view. How the mathematical algorithms have been which really look at the patterns that are in language and then how these machines are trained to replicate patterns that have existed in ⁓ and reams of of literature that they're programmed on, that they're trained on. And understanding that, you can get a better idea of why these things can appear to be so articulate and also so naive in many, many ways. Part of what we're going to be doing at Björklunden seminar is talking about the underlying technology, how it works, what is some of the mathematics in everyday language, how it works, why it works, and what are its limitations. that important to understand with the current generation of AI tools. This generation of AI tools is completely different from the tools that existed even six months or a year ago, where the logical capability, the knowledge base, the common sense is much, much more improved than it was a year ago. And they're invoking new architectures and new designs for these AI tools that are making them much, much more correct and much more trustworthy when material ⁓ the ⁓ that are used to extract that information are properly phrased. They can also be used to create deep fakes. And I think there's been a significant amount of evidence about that and both in terms of imitating voice, imitating videos of people and creating very realistic videos and ⁓ these things are going to be a challenge to our society and to our legal framework. So there's lots of challenges that are coming about. And again, that's part of what we're going be talking about at Bjorklund, where we've really tried to gather together representatives from different aspects of society, basically, where we feel these tools are going to have a tremendous impact, and will then be the instructors in the course. Myles: Yeah, you talked about the Bjork-Linden seminar and you're going to be joined by Joseph Bruce, a former Illinois Associate Circuit Judge, Michael Clayville, a Grammy nominated musician, and Mark Phelan, a philosophy professor. And when you look at those, you might say, what do each of those have to do with AI? ⁓ more and more, we're realizing it's everywhere. It's not just a ⁓ Valley conversation where you're at out in California. Tom Baer: Yep. way with you all. Yes. Myles: gonna I was just listening to a great conversation Ezra Klein's podcast he had ⁓ an advisor to Donald Trump's administration ⁓ an advisor who talked about something that I didn't realize with AI in that where it Tom Baer: Mm-hmm. Myles: where it gets troubling is that we have a lot of laws on the books, but they're not all enforced all the time. he talked about how AI tools will enable, if wants to, if law enforcement wants to, much more enforcement of the law. Because people are breaking laws all time. And we've seen this a little bit with red light cameras, with speed cameras, and now you might see that exacerbated. He talks about how Now, one of the limits on law enforcement is just sifting through these massive reams of information that they're getting at any one time. And now you could put an AI agent going through a company's documents quickly and find flaws. You could ⁓ go through social media or purchases and ⁓ the lies or find little infractions or parking infractions very simply by saying, your car was tracked here and it was shown that it was parked in a no parking zone. and it could be feeding that. So that's just, maybe just, you thought much about that from a legal framework, from a justice standpoint? Tom Baer: Well, yes, I have and have had a number of conversation with the Honorable Joseph Bruce, who is a classmate of mine and in 74. ⁓ And we were in grade school, high school and college together. So I've known him for a while ⁓ and has been a judge for 25 years ⁓ and AI and is now really looking into how AI is going to affect the legal profession across the board, not only in just the enforcement side, but how you get represented by an attorney who's assisted by AI, a judge who is assisted by AI, and maybe even having the jury assisted by AI. What does that mean? How does that improve the enforcement, improve justice? And how does it take the human element out of it? I want to pause here and make a correction. Michael Clavel, by the way, is a Grammy Award winner. He won a ⁓ Award a ago. And so just a brief update to his resume. ⁓ Yes, yes. We were very, very pleased to hear that. And Michael has been using AI tools as a way to help in the appreciation of music, help in the understanding of music, and also in the creation of music. And so he's gonna come at it from the creative arts point of view. Myles: No. Congratulations, Michael. Tom Baer: and talk about there are also some legal ramifications there about all of the training that's been done with and without permission of the artists. And he'll be talking about that as well as showing how it can be incorporated as a tool, not just to make novelty songs, but to really enhance one's appreciation of what music is. And so we're looking forward to hearing about that aspect. Joe Bruce is going to be talking about, again, where are envisioning the AI tools assisting in the whole branch, both from a judge's point of view and again, from a witness point of view, the whole spectrum. then Mark is a philosopher and his specialty is the philosophy of the human mind. And so he ⁓ has been, and I have been into what is AI in terms of its... presence on the spectrum of consciousness. Where does it lie in terms of actually having an awareness of self the way we do, having an ability to respond to novel situations, maybe even being able to be programmed to have the equivalent of emotions and to have a real human-like consciousness, or maybe a consciousness that is totally inhuman, that's alien, and that that might be a unique perspective for us to view our own self in sort of a mirror of another consciousness. ⁓ And so there's a whole raft of interesting topics I think we're going to cover during this week long up at your Björk Linden. Myles: those who are unfamiliar with the Bjorkland and Setup, what they do in these week-long seminars is it's a three-hour, 9 a.m. to noon every day for five days. So you really get a deep dive and an immersive experience in this. And obviously their campus, I shouldn't say obviously, their campus down in Baylis Harbor on the water is beautiful setting, a beautiful experience for ⁓ learning meeting other people interested in the same topic. What do you use AI for, if anything, in your day-to-day right now? Tom Baer: Well, I use it all the time. You know, I'm a scientist. I'm a physicist. I do research, both experimental and theoretical physics. And AI tools are just revolutionizing how you do, how I do what I have done for the last almost 50 years, more than 50 years. I've been programming computers and using them to do science, using them to control and automation, build, I've designed and built dozens of commercial products that incorporate microprocessors. And it's revolutionized that whole field. Now you don't necessarily need to write code anymore. The Claude code and other AI tools will essentially do the program. You tell them what program you want written. ⁓ And the program, the Claude or other code-based AI tools will write the code for you. Then you can troubleshoot it. And if there's an error, You can copy the error and the AI evaluates it and corrects the error. It's a totally different iterative process and it allows greatly expanded capabilities because the domain knowledge across science of these tools is just enormous. They have basically been trained on dozens of journals that go back decades ⁓ have this at their fingertips. ⁓ ⁓ is a remarkable tool to use in that regard. with the caveat that it also hallucinates. And so it will make up references that it thinks is rough relevance that don't exist. you have to be able to be careful about what it does. But I use it for programming. I use it for understanding different elements for tutoring in areas of science and technology that I'm not as familiar with. ⁓ And always double check it either against another AI or my own sources. But it's ⁓ the way I do science and the way I think most physicists and most scientists are going to be doing science going forward. Myles: Do you worry about it eliminating the next version of you coming along who's 23? Tom Baer: Not me, there's, there, there are, I would be concerned about people who've spent a lot of time learning how to do coding. It's useful, but you won't be doing it the way you probably learned it, which is by typing in a lot of lines of code and troubleshooting it yourself. You'd be using a completely different tool. And, uh, but I'm an experimentalist and I'm a builder and innovator. So I build things and right now AI does not have that capability. Myles: Mm-hmm. Tom Baer: And although it can do theory pretty well, unless you compare theory with experiment, you don't know if the theory is right or not. And so you need to be able to close the loop around a scientific investigation using both theory and experiment. And experimentalists, think, are going to be human-based for the time being with an assist by the AI. So no, I'm personally not worried about obsolescence in that sense. I do feel that it is going to be absolutely critical for me to become very fluent with how AI tools are incorporated into scientific disciplines, because that, you're going to be left behind. So in that sense, I've got to learn the tools, and I am learning the tools, but ⁓ they be just tools that I'll be using and be productive as a scientist going forward. Myles: One of those things, so you have a background in biomedical research, correct? And so one the things that's pretty... Tom Baer: ⁓ yeah. Myles: exciting to think about with AI. There's a lot of parts of it that are scary and worrisome. There's a lot, you know, when they talk about job loss and up all these coders that we just ⁓ 20 years telling every young kid in America to go become a coder and this is your guaranteed job and this is, you know, forget your dreams. If you can just focus on this, you're going to have a job and a good paying one and ⁓ they can't find jobs. So there's a lot of scary parts of it. And I think the coding that you're talking about, that's what they refer to as Tom Baer: That's right. Myles: kind of vibe coding now, right? Instead of having to write the full architecture you say I want an app and you got to be good at making your prompts You have to be good at instructing it But if you can give it the right instructions and you're using the right program you can say I would like to design a website that will do X Y and Z and it's gonna be somewhat dirty code is from what I understand but it's gonna be Workable it'll be a workable site. That's 80 % of what you want it and then you got to refine it maybe ⁓ Sort of like Tom Baer: you Yeah. Myles: editor refining someone's copy, which is basically what I've seen that if I tried to get AI to write an article for me on something relatively basic, I obviously couldn't do something where I had to go interview somebody tell someone's life story, but it could do the basic information of creating a listicle or something like that, but then I would still have to go and edit it to make sure it's right. And I think that's where a lot of it probably lies and what you were saying as well. In the medical world, there's been talk that like, you know, the most, the very optimistic people think be a matter of years, not decades, for AI to help us come up with breakthrough treatments or cures even for certain devastating diseases. Are you that optimistic on that front? And why would it, why would it advance this so quickly? Tom Baer: Yeah, well, a couple of responses to what you said. Dirty code is what the computers did six months ago, and they don't do that much anymore. The code is actually very high quality, well documented. and probably better than anything I could write, even though I've been programming for almost 60 years. And so evolution ⁓ the capabilities the coding front have been astounding. so again, AI is changing so fast you have to be careful how you characterize it because characterizing it in one week and the next week it's now ⁓ advanced beyond its limitations were. ⁓ Myles: Right. Yeah, your seminar might have to, you might rewrite this and reorganize this thing five times by time June rolls around. Tom Baer: Well, you know, two years ago I gave an initial seminar. It was the class of 74's 50th reunion. And I gave a seminar there where I went through some of the introductory material that I'll still use. it has now, comparing to where the... where the field was two years ago, there's just no comparison. so it is, and the changes are accelerating. So it's not slowing down. But to get back to your comments about medicine, I've worked in the biomedical field for now for 30 years, developing inspirations, working closely with doctors across the world, actually, and also at Stanford and the School of Medicine there. ⁓ I have a pretty strong ⁓ background in this area. And I would be cautious about the impact that it's going to have in terms of breakthrough cures and things of that sort. Where it is really, I think, going to have a tremendous impact is democratizing the quality of care across the world. Where now there is going to be an AI tool that helps in the diagnosis and that can be at a very expert level. Again, there are hallucinations. You got to take it carefully. But for the most part, it's going to provide access to a ⁓ variety of ⁓ material that you just don't get with the, your doctors across the So I think it's going to have a major impact there. ⁓ When I talk people here at, at both at Stanford and other doctors that I know, they say one of the major aspects of it is just facility. Sorry about my dog, my son has. two dogs in the background here. Both of them are 80 pound dogs. Myles: That's all right, we've got an office dog here that I'm surprised has not interrupted this conversation yet. Tom Baer: Yeah, and then the gardener is cutting the lawn. So we may have to edit this portion of the podcast. So the dogs are reacting to that. Myles: ⁓ our AI editor might figure that out. Tom Baer: Yeah, yeah. the where the docs have said that it has the greatest impact is in facilitating their own diagnosis by giving them current references in the field that they can go look up and say, OK, based on this description of the symptoms, here are the most resources or references ⁓ which facilitate your new diagnosis. ⁓ And say that is extremely helpful. The other aspects of it are, about the background. ⁓ other aspects ⁓ of that they see is very helpful. It's just in transcription. So rather than having to type in the comments the patients, ⁓ they get from patient ⁓ and the AI records the conversation, summarizes the doctor's notes. They say that this a tremendous improvement in their ability to focus on the patient. And so I think... Myles: It's okay. Tom Baer: Those are the two areas, facilitating the diagnosis and helping just in the day-to-day patient interaction that is going to be helpful. And moreover, then also for people who do not have access to high quality medical care, the doctors can use the AI to help facilitate their diagnosis. So I think it's going to have a major impact. ⁓ far as discovery goes, ⁓ ⁓ of medicine is so complex ⁓ involves so many different aspects of testing ⁓ verification. and the sorts of things I do as an experimentalist, that I think it's going to help, but I don't think it's going to see the impact in medicine that you see in computer programming in other areas where AI is just going to have a much more immediate impact. There are some exceptions. There's been some remarkable inventions based on using AI tools. Alpha Fold is what was a technique that was invented to describe how proteins fold. And that has revolutionized that particular area of protein structure. And again, what's interesting when you talk to people who use it is that it works most of the time, but you always have to test it and you have to do an experiment to verify it. So it is extremely helpful, but it's not the final word. And so a lot of people who are doing protein crystallography and other types of very sophisticated techniques for measuring protein structure, they're still going to have jobs because you can always have to verify. Myles: Yeah. Tom Baer: as I said, the predictions. Myles: one of the things that, and correct me if this characterization is wrong, but one of the ways I've thought of AI has been that it can do a lot of these things, can enhance a lot of these processes. If you're making a movie or doing something creative, ⁓ someone like who's not ⁓ a sound editor or anything, but if I wanted to produce something, it's gonna make me able to do that and accelerate that in a much better way. But it can't necessarily, like the iPhone, artificial intelligence is built on what came in the past and when the iPhone was produced, nobody thought a touchscreen could be a viable way to have a phone in your pocket all the time. So my thinking is something like the iPhone is not something that AI would create because it wouldn't have that vision of the thing that wasn't invented. Tom Baer: You Myles: that fair? Tom Baer: ⁓ evolving so quickly that I that and ⁓ given ⁓ different they use for generative AI where it can come up with unique images, unique ideas in many ways ⁓ in different combinations. I think that having the AI actually be ⁓ creative source new ideas and innovations like iPhones. is effect, I would predict that that's the case. Is gonna be super intelligent more? You don't Again, the domain knowledge ⁓ far surpasses human being could have. It basically is with the ⁓ contents of the of Congress. So mean, ⁓ nobody reads millions books. It's fluent in dozens of languages. ⁓ And so ⁓ has capabilities humans don't have. And they're building into it the ability to be creative to... use processes which create unique and original ideas, images, sounds, and music. And so it has that potential. There are elements that it doesn't have. It doesn't know how a puppy feels. It doesn't know how a person feels when it touches something like that. It doesn't have the human experience, but it has different capabilities and skills that I think are going to be very informative and allow us to explore. different ways of attacking and evaluating situations as well as coming up with new solutions. So I expect it's going ⁓ ⁓ able to create the equivalent iPhones or ⁓ other inventions that sort. And it'll be a tool which we will use on its own as well as with its own ⁓ capabilities. ⁓ Myles: ⁓ So I was going to ask you about electricity. So one of the things with AI and all these things that it can do is doesn't come without its drawbacks. One of those being the massive energy needed to operate these data centers. And now in Wisconsin you've seen this. The comedian Charlie Barron has been ⁓ speaking out some of these data centers. You know, there are estimates that... Tom Baer: Yeah. Myles: just a couple of these new data centers would use more electricity than every residential household combined in the state. And then you have the water usage, which we had a long, like a decade long fight to control the Great Lakes water so that they couldn't take it to the southwest. And now there's a lot of companies wanting to locate data centers on the Great Lakes because of that access to water. So Tom Baer: Mm-hmm. Right. Myles: This tool can do these incredible things, but do we have the energy infrastructure and the water infrastructure for it to continue to grow at the rate it does with the use that it entails? Tom Baer: Excellent questions. Just to ask the background, I've spent beyond what I've done in biomedicine and physics. I've also spent a lot of time in working in climate mitigation, climate change mitigation, and looking very carefully at energy needs in the country and across the world. So this is an area that I've been very, very concerned about and have considerable experience in modeling the different aspects of energy demand. And I think the AI needs to be put in its energy has to be put in perspective. The analogy I use is Bitcoin. Bitcoin right now ⁓ has and is still using more energy than ⁓ all the generative AI tools that exist right now. And we've put up with that, okay? ⁓ And one really seems, occasionally it's complained about, but the amount of energy that Bitcoin uses to do their calculations to determine if the blockchain is valid. is just doing nonsense. It has no ⁓ contribution to society, not solving any problem. It's just basically solving a riddle. And the who solves the riddle gets the Bitcoin, gets paid in Bitcoin, ⁓ and pays for the electricity to run the Bitcoin machines. And this has been going on all over the world, using energy that's equivalent to what gets used in a country, not just the state of Wisconsin. And so, you have to view AI in that context. We already use technologies ⁓ applications ⁓ I'm not sure have a lot of intrinsic worth. You have to look at how much energy is used when you download a video on your cell phone, which is substantial, actually. And so use the of energy that AI is going to use is a small fraction is using right now, not what it is going to use. What it is using right now is a small fraction of what we use for looking at YouTube videos and just using the internet and electronic commerce. So it has to be viewed in that context. And I don't want to minimize the impact that this can have on local energy prices or on resources like water access. But it's not as bad as it's been played out to be in a lot of the press right now. Is it going to be? Well, it depends on how much it gets integrated into our society. Overall, We have to go through an energy transition anyways. As we move to electric vehicles and electric cooking, all the other aspects, the heating and cooling, heat pumps, et cetera, we're going to have to redo our ⁓ energy ⁓ in to move away from fossil fuel use to the use of renewable resources anyways. And we need to incorporate into that the idea that ⁓ AI is to take a non insignificant amount of energy. But I think compared to what we use for transportation, what we use for heating and cooling, it's going to be a very small fraction of that. So it has to be viewed in that perspective. we need to plan for it? are there places where it is totally inappropriate to put in these ⁓ ⁓ and resource-hungry data centers? Absolutely. But can they be integrated into our environment and be essentially a not a nuisance, absolutely it can. And not terribly concerned about it. In spite of the fact, I am very concerned about overall climate change ⁓ the use of fossil fuels going forward and the rest. The AI energy needs and water needs are things that I think can easily handled. Myles: Might be more encouraging if we were in parallel also pursuing some of these renewable goals that you mentioned there, right? Tom Baer: Well, that we have to do and ⁓ there considerable pressure against doing that. And that's a problem. And is ⁓ a lack of concern to a large extent for the environment and for what are going to be future energy needs. And those things, we can't ignore it. Is it I think absolutely it is handleable and we can absorb the ⁓ extra energy needs society. Myles: Yeah. Tom Baer: with the AI tools, it's not going to be a problem. It's going be much more of a problem shifting from gasoline burning cars to like EVs. That is going to be a major impact on our ability to generate renewable resources. But the AI tools, it's going to be something that needs to be planned for, but I think it's very, very acceptable. And we've accepted the fact that we're burning up all this energy for nothing for Bitcoin, unless you're a believer in Bitcoin as an alternative. Myles: Yeah. Tom Baer: a currency source and financial market, I'm not convinced it needs to be there, but other people are. other that, again, if these calculations for Bitcoin were doing something useful, could almost it. But the AI, ⁓ uses of AI going to be, I think, ⁓ imminently useful. ⁓ And so ⁓ think it's we just plan for. Myles: you Well, final topic I wanted to touch on is, and I'm ⁓ guessing followed this a little bit, what AI in warfare. ⁓ And we've Anthropic and the Pentagon in these very contentious, ⁓ Anthropic being one of the larger AI companies ⁓ and the one think most consider as like the fastest moving on this front, but they, Tom Baer: ⁓ yeah. Myles: They had a content negotiation with the Pentagon that I'm not smart enough to get into all the details on, but the Pentagon basically saying, and Pete Hegsteth basically saying, hey, we want to use this however we want to. And saying, we want to put some guardrails on it because essentially we don't quite trust our technology quite yet for you to unleash it in warfare. And... Tom Baer: Mm-hmm. Myles: there's also allegations and questions about whether or not it is playing a role in the current conflict in Iran. It almost certainly is being used. I can't imagine the military not using but there is speculation that it might have been used in the bombing that up hitting. Iranian school allegedly a US bomb hit Iranian school and killed more than a hundred young women It raises some real questions that are beyond energy and costs and trade-offs but like Can we our own use of this if it then becomes something that's used in warfare? Tom Baer: Yeah. Yeah. Well, I think these are all good questions and there's articles now in local newspaper and in ⁓ national describing the And what I do applaud is that we're having the dialogue, we're having a conversation, and these are very serious topics. And there's nothing ⁓ we should ignore assume that we can do these things. So the fact that Anthropic has said, let's put the brakes on this at least temporarily and let's have this discussion is something to really applaud. And I hope the government will, in this administration, will take it seriously to sit back, get people involved from the community, get a cross section of society involved in what the ethical use of these technologies. It gets to a broader question about science and scientific ethics. How do you introduce new technologies into society and who puts the guardrails on? And ⁓ I think scientists by themselves should do it. ⁓ Myles: Mm-hmm. Tom Baer: any one sector of society. And part of what I've always been in favor of is having the made by a cross section of the that's going to be impacted by this technology. And so I think the fact that these conversations are taking place ⁓ and I Anthropic for really forcing the issue ⁓ and where decisions will end up is really beyond me. I don't know the technology well enough to be able to determine what is its ethical use. I'd love to be in conversations with ⁓ with this issue, know more about the technology, that aspect of it ⁓ than I do. And then involve from all classes, ⁓ across of society to try to set a policy that is consistent with ⁓ our and moral ethics here in the United States. And hopefully we can set an example for the rest of the world. ⁓ So don't know if that answers your question, but Myles: No, that's, you know, there isn't an answer to that question, right? Not yet, anyway. ⁓ Tom Baer: Yeah. Yeah, no, very true. Very true. what I've heard, ⁓ the the what in the paper just today that I was reading the local market news. ⁓ Myles: Well... Tom Baer: is that the ⁓ targeting strategies across aspects of this ⁓ conflict ⁓ the Israeli US side have ⁓ using AI and anthropic tools in particular ⁓ to the targeting strategies. ⁓ And using data from satellites that has, Silicon Valley company has ⁓ an interesting reputation, put together and really looking for patterns that ⁓ AI is specific specifically suited to look for and that those targets have been really determined by the AI, then approved ⁓ the human ⁓ structure. That's what was in the paper. And so I don't know if that's way it is or not. That's sort of the state of things. ⁓ boy, I am sure glad we're having this conversation because I think it's absolutely critical for us to have. Myles: Yeah. Yeah. Well, and what Dean Ball, the former Trump advisor for AI had said is, legal structure and our political structure is lagging so far behind in this conversation. And by the time ⁓ of seminar in June, I would guess that as if voting wasn't complicated and hard enough, Tom Baer: Yeah. Myles: I would think, would guess that AI is going to start to creep into ⁓ discussions about the campaign. ⁓ not, I think anybody talked about it in the last election cycle that we covered. I don't think anyone's really talking about it now from a political perspective, because they don't know enough. I don't know enough, right? Like very few people actually know enough. So. Representatives are very afraid to put this on the table, but it's at a certain point, it's going to bubble to the top because you're going to start worrying about jobs, you're going to start worrying about how it's used in warfare, how it's using energy, how it's using water. They're going to have to... Tom Baer: Yeah. Myles: they're going to have to brush up on this pretty quick because it's going be, I think it's probably going to elevate into the campaign cycle for the 2026 cycle. ⁓ Tom Baer: Yeah, I think not only as a topic, but unfortunately, I think we're already seeing evidence that it's being used to finance the political campaigns of the various ⁓ And that's what I would encourage your listeners to pay close attention to. Who is funding what? And because there's tremendous amount of money that's available right now, given the market capitalization of these AI companies, and they are throwing their financial weight around. ⁓ that's of the issue here is that ⁓ lot the Myles: Yes. Tom Baer: political decisions are probably going to be determined by campaign contributions. And I think that needs to be watched very carefully because we have a democracy and the power is in the people and we can the course. And we shouldn't let the company set the course. We shouldn't let the Defense Department set the ⁓ The ⁓ should do it. ⁓ we've to teach ourselves what the are, how it works, and understand what the implications are. And that's a responsibility. that we have as citizens. And it's a particular responsibility I have as a scientist, which is one of the reasons to get back to New York and then I'm doing this seminar. I'm trying to teach, to educate people as to what AI is, what the power consumptions, the water consumptions are going to be. We'll be talking on, touching on those topics. What it's going to bring us, what the ethical challenges are going to be. As a scientist, I have been involved peripherally in the invention of these tools. because of that, I have a responsibility to... educate society so as a society we can decide how they're going to be used. It shouldn't be left up to the companies, should be the government through the democratic process and it definitely should not be left up to the scientists. It be a combination of all those elements that make these decisions ⁓ and is one of the most important ethical challenges ⁓ we're going to facing because of the ⁓ true power of AI tools. Myles: and just how fast it's moving. One of the things that they've talked about is, you know, it usually takes decades for a new technology to really permeate the culture and ⁓ from what they're saying is this is just nothing's ever gone this quickly. So. Tom Baer: So moving. Yeah. No, it's what I said right in the beginning and I couldn't agree with you more. It is ⁓ over so many aspects of our day to day lives that I ⁓ not taking but it is a tool which we can use to enhance and to explore different aspects of ⁓ our daily ⁓ And ⁓ people be using it to explore our lives as well. there's a lot of questions. And again, at the Björklunden seminar, we hope to go through those. We've we've a musician, Grammy award winning musician. We've got Joe Bruce, who's been a judge for 25 years. We're going be talking about all these topics. Myles: Yeah, well that seminar is June 14th through the 19th and in addition to the seminar, even if you don't take part in that seminar, there will be a roundtable discussion on June 16th as part of this week. be at 7 p.m. that's open to everybody, but this sounds like it's going to be a really fascinating week and that roundtable sounds like it's going to be really fascinating. So thanks professor for your time and talking about this and exploring some of these questions and we could probably do this again in 48 hours and have a totally different discussion. Tom Baer: Yeah, ⁓ agree with you there, Myla. Nice to meet you and look forward to meeting you hopefully in person when I'm out there indoor county.