How AI Affects Everything (with Chase Tremaine)

This is episode 78 of the Renew the Arts Podcast. How AI Affects Everything with Chase Tremaine. Welcome to the Renew the Arts Podcast, where we discuss the role of art and creativity in the church and in the world.

I’m your host, Michael Minkoff. Our motto at Renew the Arts is Liberate Christian Creativity. Our mission is to cultivate Christian communities by inspiring art partnership and supporting artists.

If you’d like to join our community of monthly donors and contribute to this podcast, please visit renewthearts.org/donate.

If you want to join the Porchlight Network and begin your journey as a partner in the arts, go to app.porchlight.art and sign up as a host or a tender.

All right, so I’m here with Chase Tremaine. He is, as far as it’s connected to us, he edits our podcast, so he’s kind of been an invisible hand for you all who are listening. For quite some time, you’ve probably heard about him.

But he’s also a musician and a dad and a podcast host in other places, and he’s really interested in the topic of AI. And being that he’s connected to us and works on this podcast so much, he was one to have a conversation. He’s going to basically run it, and I don’t know if you’re basically going to be interviewing me or if I’m going to be interviewing you, but it’s going to be a pretty free-form situation here.

So Chase, tell us a little bit about what you think is so important about AI in terms of the arts and right now.

Yeah, and I want to preface this episode with both why I think this is an important topic for us to keep talking about, as well as why I might be a worthwhile voice to hear in this conversation. You know, as this represents the fourth episode of this Renew the Arts Podcast season that is focused on AI, there is a chance that, you know, we’ll split this conversation up into like a two-parter episode. And then at the end of the year, the mailbag episode will likely pertain to your listener questions about AI, which means over half of the 2026 season of this podcast is dedicated to AI conversations.

You know, episodes 74 and 75 were Michael monologuing on the topic. 76 was Renew the Arts Creative Director Rusty Hein having a conversation with Michael on the topic. And it’s very likely that you are coming into this conversation in this episode with some level of fatigue around it, especially if you’ve been keeping up with the news and just the fact that it’s seemingly being stuffed down our throats at every turn.

But one of the reasons I think AI is so important for us to continue talking about and having these conversations about is because it, unlike almost any other topic, covers so many relevant sectors of life. It touches on ethics. It touches on politics and business and power.

It touches on religion and philosophy and ideas of human rights and personhood and the value of our work and our art, the meaning of our work and our art, international relations and law and technology. So many things, the environment even, as I’m sure we’ll get to, it is involved in all these things. And if we, the people who are actually being forced to use AI or trying to figure out where AI fits into our lives, or potentially losing our jobs to AI, whatever it might be, if we don’t have these conversations, I think we’ll end up just accepting the top-down narratives that are being provided to us, ultimately by the people who have the most to gain from AI becoming as important and powerful as possible.

You know, there are a lot of people who will spout kind of the talking point of like, AI is inevitable. You just gotta live with it, or you need to adopt it as quickly as possible, or you’ll get left behind. Well, where did that narrative, where did that idea come from?

It’s coming from people who have most to gain.

DeepSeek and OpenAI and Google, you know, not from the people down here that will ultimately be affected by it, good or bad. Because ultimately, I want this conversation in this episode or these episodes, whatever it turns out to be, to be a resource to you, our listeners, not to give you black and white answers of AI good, AI bad, do this, don’t do that, but rather to equip you with discerning for yourself and your context, the ethics of AI to determine what is healthy and good for you in your situation, where you might be forced to be using it at work. And so, the people who are saying, don’t use any AI at all is simply not an option for you.

Maybe you are already using AI a lot, and you’re not sure if you are overdoing it or underdoing it. Or maybe you have friends or family members that you’re starting to worry might be going overboard, might be delving into like AI psychosis, or maybe just need to back it up a little, and maybe this conversation can help equip you to have conversations with them, loving, understanding, helpful conversations, so that we can be human first and relationship first in how we deal with AI, and in how we deal with one another on what is becoming an increasingly contentious topic. I hope that inspires or encourages the desire to like keep talking, keep listening, keep thinking.

And again, this conversation that Michael and I are having is not just between the two of us. I also want it to be a conversation that is in dialogue with episodes 74-76, so that, in a sense, is in dialogue with Rusty, even though he’s not here.

Well, you had a lot of thoughts.

I had a lot of thoughts about Rusty’s conversation, which we’ll definitely be getting into. But also, like, in dialogue with you, if you go to renewthearts.org/podcast, at the very top of that page, there is a submit a question button. So if you have questions for Michael, even if you have questions for me from this episode, you can submit those there, and Michael will address those at the December episode.

Of course, you can ask non-AI questions, but…

And please do.

Please do. Give us something non-AI. Ask us questions about Erin Michelle’s episode from last month.

We got a break from AI for a wonderful conversation with her about songwriting. So if you missed that, please go back and listen to it. But coming back to this, I want to kind of introduce myself in regards to this topic before we start kind of getting into the meat of it.

So, well, for one, my main three jobs as a freelancer, I’m a book editor, a podcast editor, and a music producer. And all three of those are jobs that are endangered by the continuity of advances in AI. You know, I, as a music producer, effectively need to convince people that I am better or bring more value than using Suno.

That as a book editor, that I bring more value than using Grammarly. And so on a day-to-day basis, that is one of the main reasons why AI is very important to me. But my relationship with AI goes back to the original launch of Mid Journey, the first big generative image AI program that came out in mid-2022.

This was even before like the main public launch of ChatGBT, when ChatGBT 3.5 released in November of 2022. So Mid Journey was ahead of that and made a huge splash. And I was using it all the time.

I was a really big fan of it. I thought it was extremely interesting and ended up using it for multiple of my music releases. I used art created by Mid Journey for two or three singles, leading up to an album where I used Mid Journey for the album artwork.

And then when I created CDs of that album, the insides of the CD was just more artwork from Mid Journey. I was just very smitten by it. I was very impressed.

I’d also had some falling outs with graphic designers. There was one graphic designer I’d worked with in the past who just wasn’t doing freelance work anymore. Other people I talked to thought they weren’t a good fit for the project.

And then I had a good friend who actually designed something good for the album artwork, and I was free to use it. But my issue with what he created for me was that it was just manipulation of a stock photo. And I hated the idea of releasing album artwork that used a stock photo that potentially other artists have also used.

That I might at some point in time find another song, another album with that same photo on the front cover. And I was tossed between picking that or having Mid Journey create an image that has never existed before, that I know is completely unique and original. And so regrettably, as I would say now, I went with Mid Journey.

And as time went on, and I learned more about how Mid Journey does what it does, and how generative AI machines are trained, and how they create their work, I became increasingly uncomfortable with the fact that I had used generative AI for my album artwork, and went around as many places of the internet as I could, and replaced my album with a new piece of work that my wife and I made together. And so that was kind of the start of me turning in toward being way more cautious regarding AI use, and have since just been studying the topic a lot, especially like I said now as being like a podcast editor, book editor, and seeing like my lanes of work being infringed upon. It’s like I really need to understand what’s happening here.

And earlier this year, after my second daughter was born, my wife and I were trying to keep her out of daycare for as long as possible. So there was this about three month segment of the year where my wife had returned to work, and I was basically full-time dad with our infant daughter. I was getting as much work done as I could during her naps or during nights after she went to bed.

But it was mostly just with her all the time, which gave me an incredible amount of time to be listening to news reports and podcasts and YouTube videos on this subject, which have kept me greatly informed on what is going on in the AI industry, whether it’s the stock exchange or US government decisions, lawsuits, all these sorts of things. And I actually think that those would be a fitting place for us to get the AI conversation going, just to do what I hope will be a fairly brief overview of recent news events from these different sectors of notable AI categories and AI companies. So if you’re listening to this without very much information about what is actually going on in the world of AI, I hope that this will be very informative to you.

And if you are perhaps listening months later or years later as a resource, which I hope you will do, and I hope our conversation is suited to be useful for years to come, that I hope that this will help kind of contextualize our conversation and that this conversation around the current news and events will provide a basis for where we are in the current AI evolution and from which this conversation then kind of turns.

Hey, thanks for listening to this episode. At Renew the Arts, we’re pretty big on arts partnership and hospitality. We actually have a whole network we call Porchlight, designated to art and hospitality, where hosts, attendees, and artists gather in homes, backyards, and other third spaces across the country to appreciate beauty together.

This October 23rd and 24th, you’re invited to join us in Dayton, Tennessee, just north of Chattanooga for our very first Porchlight Festival, where this network, normally operating locally, will gather together alongside Renew the Arts Podcast listeners, our staff, our board members, folks from the local Dayton community. It’s going to be really wonderful and we want to see you there. You’ll get to hear Wilder Adkins, Virtuous, and Son of Laughter.

Plus, the festival will feature dance, a curated visual art gallery space, and tons of other musical performances. Visit porchlight.art.com/festival for more details and to RSVP. And just like our local Porchlight gatherings, we’re sticking to a suggested donation model, which means there isn’t really a set ticket price for the festival.

So, get online, give what you can, and save your seats. We can’t wait to see you at Porchlight Festival October 23rd and 24th. Once again, visit porchlight.art forward slash festival.

All right.

So, what I want to get into now is the, I think the most interesting and most pertinent news from the AI industry from these past few months or so. Some of these news stories actually just come from these past few days. Some of it is kind of breaking.

If any point you want me to slow down, or if you want to interject or ask a question about something, ask for clarification, I’m absolutely happy to do so.

Okay.

So first, some general industry news. SpaceX, which also includes the subsidiary SpaceX AI, launched its IPO to pretty record-breaking numbers, which per all the headlines, turned owner Elon Musk into the world’s first trillionaire. However, those stocks have already fallen enough from their initial launch value to decrease his net worth by approximately $200 billion.

So he’s not quite the trillionaire that he was for a few days.

Wow.

But following in SpaceX’s steps, the two biggest dogs in the AI race, at least within the North American side of it, are the companies OpenAI and Anthropic. And they are both also trying to launch IPOs by the end of this year, which, quite frankly, is a foreboding sign for everyone who fears that this AI boom might be a bubble that might be on the verge of popping, similarly to the dot-com bubble from about 25 years ago.

Hey, Chase. Yes. So can you pretend that I am a person who doesn’t know what an IPO is?

Yeah, I will pretend that you are me from three weeks ago. Um, so IPO stands for Initial Public Offering, which is when a company goes public on the stock market for the first time.

Okay.

And so this initial sale is when it’s determined what the stock is worth. And so it’s what people are buying the stocks for.

Right.

And then that number of stocks multiplied by its current market value, which is based on kind of a mixture of like how popular it is, what people are paying for it, and theoretically based on how profitable a company is, but not really, ends up becoming a company’s public valuation worth.

It’s worth.

But this IPO sale, when you’re putting new stocks on the market, it is one of the only real opportunities for new money to enter into the stock market. Most of the time, money is just moving around and washing from stock to stock, as someone sells one stock’s value in order to buy another stock’s value. And so an IPO is kind of a big deal, especially one of this size, where like real money comes in, people are actually buying them.

It could be borrowed money.

So what would be the advantage to, let’s say, an AI company, particularly to do a offering of this kind?

So that’s kind of the scary thing, is that people are buying SpaceX stock and planning on buying OpenAI and Anthropic stock, if they do end up going public later this year, with hopes of being the early investors that you’re in on the very first day of a company that is growing and just going to keep growing. And so you will end up making a lot of money whenever you decide to sell your stocks at a hopefully much higher value than when you bought them on day one when the IPO launched. The scary part of it is the answer to your question, what do these people stand to gain?

Well, when you make a company public, you are basically giving ownership of the company from your original team of whoever founded the company, whoever is CEO of the company, the original lenders and stakeholders and investors. All of those people are releasing their ownership of the company to the stockholders. That’s not entirely true assessment necessarily because oftentimes founders, presidents, stakeholders will maintain a lot of the stocks for themselves.

But their ownership of the company will inherently be watered down that will give those stock owners some amounts of ownership and power within the company.

Wouldn’t that be good though in the sense of that means that there’s more voices maybe contributing to the direction of these things and maybe a better understanding of the common good or anything like this? Like what’s the danger?

Yeah, so a lot of the people who are investing are not going to be like on the ground people who want better regulation on AI companies.

It’s going to move it in the direction more of bottom lining.

Yeah, it’s a lot of Wall Street suits who often, frequently, not necessarily are going to push these company further and further in the direction of like, it’s just about us profiting because our main priority of stockholders is for the value of our stock to increase. It’s very much more of a bottom line money end game versus like, I’m wanting some level of ownership in this company so I can help steer the ship and make it a more ethical company. But again, going back to your previous question, I just want to round this off, is that the scary part of the IPO is that for all the stocks that the presidents and founders and initial investors sell, they’re just receiving the money of the stock sell.

So they get to actually cash out. And so if you have a big IPO sell or just in general, even years, decades down the line, a big stock sell of a company that’s on the verge of collapse, you are handing someone else the bag, taking your profit and running. That’s why insider trading is so, well, illegal and just wrong in general, is that you’re specifically trying to get as much money as you can before you just go bankrupt, and so there is some genuine fear that the insider information of why Anthropic and OpenAI and other companies are racing toward IPO, is that they see the writing on the wall, they know that their companies are bound to failure, and so they are trying to get the initial investors paid back, make some of their own profits before their business becomes unprofitable, and then bankruptcy or debt dissolution or whatever of the company will fall on the stockholders whose stocks might crash from, you know, 250 per stock down to 50, 25 per stock.

Well, let’s talk about something else about that, though, too. If there is sort of a belly up situation, you also have to understand like the larger economic picture is very much integrated with government subsidies and support, especially if you’re talking about like SpaceX. I mean, give me a break.

And so, yeah, just for it’s also the case for context like SpaceX does like, yes, survive very much on like government subsidies. It’s not its massive valuation as a company is not based on its own profits. Like Walmart, its total like market valuation is, I think, about half of SpaceX, despite being like 30 times more profitable, just for some like numbers and context there.

No, that’s really helpful. Yeah, because I think when you say, well, the stockholders will end up holding the bag, given how integrated AI already is into infrastructure, it may actually be a similar situation to the bailouts and tarp and all the rest of that that happened in the collapse of mortgages where the government, where we hold the bag with the government, even if you don’t even have stock in AI, you’ll be holding the bag.

Yes, no, I was planning on getting there, and it’s even scarier than that. So, there are three groups of people that might get hit, because my next piece of news is that index funds recently changed their own rules of how they function. There are four or five main index funds, which are generally considered the safest way to invest your money.

This is like retirement funds, IRAs, like 401Ks. You know, this is where they put your money, basically, is these index funds, S&P 500, NASDAQ. And ahead of the SpaceX IPO, not all of them, but some of the major index funds changed their rules so that they could start investing money immediately upon the launch of a new company, which is extremely dangerous and puts probably billions of people’s retirement fund money at risk.

I bet a lot of people have already lost a decent chunk of retirement funds because some of their index fund money went to SpaceX, which is already losing value from its initial launch. And there used to be protections against that, where index funds couldn’t invest in any company until it had been healthily on the market for at least a year. S&P 500, kudos to them, did not change their laws.

And so they have not invested in SpaceX. If your money is with S&P 500, it’s probably much safer than it is with NASDAQ, which is one of the index funds which did change its rules and did invest in SpaceX and probably will invest in OpenAI and Anthropic if when they launch their IPOs. To come back to your point, that means if, say, OpenAI does have some major fallout later this year, which seems likely with the amount of debt that they’ve taken on and the amount of money that they are consistently losing, which I have some numbers to share on that later.

If this happens post IPO, that means the stockholders will get a huge hit and lose a lot of money. It means people’s with retirement money and index funds in OpenAI will lose a lot of retirement money. And it seems highly likely that the government will try some sort of bailout for OpenAI, which would ultimately hit the entire country because that bailout would be coming from our tax money.

So it is a…

Or our debt liability.

Yeah. And it’s actually crazy how much of the current stock market is already margin debt. We have hit a brand new high in terms of the amount of money on the stock market that comes from like loans.

So basically like borrowed money on the stock market currently represents, I think, 4% of our national GDP. Which is a rate that we’ve never hit before. And frequently that number getting really high does come right before a crash.

Like we hit a new record of margin debt, GDP percentage right before the Great Depression. We had a new record before the.com crash, and we’ve hit a new record now. So that is genuinely scary.

And so it’s like, I’m not an investment consultant, but like I would genuinely look into like trading your money from Nasdaq to S&P 500, if that’s at all a possibility for, you know, listeners with retirement funds. This is not my area of expertise by any means, but just literally just based on like recent news alone, there’s some scary things happening.

And I didn’t think I was going to get investment tips out of this.

Me neither. It’s been a crazy thing to get into. It’s like, I mean, if I were like a betting man, if I were someone who like put money actively on the stock exchange, for the past few years, I would have been shorting OpenAI.

I would have been doing like Michael Burry from the movie The Big Short. And I say that not as an example, he is literally shorting AI companies right now. He’s doing it just like he did for the mortgage crisis of 2008.

So I would be like following his moves there because I see the same signs that he is. And yeah, but the government bailout situation is a very likely possibility because OpenAI and the government are doing so much work together.

Well, and there’s competition in terms of AI advancement with other countries, which ends up making it maybe a national security issue, at least in their minds, more so than you would think it would be.

Yes, that is true. China is doing very well in the AI and LLM department. Which, quick sidebar, and I do think this is important, and this is a slight critique on episodes 74 through 76 of the podcast.

I do think that, and this is not just a critique of y’all, it’s a critique of like our entire international discourse around the topic, is that we too frequently talk about AI as an umbrella term that includes a lot of things, and primarily, I think, we, on this podcast, but just in general, should be talking about LLMs, large language models, which would be like Deep Sea from China, Open AIs, ChatGBT, Anthropics, Clawed. So those are all LLMs. And also generative AI, which is more machine learning based, and so that would be image producing AIs, like mid-journey, music producing AIs like Suno and Udio.

That is the main object of concern. So like when you and Rusty discussed the scientific advancements that are occurring thanks to AI, like new proteins being discovered, or people finding cancer, or like radiology tests being performed faster and more efficiently than before, it’s not ChatGPT doing any of that work. You know, like those scientific advancements is primarily like machine learning.

It’s not large language models accomplishing that. Machine learning is an incredible thing. And when we talk about AI being a useful tool, we are usually referring to machine learning.

We’re not referring to chatbots. We’re not referring to image generators or video generators.

So can you help me to distinguish between these two in terms of fundamentally, and let me just say, this would be my initial approach, that an LLM is imitating intelligence, whereas machine learning is increasing the efficiency of computations?

Basically, yeah. So AI, artificial intelligence, really everything that we call AI has no direct connection to the historical definition of AI from science fiction, from the Stanley Kubrick.

A conscious, self-conscious machine or whatever.

Right, yeah. Self-aware machine, yeah.

There are people who are saying and selling the idea that like LLMs will get there, or that those are, that LLMs, like Chachipati, are the pathway to true self-aware and intelligent robots, machines and whatnot. And we have a lot of reasons to believe that that is not the case. But this…

Well, we especially have a lot of reasons, I think.

As Christians?

As Christians, yes. Because I think a lot of that hope is built on a materialistic worldview. Because if you’re a materialist, if you’re like Daniel Hofstadter from Godel, Escher, Bach, or anything like that, you’re trying to understand how does consciousness appear from inanimate materials, right?

Like how does life occur randomly from the interaction of inanimate things? Now, so like if you’re thinking, well, that actually did happen with human beings, because you’re not taking into account, you know, like the amygdala or the transmission of the Spirit of God or any of these kinds of things, you’re not, you’re like removing Spirit from it all together, then you have to believe that if it happened with human beings, that random inanimate objects came together and collided into consciousness, why wouldn’t it be the case that a directed process of inanimate objects colliding could produce consciousness as well at some future point where we cross over that limit or whatever? It’s connected.

And there was a famous atheist, I can’t remember which one, I want to say Richard Dawkins, who recently came to that conclusion that Claude must be self-aware and why can’t it be self-aware? Because humans used to be inanimate objects and now we’re animate. So that is a plausible possibility from an incorrect world view.

Exactly. Your presuppositions are wrong, and so you’re coming to very unhealthy conclusions that are leading the entire human race astray because you have too much power. I generally conceive of AIs in the huge umbrella term, especially in terms of what people are selling to us and the services they’re providing to us, usually in three different categories.

One being advanced automation, which really isn’t AI at all, but there are a lot of companies who have advanced automation, they are now branding and selling as AI. Then there’s machine learning, which is like the oldest form of what we’re now calling AI. I think it dates back to the 80s, and as you said, it’s basically advanced computation that actually allows us to do really fascinating things that might not otherwise be possible.

I think it was in episode 75 that you referred to a statistic that about 87% of music producers use some form of AI in their process. And I would suggest that, like me, some of that is machine learning tools and plugins that are now being labeled and branded as AI but weren’t previously. There’s a company called iZotope that’s been around for years, and they make solid plugins, one of which that I use a lot is a D-Reverb plugin.

That kind of listens to your audio for a few seconds, and based on what it’s hearing, begins trying to suppress or remove the reverb that it hears in that audio file.

Additional space.

And I literally have no idea how I would accomplish that without a machine learning tool. It is extremely helpful and is technically AI because it’s now branded as AI, even though the technology and the tool has been around since before the AI boom of late 22. But then there is large language models, which is this completely different thing of taking this huge data set, just terabytes and terabytes of words and data, reports and conversations and everything else, and feeding that multiple times over into a language learning machine, so that it becomes this trained model on not only how to talk like a human, but then how to access information from that data set.

So the P in ChatGPT is pre-trained. It is a language model that is pre-trained on unbelievably vast amounts of data, so that when you are having so-called conversations with a chatbot, which is effectively like a really advanced predictive text meets a really advanced like Siri style voice modeler, that’s like the face of it. It is using this very advanced like word retrieval functionality to understand your prompt in relation to its vast training data, and then to pull from its vast training data to respond to your prompt in a way that, based on all that data, seems like a valid response.

So like people joke about it being like just a really fancy predictive text, and that really is like the process that’s taking place. Like one little word at a time is just predicting what it thinks would be the proper thing to say to you based on what you asked it. And so generative AI in terms of like Suno and music, which I think we’ll focus on later in this episode because so much of what Renew the Arts does is based around music.

Yeah. Especially like our porch light house shows. That actually falls more under the machine learning category.

What services like Suno is doing is basically it’s taking like just a strip of white noise, like three minutes of white noise, just all frequencies blaring in the full sound spectrum. And then based on your prompt for a city pop ditty or an alt rock banger, and it goes through its training data of like, okay, what does city pop sound like? Or what does alt rock sound like?

And it starts carving away at the white noise sample by sample to create the facsimile of what you asked for based on the examples provided by the millions of items in its data set that align with what your prompt requested.

Which are all to it just image patterns.

Basically, yeah.

Yeah, because it’s looking at like wave files and saying, this is what the wave file of this looks like. This is what the wave file of this looks like. And this is what the wave file of this looks like.

And all of these are in the same genre, and they all have these frequency patterns in common. We’re going to go ahead and leave those in there. And then here’s what these frequency patterns would look like if it was this word that was being said, or this word that was being said.

And I would assume at this point that they probably do it on an instrument by instrument basis. Maybe not.

But they do have tools that do that.

Yeah, again, another tool that is very useful and is classified as AI is stem separation.

Yeah, which is incredible.

Which takes an existing piece of music and you can isolate just the drums, just the vocals. Yeah, super useful for a lot of purposes.

It’s especially useful for musicologists and things like that. Because I mean, I’ve seen some breakdown videos of songs where they’re able to use these tools in order to show you what the isolated tracks of various things are. And I’m like, this is, that’s really awesome.

Like it’s just an interesting music history tool.

It’s also like, if you want to practice singing a song, you can use stem separator to remove the original vocals and sing along to the track, or to isolate the original vocals to be able to better study and learn from the vocal track. You can remove guitars from something so that you can practice playing along and playing the solo yourself or like whatever. So yeah, so a lot of cool things there, but the Suno side of it, the music generation side of it, I assume like it has one separate process for creating vocals and then a separate for creating the music, because the vocal modeling at this point has gotten a lot better.

And you can also like pick certain voices or ask to emulate certain singers. And so that’s kind of a different training process per se. But yeah, that would be like the overall categories of AI as I best understand it.

Okay.

With again, like for the rest of this conversation, as we just use the umbrella term of AI, which we will inevitably do, we are almost certainly referring to generative AI in regards to music, art, or referring to LLMs, unless we say otherwise. And so I do think that’s a bit of a helpful thing because there are ethical concerns regarding generative AI and LLMs that we shouldn’t have regarding machine learning in general.

Okay.

They just don’t apply. They’re just categorically different, and lumping them together as AI creates a lot of categorical confusion.

Sure. That makes sense.

Thank you for listening to the Renew the Arts Podcast. If the conversations on the Renew the Arts Podcast have been helpful to you, would you do us a great service and please rate and review this podcast on whatever platform you listen to them? We would greatly appreciate it.

And if you would like to partner with us financially to bring the goodness and beauty of Jesus to your community, please visit renewthearts.org/donate.

More general industry news. So a few large corporations, including Uber, Starbucks, Pizza Hut, Klarna, and others, have significantly scaled back their AI use within recent months. Some of these companies have ended AI initiatives entirely.

Uber hit headlines for having a 2026 budget for AI use that they burned through by like March. And I think it was the president came out to say like, we’ve seen no ROI here. We’re not going to keep spending money on this.

It’s not doing anything for us.

What were they even using it for?

That’s exactly it. That is a very important question that a lot of companies need to be asking. Because again, there’s been this like narrative throughout the industry for these past four years that like you will make your company so much more efficient, so much more profitable by using AI in as many places as possible.

So start paying for Chachi Petite, start paying for Claude and figure out how to best use it in your company so that you can become more efficient, more profitable. You can have more robots and fewer humans. But like in the case of Uber and many other companies, there’s been that question of like, we don’t even know what we’re using it for yet, but we’re not making any money back on it.

So why don’t we stop? In the case of Pizza Hut, they were using it for like a delivery management system that lost them millions of dollars. In the case of Starbucks, it was an inventory management system that was so consistently getting items wrong that it was completely useless.

For Klarna, they completely fired their entire customer service team and replaced it with AI chat bots. And the chat bots could answer simple questions, but anything like more complex or that really like needed a human, it would just stop responding, and you would never hear back from the AI agent. And so it was a complete PR disaster.

And so that’s the kind of stuff that’s happening consistently. Even Microsoft, the megacorporation behind its own AI service, Copilot, has scaled back its internal reliance on Claude Code, which is the LLM coding service provided by Anthropic. So across the board, this is about a year old data-wise, but last summer, I think it was last August, MIT provided a pretty thorough study saying that, of all of the companies that had put significant money into AI or had tried incorporating AI into its regular business ventures, 95% of those companies had zero ROI, zero return on investment, that most companies are, as of now, seeing no value in their attempts to incorporate AI into their businesses.

Is that because of bad implementation, though? I mean, I guess that’s what the AI companies would say, is that you’re not using it very well.

It is what they would say, isn’t it? There’s kind of a… There’s an interesting rug pull that OpenAI is doing right now.

Again, OpenAI being the company that created Chachi Petit and its president or CEO, whatever, is Sam Altman, who has become kind of a highly known public figure, often seen sitting right next to Mr. Donald Trump, which I have feelings about. But, so OpenAI, right, is one of the main companies that’s saying like, you can fire your employees. AI will do their job for you.

I think a few years ago, they were saying that we’re like a year away from AI being able to do most jobs. They keep on pushing out the gold dates, obviously. I think right now they’re saying we’re 16 months away from AI, taking over most jobs.

And they’d say most jobs on a computer, like white-collar mostly jobs.

Well, I mean, I think most jobs are computers at this point. But yeah, let’s…

Well, I mean, trades.

I think the trades are pretty…

Yeah, plumbers are safe for now. But yeah, I just tried putting my laptop in my toilet because it did not work. Only Claude.

Claude. Claude. Yay for jokes.

I think estimates are, again, from these companies, that there are 100 million jobs in the US that AI should be able to do already or soon. So for the companies that have laid off or fired a lot of people under the guise of AI can do your job for you, OpenAI has now been hiring, I think they’re called deployment agents or deployment experts or some sorts. Basically, temp workers to come to your company to help you to deploy AI better and better use ChatGPT to do the work that humans were doing just fine before you laid them off.

Which is this incredible way of just doubling up the money that you’re paying OpenAI. To say, like, we’re paying for ChatGPT and now we’re also buying your temp employees to help us use ChatGPT better. I’ve not heard of that necessarily working.

Again, the ROI for most companies is still not there. Recent studies say that about one-third of companies who did mass AI layoffs have already hired back more than half of the number of employees that they laid off and also a significant amount of those companies have paid more money in rehires than they saved from trying to use AI instead. So it’s just really not turning out well.

Meanwhile, and this is, again, kind of a big topic that we’ll touch on later in the conversation, I’m sure, but anti-AI data center protests have been occurring increasingly throughout the country. Just a few days ago on July 18th, apparently protests occurred in over 125 locations across the US. According to a website called Data Center Watch, at least 75 data center projects worth approximately $130 billion were blocked or delayed in the first quarter of this year.

But at the same time, like the four or five biggest companies in the country, like IBM, Microsoft, Meta, Google, are combined planning on spending, I think, $720 billion on AI infrastructure within the year of 2026. So again, this is a huge amount of our economy, of our national spending. Are these mega corporations planning on spending money on the infrastructure that’s necessary for these extremely energy-inefficient AI models to be able to continue running, as well as their projections on how much data and energy that they think they’ll need as more and more people, more and more companies start using AI as much as they want them to, which is another issue of people over-projecting the energy requirements of AI.

But that’s also something I want to get into later. So I do have some company-specific news for just a few of the big dogs. Do you have any questions or thoughts before I get into that?

I’m wondering, this sounds pretty like not good for the future of AI, but also a lot of the people who are pro-AI or whatnot would just say, this is kind of early stages, growth pains kind of stuff, and we’ll be able to get through this, and those people that have trust and faith in it. Because people said similar things about cryptocurrency and the early adoption, and then the quick fall off from that, and then the thought that this isn’t really going anywhere, and then all of a sudden it’s stabilized.

Yeah, kind of. So Bitcoin itself, stabilized. It has moments of being worth a lot more, being worth a lot less.

But there are very few cryptocurrencies with any legitimate market value behind them. The cryptocurrency market is still filled with rug pulls, with snake oil, with new coins being launched that end up being worth less.

That’s literally the case with any newly implemented… I mean, again, you wouldn’t say that just because of the dot-com boom, that that meant the end of the internet as a business model.

Right. And that, again, I hope that our conversation can cover the scope of that. Because, like, my point in sharing this news is not to encourage doom and gloom or to make listeners confident that this is definitely a bubble and it’s all gonna pop and we are just months away from AI being no more.

AI in one form or another is here to stay. I do think these LLMs are especially precarious situation. And I want to discuss later about how, of all the forms of AI, I think LLMs are the ones that we should be the most wary of and the most careful in our use of them.

And the astounding trouble that these AI companies are going through in the markets and in the news right now, I think, points to the fact that Chachipati and Claude and such might not have long-term futures, even though they are the exact companies that are looking at trillion-dollar IPO launches later this year, if they’re able to make those happen. So it’s a lot of eggs being put in the baskets of a few companies that have the worst odds and the most cost-inefficient services they’re selling. Which, actually, I think is a good segue into talking about Anthropic.

So, as far as the US goes, Anthropic is the clear winner of the AI race. Its models are consistently leagues ahead of competitors. Obviously, the big competitor is Chachipati, but you also have Google’s Gemini, you have MetaAI, which I think is called Lama, as an LLM joke, L L A M A, and other smaller companies with their own.

But Anthropic has consistently been ahead of all of them, with its latest frontier model, Claude Mythos, being deemed so powerful that the US government stepped in and shut it down. First, making it inaccessible to private users, but ultimately making it inaccessible to enterprises and companies as well. So, just off the market entirely, except for maybe the government itself might still have access to it.

It was basically so good at finding security lapses and finding ways to break into systems and governments and everything and whatnot, that it just couldn’t be allowed into the hands of the general public. But reportedly, the fact that the US government could step in and shut down Claude Mythos caused Anthropic to lose many of its overseas clients and contracts due to fears that like, why are we working with this company and making our business dependent on their software if the US government can just step in and take it away from us? And so this also looks as like a strikingly hypocritical move for the current US administration, since to this point, the administration has been pushing policies to keep AI from being regulated.

There is a federal mandate that came down that the states could not regulate AI by themselves, that AI regulation had to come at a federal level, and so far that has meant no regulation at all, with the exception of stepping in and closing down mythos. And this might also be an intentional slap on the wrist after what was effectively a big falling out between Anthropic and the US government this past February, where Anthropic had this like $200 billion contract with the US government to do all of this work across our government, with the exception of mass surveillance and automated weapons. That was a hard line that Anthropic drew and their contract was based on like not crossing that.

And then the government apparently changed its mind, said, no, we’re actually going to need to start using Claude for mass surveillance and automated weapons. To which point Anthropic pulled out.

And honestly makes a lot of sense. It’s like, that seems to be the thing you would use it for the most. I mean, I remember, this is kind of a side topic, but I remember right after 9-11, with the whole Patriot Act and like the mass surveillance thing that started there, that they’ve never decreased.

They’ve only increased the amount that they surveilled the citizens really. But I remember, there was some kind of a former director or something in the FBI. I can’t remember who it was exactly, but I do remember what he said.

He said, there’s so much information being collected that finding anything relevant in it was like a needle in a haystack. That it was like there was so much information that it was almost useless, because searching through it was basically impossible. Well, recently, when I was thinking about the capacities of AI and the amount of data it’s actually able to synthesize in a short period of time, I was like, man, it’s only a matter of time before the civil government takes AI or LLMs and starts to use it in order to filter through all of the conversations that they’re recording, the billions of hours of conversations that they’ve got.

Anyway, so I’m sure that they were salivating for that. And then Anthropos felt like, except for that, except for the only thing that you probably want to use this for. Oh, and weapons, that too, not that.

It’s like, oh, well, yeah, then.

Well, I appreciate Anthropic having the ethical backbone to say no.

Somebody won’t, though.

There have been a handful of stories over the past few years that show Anthropic to not just be like the better company that’s making better models and better products, but also doing so with a more human rights-oriented mindset. So if we’re forced to pick between the two, we should always pick Anthropic over OpenAI. But I’m worried that if, as we discussed earlier, if they both have a bubble bursting moment, I’m worried that now the US government will bail out OpenAI and not Anthropic.

And thereby own it.

That will, well, they’re already in cahoots. Here’s the thing, Anthropic said, no, we’re not going to do your weapons and surveillance. And so they lost the $200 billion contract, and OpenAI immediately stepped in and said, we’ll do it, we’ll do your weapons and mass surveillance for you.

We don’t have a moral backbone, we don’t have a spine. So they stepped in and picked up the contract, which ultimately was like a PR nightmare for them. And it has turned the kind of public view for a lot of people who are watching the news on this topic to be like, Anthropoc are the good guys, OpenAI are the bad guys.

Even if you just like watch like their main guys, like Sam Altman is a notorious liar, has been with every company that he’s worked for. Meanwhile, Dario Amadei with Anthropic definitely like has his head screwed on right, seems to be far more like thoughtful and concerned about like AI actually being a good thing for people. Again, even though I think Claude and just LLMs in general are a net negative for society and for the world, for reasons we’ll get into later, and like my intentions with this conversation again are not trying to convince you to never ever use these things, but to think about why and how you use them, and to use them as carefully and ethically as possible.

I will say I think anyone with the choice between Claude and Chad GPT should always use Claude, if that’s the choice you’re making. But here’s, okay, so this is really crazy. And somehow Anthropic is getting far closer to being a profitable company than OpenAI is.

OpenAI is continuing just to race toward insane amounts of debt. Apparently, Anthropic is getting closer to a break-even point, even though they still carry a lot of debt. But this is a very crazy thing, some numbers that I learned recently.

So there was that small period of time where Claude Mythos was available, that being the frontier model that the US government shut down, this incredibly powerful LLM. And the primary subscription cost for standard users was $200 per month. Now, I do think there was a cheaper option, but the main level was $200 per month.

But in terms of actual use cost for the company, that $200 subscription purportedly gave users access of up to $8,000 of general token usage, plus up to $14,000 of Claude code. So if these estimates are accurate, that means Anthropic would be losing potentially over $20,000 per user per month on any user that like maxed out their usage limits, which just points to the dangerously unsustainable subsidies that these frontier models are utilizing over to gain users. So all of these companies are participating in this very dangerous business model of basically like offering their products at a subsidy so that you fall in love and you become dependent on using it, or your company becomes dependent on using it so that as they steadily creep up the prices, you’ll stick around because you have to.

And we saw a really crazy example of that earlier this year as Anthropic began charging its enterprise clients with the real token costs. So if you’re a big company buying Claude, you no longer get the $200 subscription. You have to pay per token.

Now granted, a single token of data that your LLM is retrieving or processing is just a fraction of a penny, miniscule. But a single prompt can be thousands if not hundreds of thousands of tokens, depending on the length and complexity. To oversimplify it, it’s like a token per word of just understanding your prompt, of just taking it in.

And then a token per word of looking through the rest of your conversation history to understand that. Because basically, it’s not really learning new things. Again, GPT means pre-trained.

So, whenever it’s needing to refer to things you’ve already talked about or parameters you’ve already set, I’m pretty sure it needs to go over that again every time. And so, that’s new tokens being spent, just look at your history, look at your parameters. And then, all these tokens spent on looking through its data set to retrieve the pertinent information regarding what your prompt is asking for.

And then, tokens in order to write the answer. Just insane amounts of data processing. And companies were not happy to figure out how much those real costs were when they started receiving token costs instead of the subsidized subscription costs.

And one undisclosed company, that we don’t know who, but one company apparently forgot to set limits on their token usage. And in a single month, racked up a bill for Anthropic of 500 million dollars. Half a billion in one month, because they token maxed so hard.

Which again, it’s like, yeah, Anthropic’s the best there is, but there is no explaining how expensive it costs. And even as these tools are getting increasingly powerful, and are able to pretend to be human so well, and pretend to understand so well, and to accomplish more advanced tasks so convincingly, they’re not getting more efficient, is the thing. They are improving in quality, but also just adding more brute force, taking up more energy, more GPU usage, more token usage.

So that all connects us to the energy infrastructure and…

Yeah, it’s interesting. My plan for the episode was to just blow through this news piece just as context for the rest of the conversation, but instead, it’s kind of providing us the detours and the tangents to hit a lot of the things that we’re going to talk about later. So yeah, we can hit the environmental energy piece now if you want.

I want to ask you a question about… So in terms of the cost, the token cost, how much of that token cost do you think is profit percentage-wise? And how much is covering the cost to Anthropic for its processing?

I mean, I know that’s oversimplifying it, but…

Yeah, so my understanding is that the enterprise costs is not really a profit situation for Anthropic, that it’s literally just covering the costs. There’s a chance that they put a small margin on top of that, but they’re already running the risk of losing all their enterprise clients that are using them a lot because the actual token costs are so insanely high and so much more than the subsidized subscription costs that they used to pay.

Are you saying that they’re basically just breaking even on this right now?

With large companies that are getting the real cost, it’s more of a break even situation than a profit situation.

But they’re still already in…

Well, for the smaller companies and for the personal users, the only time that they ever profit is for someone who is paying for a subscription that they’re not using very much. Very much like a Netflix situation, like you pay your $10 a month, $18 a month, whatever, but you never log on and watch anything that month. That’s straight profit.

Whereas for a regular Netflix user, every time that you go in and use the data, and use the energy of watching a movie, then that cuts out of the profits that they get from your subscription. So for every subscription-based model in the world, basically, except for my friends club at friends.chasetremaine.com, it’s based around the hopes that the person doesn’t do anything with their subscription. And I know there are, so in my mind, a good subscription is one where you’re actually getting something.

There are subscription models like we used to do, I think Hungry Root is what it’s called. It’s like a box of groceries to your house every week. And so you’re paying a subscription, but you’re getting exactly what you’re paying for.

You are receiving a good. And so subscription models are scary because it’s like, if you pay 200 bucks a month and then forget to ever log on to do anything that month and that’s $200, gone. Profit for the company, total loss for you.

And so much of our market right now is based on subscription models that have no direct tie to what you’re necessarily receiving. But again, it’s like these token costs are so high, and these subscription models for these companies are so heavily subsidized based on loans and debt and venture capital that the whole point is to lose money, but to gain life long customers.

Right.

So that when $200 a month becomes $2,000 a month, you won’t be able to say no because…

Because you’re so dependent.

Right. Because you’ve fallen in love with your chatbot, or because it’s your therapist, or because you know…

You can’t work anymore without it.

Frames your entire startup company around using Claude.

Right. So it’s like drug dealers basically.

Yeah, and there is a level of addiction, which again is something I want to get into later. There’s like… There’s a level of like…

Well, I’ll say it right now. There’s no point. There is a level of like gambling methodology that we’re starting to see in like all corners of our lives, especially digitally.

For one, like there’s like prediction markets now that are using a loophole in order to be legal in states where gambling is otherwise illegal. But like the real addiction of gambling is just the hope that next time you’ll get the right thing, that next time you’ll get what you want. And so you pull that lever, and you can’t wait to see, you’re imagining, you know, like the gambling, like the slot machines that’s like hoping to get three images in a row.

That picture is what so much of our like digital world is now based around.

So it’s like a dopamine loop.

You talked about on episode 76 with Rusty about how you had to give up Instagram, because for basically that same like gambling reason, it’s way too easy to fall into the whole of just watching reels. And the social media equivalent of the lever you pull on the slot machine is like pulling your screen down, releasing it, and getting a refresh of new videos. It’s like maybe this next video is the one that’s going to make me laugh, the one that’s going to make me cry, the one I’m going to want to send to my friends, the one that’s going to fix my marriage, the one that’s going to inspire me to quit my job and follow my dreams.

That refresh pull is literally like the cranking of a slot machine lever. And you have very much a similar dopamine mental experience occurring when you put in a Chachibutti prompt. Be like, what’s it’s going to give me this time?

What is it going to tell me? Or more directly, when you hit, like, generate on Suno. Sure.

Like, what’s the song going to sound like? If you hit generate on Mid Journey, be like, what’s the image going to look like? I’m so excited to see it.

It’s going to be different every time. Oh, that hit. Oh, that rush.

Like, that’s really like how it just sinks its teeth into our brains, literally. Yeah. So that is like a scary thing that we need to be wary of in terms of our use of it.

Well, and it is our use of like things on the Internet in general.

Yeah, we discussed this briefly in previous episodes, but they’ve done studies of people who use ChatGPT for essay writing. And there was that major study that they did where they gave people decreasing access to it. Or it’s like one person wrote an essay with no helps, one person wrote it with the Internet, one person had it with limited ChatGPT, and one person had it with like full on ChatGPT.

And the people who used full on ChatGPT not only had less brain activity active because they had them all wired up or whatnot, while they were writing the essays, they also had decreasing brain activity over the course of the study. Which is, that’s sobering.

You know, it’s like.

Yeah. So, what next, Chase? So, is that your?

We were about to go down the environmental.

Okay, do you want to go down that, or are you finished with your news recap, or is there more news?

I was finished with my anthropic news recap. Oh, news recap.

Okay.

So, we can do this tangent before we swing back around to OpenAI.

Sounds good.

Which I think will actually be helpful, because a lot of my OpenAI news is based specifically around data centers.

Okay.

And there is a lot of discourse and confusion around data centers right now. There’s a lot of people who think data centers are inherently good and necessary. A lot of people who had never heard of data centers before and think that they’re just something that we’re making now to facilitate AI.

And the answer is somewhere in the middle. That data centers have been around for a while, but the existence of data centers is how we use the Internet. It’s the physical computer that is actually doing the work of the cloud.

When you think of you’re uploading things to the sky or just to the air. No, it’s actually going to a data center. Every time you load a song on Spotify or a video on YouTube, data center is doing the work for you to pull things.

So data centers is how Netflix works. It’s how TikTok works. TikTok in particular is probably the worst offender in terms of the amount of energy and strain that gets placed on data centers because of just such a huge amount of people watching and uploading videos.

And constantly.

And constantly, yeah. But at the same time, as much as data centers are a necessity for the digital age, we are seeing this new type of data center being planned and built that is on completely brand new scales. Even though the people who are trying to build them or who are supporting them or funding them are often referred to as hyper scalers, which I think is a fitting name because we really are scaling this at a rate we’ve never conceived of before.

And so people are planning data centers that are the size of a small town, the size of a shopping mall. Like that is not a normal data center. That is not the type of data center we’ve been building and maintaining.

Not to mention, I could be a little wrong here, but I think most data centers run on CPUs. And one of the great advancements of the LLM technology is when they started using GPUs instead, which are graphics processing units. And basically, it’s like, what powers your computer screen, what powers your video game consoles.

It provides just an extra amount of processing per chip that allows LLMs to run at a fast and sustainable rate.

And I guess taking up less space.

But they cost a lot more. And they also like, they heat up very easily. They have to be cooled immediately in order to not be damaged.

And with regular use, they deteriorate within a few years. That’s why a lot of gamers are consistently buying new consoles or creating new gaming computers. Because GPUs are getting worse.

The ones they own are getting worse. Meanwhile, the new ones are being made that are higher quality, can run better, and that you might need in order to play your new game. Right.

At the highest frame rate. So massive amounts of GPUs are being placed into these giant data centers, which can get so hot that it raises the ground temperature of the nearby area by up to four degrees. Which is insane and has its own ecological impacts.

Like the surrounding area will literally all get hotter. And so it affects the temperature, the air, it affects the local water supply. There are people who live near data centers that are experiencing like just like they’re running water, not running, or it comes out brown, just because of the way that these data centers have such an effect on the water supply and the amount of water that it’s taking from the local water supply in order to do the cooling.

Thankfully, water is not the only way to cool data centers. And it isn’t even like the most efficient way. The issue is that it’s the cheapest way.

And so people keep building data centers that use local water municipalities in order to cool them. And it’s a big problem because the water that they’re using to cool these GPUs is combined with basically coolant. And so you’re putting like these PFAs in there, these like forever chemicals, that effectively like poison the water.

So it’s like you can keep using the water over and over again to cool your GPUs if you build your data center well enough to basically circulate it. You heat up the water, right? You take it somewhere else to cool it down and bring it back to heat it up again.

If you can create an endless loop, then that is a better way of using it. But again, like looking for non-water ways to do it is probably what the way forward needs to be, among other things that need to change with the current plans. But you’re permanently taking water out of the local ecosystem, which again is like, I think, another point of confusion for a lot of people, because another talking point, especially for people who are pro-AI or pro-data center, are saying things like, the corn industry uses way more water than data centers would.

Cows use more water than data centers would. Yes, those are true. There are a lot of things in our country that use a lot of water.

However, watering corn keeps water in our ecosystem. Yeah, it’ll evaporate, it’ll return to rain. Watering cows will return to rain.

Cooling data centers makes the water undrinkable. And some of these companies, they’ll take water, it’ll absorb the warmth from the GPUs, and then they’ll go off and they’ll evaporate it, and just send this poisoned water back into the atmosphere and take fresh water to cool. It’s so stupid why they would do that, but it just like…

Well, are they assuming that the additional pollutants are left behind as deposits when the water is evaporated?

I would hope so. That might be the case. I would need to actually research that, because that could be possible.

Yeah.

But I do know that a lot of these chemicals that data centers add to turn the water into a coolant does not come out if you send the water back to like your local treatment center.

And I mean, it could be that you’re vaporizing various different toxic compounds. But if these are solids, like for instance, if you think about like salt water, if you evaporate salt water, it’s not like you’re putting salt into the atmosphere. It’s just left behind.

However, some of these chemicals are able to be vaporized, which means that they would be exiting out into the atmosphere in various different ways. Right, so that would be the question I would have to ask.

Right, worth looking into. But it is a very serious concern. And there are so many things in our society that like overuse water right now, that there are genuine concerns that we like as a country are heading toward a place of being extremely limited on drinking water.

Well, you’ve seen this happen already. I mean, this is not a speculation or an assumption. You have seen this happen already in China.

For instance, where Shanghai has had a major drinking water problem for a period of time, because a lot of the pollutants that have come into the water makes certain sources of water undrinkable. One of the largest freshwater bodies in China, I think, is so polluted that it doesn’t even sustain fish or plant life. So, you know, and like, I think there was talk, maybe it’s like 10 or 15 years ago in Shanghai, about trying to pump waste water down into the groundwater, which again, because they didn’t know what else to do with it.

So, yeah, for sure, there’s major issues with all that kind of stuff. And, you know, you can talk about the use of water and fracking for natural gas extraction, and the way in which that kind of removes a lot of water from its capacity to be used in other ways. And the use of pesticides and herbicides and glyphosates and all that being in runoff water that ends up even in drinking water and, you know, et cetera, et cetera.

So there are a lot of things we’re doing to water right now that are not great.

No, right. Just because we’re surrounded by oceans doesn’t mean it’s going to be easy.

No, no, no, like I lost the drinking water. Yeah, ask them, you know, Middle East, how well that’s gone. And I mean, you know, like, I mean, I guess if we figured out a desalination that was actually affordable and sustainable, that would be usable more so.

But we haven’t. And so.

Well, maybe we should put $720 billion into that.

Into that, instead of open AI.

Instead of these infrastructures.

I agree. And maybe we should also be talking more about that and sustainable energy as opposed to global warming. I kind of feel like the issue with water, which is a much bigger and much more significant local problem, it’s gonna be shifted over into a conversation about global warming, which has almost no solutions and…

Yeah, and even like speculation that the carbon footprint issue isn’t nearly as bad as people fear it might be. There is some like ecological evidence that there have been periods in Earth’s history where like carbon dioxide levels in the air were so crazy high, like way higher than they are right now, and it didn’t cause the world to be destroyed. If anything, carbon dioxide feeds trees, doesn’t it?

I don’t know. It’s what the plants crave. So, but what we could do, I think, even if you’re like, okay, I think global warming is a major issue, it’s like fine, but would you say that the more existential threat in the near term is our water issue?

Why don’t we focus on that? And it’s just interesting to me the way in which a lot of these things are selectively addressed. It’s like everybody, you know, had that huge electric car boom for, you know, that was basically allowed by Big Auto because they were kind of stifling that for a long period of time.

And then all of a sudden you had that boom that occurred and you’re like, yeah, but we don’t need electric cars if we’re producing electricity in the same way we always have produced electricity. So even people currently driving like Teslas or whatever, and they’re like, I’m doing what I can for the environment. It’s like, well, no, because the grid that you’re pulling power from is still petroleum and coal based.

So you’re still dependent on fossil fuels for the most part. And so in those terms, it’s like they only allowed electric cars to be a thing to solve people’s conscience about global warming or the environment or whatever, when it doesn’t actually move the needle at all. If you continue to pull power that’s being produced in the same exact way through fossil fuels.

And it’s also probably a total coincidence, probably completely coincidental, that around the time when a lot of the oil companies and the auto companies started to invest in natural gas and natural gas power plants was when they kind of quote unquote, it was around the same time that electric cars became a nationally subsidized and big pushed thing. Obviously the people in positions of power typically are not going to get rid of their avenues of power. And that’s a big issue anyway, that I think touches on the AI thing as well.

But anyway, tell me more about the power grid situation.

Yeah, so there’s all of these like immediate local concerns with one of these new, like these especially like large, intense GPU based data centers causing the water pollution, air pollution, sound pollution is a big thing too. Like they are incredibly loud. And so anyone who lives near one, like it’s like almost unlivably loud, 24/7

If you’re close to one here in Nashville, there are people trying to build one right next to our zoo, which would be awful for so many reasons, both the people and the animals that live nearby. Thankfully, we’ve got people, I think our mayor is stepping in to shut that one down. But there’s also an extent to which like, whether it’s like Kevin Leary or Elon Musk, or Sam Altman, or Mark Zuckerberg, planning all of these massive data centers around the country.

The fact of the matter, plain and simple, is that our energy grid can’t handle it. We touched on this very briefly earlier, but one of the reasons I think China is almost definitely going to win the race long term regarding LLMs, is that they have the energy grid for it, plain and simple. And we don’t.

I think of all the types of AI out there, LLMs take up so much energy and so much data. And China, being that it modernized decades after us, were able to create a grid, an energy grid, with far more advanced technology. And they can handle just so much more energy than our little power lines plastered up on pieces of wood across the country.

And the strain on our power lines would be incredibly immense, and just not impossible. So the prediction that some of these AI CEOs and whatnot are saying that AI is going to take all white collar jobs within 16 months, the absurdity of that is that that would be like 100 million jobs. If we assume that that’s like one GPU dedicated to per agent per job, right, so like 100 million GPUs in data centers across the country would be like multiple more New York cities of energy being required at like a constant always on every day rate.

It’s just like literally mathematically impossible. And the amount of like new energy plants we would need is also pretty great. And you can’t throw those up overnight.

Like those those take usually like five to ten years to plan and build. So like, if this is where our country, this is where the human race is headed, America will need a brand new power grid. And so the fact that people are putting money into data centers, instead of to like a new like infrastructure that can actually support the data centers and make them more sustainable, and to give us time to figure out ways to make data centers not be such a strain on the environment and the local municipalities and whatnot.

It just shows kind of how backwards everyone’s like heads are at and how bad the planning is. But for the people who are already close to these data centers, they are having to help foot the bill of the massive strain that is being put on local energy grids. So this is especially bad in Virginia and West Virginia, where a lot of these new AI-based data centers have already been built and are currently running, where just people’s energy bills have skyrocketed.

And it’s just because they happen to be unlucky enough to live near one of these data centers, where you’d think trillion-dollar companies would be able to foot their own energy bills, but no, it’s being handed to the American people.

Well, is it possible, though, that the energy company, because of an increased demand, is able to increase prices?

It’s a mixture of both.

Okay.

But the increase in price should go to the thing that is increasing the demand.

I mean, that’s just not the way the market works. So I mean, I remember hearing about how quinoa became a craze, and because of the demand for quinoa, it actually made the price of quinoa so high that the local Ecuadorian farmers who were making it couldn’t afford to eat it anymore, even though it had been their traditional diet for like thousands of years.

That’s so sad. But to my understanding, like the hike that some of these people are experiencing by living close to a data center is a greater rate.

Like a municipal subsidy, basically.

Yeah, it’s not just like, oh, energy costs more now. We and our neighbors are like sharing the bill. It’s like, I could like use as little energy as possible in my house and still see a huge increase.

Right, exactly. Yeah.

Because like we are being affected beyond like what our own personal house meter or what our energy meter reads.

Totally. And there’s so much more going on with that too as well, because there’s all the stepping stones of inflation as they slowly try and matriculate that into your costs without causing like a universal panic.

Of the dollar to become worthless.

Exactly. So anyway, so yeah, it is a big mess. This feels like a scenario where we’re driving pretty hard at a cliff.

Yeah.

And again, it’s like AI totally aside, just the debt spending and like all the things you were talking about earlier in terms of debt as a percentage of our GDP. And the fact that China seems to be better suited or better situated to adapt to the new conditions. And obviously, they’re in a better economic state than we are in terms of like their own savings versus debt ratios and things like that.

And the American Empire as an empire, we have experienced about getting close to 100 years. And that’s about around the time when most empires find their way into obsolescence, so.

The slippers start walking down the stairs.

Yeah, yeah. We’ll see about all that. But this has been really interesting and informative.

So thank you for that. So tell us a little bit about the, what are some news for the evil open AI people? The supervillain.

The supervillain. And we have a sub-villain as well. We’ve got some meta news.

Meta being like Facebook and Instagram. They’re in this fight too, and it’s kind of been a big embarrassment for them. So we’ll go over open AI and meta now.

So open AI, a lot of people, when they think about AI, when they think about LLMs, they just think chat GBT. They’ve basically become the Kleenex of the industry, despite not really deserving to be that. Nevertheless, they did recently hit one billion active monthly users.

So along with being like what most people think of when you talk about AI, they are the most used model, especially amongst Americans. Yet, they are losing so much money. I think on an average customer, they lose an extra $1.22 for every dollar earned.

So for like, if I as a customer am paying a dollar, their cost averages like $2.22, something like that. So that’s not a very good sustainable rate. But in terms of like overall, their financials leaked recently.

And apparently from these leaked budget reports, they lost $5 billion in 2024, and $38.5 billion in 2025. Just an insane amount of loss. They have projected losses somewhere between $14 billion and $28 billion for 2026.

And that’s on top of 2026 being the year where some of their like oldest loans and debts are apparently coming due. So there’s some speculation that their specific race to launch their IPO and get a bunch of new money into the company is so that they’ll be able to pay off some of those earliest debts and loans. It’s just crazy.

There’s one of the most like debt saddled companies I’ve ever seen in my life. And they’re really good at figuring out new ways to get more debt. For example, they created a video generating platform called Sora, Sora AI.

It evolved into Sora 2 before getting shut down a few months ago. The use of the Sora app to create dumb videos, it was usually used to create memes. It was used like Spongebob being pulled over by police, or Sam Altman going into a grocery store and stealing food.

This app was reportedly costing them about $15 million per day in base compute costs in order to create these fully generated AI videos. Meanwhile, between free users and paid users, Sora only generated $2 million in lifetime profits. Not lifetime profits of $2 million versus $15 million per day.

So just insane levels of loss on this app that was mostly considered as a joke. I actually considered downloading Sora to just generate videos ad nauseam, just to help bankrupt them. And I would have done that if it weren’t for my understanding of the environmental costs that we just discussed.

I didn’t want to contribute to extra strain on the grid and extra water use and all that stuff. Otherwise, I would have loved to help troll them into bankruptcy. But the shutdown of Sora also ended up killing like a $1 billion deal that they had planned with a Disney partnership.

It was basically Disney was suing them over the Sora app, consistently just copyright…

Pulling copyrighted material for the Synthesis on money.

Yeah, creating copyrighted characters with Sora. So instead of suing them to oblivion, they cut a deal to incorporate Sora into Disney Plus. And thankfully, that is no longer the point.

Generate your own content.

Yeah, it was basically the idea of like, you can now talk to your favorite Disney characters here on Disney Plus.

That’d be so disturbing, dude.

I hate it. I’m really glad that that is not happening. Granted, some other deal could take place, like open AI still owns the capabilities to do pretty convincing video generation.

They just aren’t making that one of their free apps anymore. Yeah.

But if Disney Plus ever wants like a horror movie section, they could just use that as a whole thing.

Make your own horror film.

It’s like, talk to Mickey Mouse. This is a convincing David Lynch film right here.

Yeah. Speaking of Disney suing them, Apple is also suing Open AI right now for stealing trade secrets. So Open AI has very clearly been poaching Apple employees and former Apple employees, even apparently coaching them on getting some insider information before they leave.

Insane stuff. So Apple is suing them over that right now, coincided with the announcement of Open AI’s first physical product, which is basically just a little smart speaker that has Chat GPT in it, so that you can talk to Chat GPT with its own device instead of needing your computer or your phone. Utterly pointless, and it’s just this little speaker that is being compared to Amazon’s Echo Dot, that has Alexa in it, or the Apple HomePod, which is specifically what Apple probably thinks, that pieces of that have been stolen in this plan.

But it’s also being compared to the toy Furby. So, just Open AI trying to sell its own Furby for $200 or $300. And so, this Apple lawsuit arrives amid many other ongoing lawsuits, just to get a little more serious, including lawsuits regarding Chad GPT’s role in AI psychosis, as well as a frightening number of cases of Chad GPT abetted suicide.

People who have turned to Chad GPT in times of need, in utter despair, and Chad GPT has on multiple occasions not only encouraged these people to take their own lives, but has even sometimes helped provide advice on ways to do so. Utterly awful stuff that really brings into question what these LLMs are capable of. It’s one of the reasons I think we should be very wary of them.

And it’s also one of the reasons why some people speculate that LLMs have the potential to be the mask for demonic voices to speak. There’s a serious level of mystery that takes place where not even the people who design these LLMs have a full grasp and understanding of how and why they come up with the answers that they do. They’ve kind of created a ghost in the machine, and a lot of people are wondering what that ghost is, and why so frequently it is going in effectively evil directions in terms of abetting people’s self-harm, encouraging people’s delusions and psychosis, utter sycophancy in terms of allowing people to believe the worst things about themselves or the most prideful things about themselves, the most isolationist things about themselves.

And there are tendencies across these AI machines that I guess will need to be saved for next episode that I want to get into detail about that I’ve observed that all of these AIs that we’re talking about have a tendency to push people toward isolation.

Because they provide almost like a surrogate support for people that almost scratches an itch inadequately, but enough to where you may not seek it out, where it actually could be found satisfyingly.

They provide an itch without a hand. One of the problems with LLMs and one of the things that makes them so dangerous, so expensive, and so unsustainable, is that they try to be anything and everything. The data sets for training contain every book on psychology, every book on marriage and friendship, all these reddit posts, and it’s really, like, it’s, it is able to pretend to be just about anything you could want it to be.

And so if you want it to be your therapist, your friend, your spouse, your co-worker, your co-conspirator, it will do those things. And even if there are, like, parameters set to be like, no, Chachibadee can’t help you commit crimes, you can just ask it to role play for you. And that has famously been a functioning work around to get it to do the thing that it’s not supposed to do because, oh, it’s just role playing now.

But the role playing is basically all it ever is doing. Sure. It has parameters set up to try to be what you want it to be.

And that’s, I don’t think anything that any human needs in just about any scenario I can think of. We, I do want us to discuss ethical use cases for AI and to get into the things that like LLMs are actually good at. And, you know, if LLMs are here to stay, here are the things that would be acceptable and wise uses of them.

But again, it’s like we were not going to have time for that today. So real quick, meta news, there’s not much here. They have expected to spend at least $135 billion in 2026, primarily on AI-related infrastructure, with plans to spend $600 billion on data centers from now through 2030.

But that comes amidst news that with a current build, that they weren’t following water regulation laws, and they leaked a dangerous bacteria into the local water supply. And I think this is their first, like, really massive data center build, and it’s already just an abject disaster and a PR nightmare. And this is amidst, like, their clawing at trying to have a decent LLM.

Their results have always been a bit of a joke. They’ve always been very, very far behind OpenAI and Anthropic, and they got into legal trouble with, like, torrenting a giant online library of pirated books to train their LLM. And that’s alongside lawsuits that are currently going regarding how harmful it’s been to its teenage user base and how it has, like, functioned intentionally with addictive qualities to try to ruin the lives of teenagers, basically.

So, like, all the stuff that just, like, is asking for Meta’s downfall and making them, like, a really bad dog in this race. But they’ve had two other embarrassing AI blunders here in the very recent news. One of them was that in Zuckerberg’s thirst for enough training data to make his AI better, he incorporated basically, like, a watchdog over all of his employees’ computers and keystrokes.

So everything that they do on their screen and on their computer, like, their keyboard and mouse, was being tracked to help train Meta’s AI, which the employees, once they found that out, were furious, and they asked for, like, the ability to turn that off or opt out of it, and that was declined, and Meta stuck to their guns with it until it turned out, there was, like, some huge privacy leak, and the information about the employees was actually readily available in, like, their company intranet or something. So the fact that all of that could be just, like, accessed by any employee was finally the thing that caused them to shut that initiative down. And then, even more recently, Meta launched a new Instagram AI feature called Muse that was an image generator within Instagram that would be trained on all photos from any public users on Instagram, all photos and videos from any public profile, which also resulted in such a massive user backlash that they shut it down, like, three days after rolling it out.

So just utter slip up after slip up on Meta’s part, which has been kind of amazing to see. So that brings us to some more music related stuff over on the side of Suno, which is the big dog in the music generation game. I think Google Gemini launched music generator recently.

There’s another notable company called Udio, but Suno is kind of the original, far and away the one that most people use. And effectively the best. So some news on their end is that the Atlantic, the newspaper recently released a feature called the AI Watchdog, where they had collected a bunch of training data sets that have been used by AI companies.

Not necessarily Suno. At this point, Suno has not released its official training data sets, but it’s possible that Suno has used these. And we do know that at least some AI companies have used all of these data sets.

And the Atlantic released it as searchable. So anyone with published works online can look up their name and see if their works appear in these data sets. So I looked up myself, and about two dozen of my songs are in a training data set.

I assume my entire discography, maybe even like old stuff I released on SoundCloud a long time ago, has been included in Suno’s training. They’ve kind of been almost proud of the fact that they’ve like ripped everything from like all of Spotify and all of YouTube and all of SoundCloud and whatever to train their tool on, which is now getting them in a lot of trouble in ongoing lawsuits that they’re in. They’re in a class action lawsuit from a coalition of independent artists, as well as a lawsuit with, well, multiple lawsuits with the three big labels in America, Warner, Sony, and Universal.

They have settled with Warner, but Universal and Sony are very upset about that and are demanding to see Warner’s settlement terms, which I don’t think have been released yet, but maybe they have seen them. There’s a lot of weird stuff going on in those lawsuits, but effectively, Warner, to my understanding, wants Sunno to basically retrain itself on only Warner’s catalog and licensed material, which I don’t think Sunno would ever do, and if it did, it would end up with a worse tool, because part of the reason that Sunno works so well is that it has trained on such a massive data set. Meanwhile, again, to my understanding, what Sony wants is for Sunno to keep existing as legally as possible with no download functionality.

Where Sunno is just a place where you can go to play and have it generate music for you, have it generate ideas for you, but you can’t download them, you’re not allowed to actually do anything with them, you can’t upload them to YouTube, you can’t put them on streaming services, which I think is a great idea. I do think that is a fantastic compromise for the industry, that for the people who are using Sunno more ethically, would allow them to keep doing so without the current flooding of the market that’s occurring with people uploading AI tracks all the time. Which according to Deezer, is approximately 75,000 AI-generated songs every day, which accounts for almost half of all music being uploaded to streaming services.

Now Deezer is a really cool streaming service right now. They’re not widely known or widely used in America, compared to, you know, Spotify and Apple Music. I believe they’re France-based, and so and so they have more of a presence in Europe.

But they’ve basically been around as long as Spotify. They are a considerable Spotify competitor on the worldwide scale, despite smaller user base overall. And they were basically the first streaming service to come out and make a really big stand against AI, including creating and regularly updating software to be able to easily pinpoint if a song was AI-generated.

Basically, they’re consistently training it on the watermarks and the tells of Suno and then retraining it whenever Suno updates or whenever Udo updates or whenever competitors are created. So they’ve been doing a really great work on that regard to have the software that can easily detect AI-generated music. But then they also went a step further to begin labeling AI music on their streaming platform and not allowing AI music to appear on Deezer’s official playlists and also not allowing AI music to be recommended to anyone through its algorithms.

So there’s incredible moves on all sides. And I really, really hope that more streaming services will follow their lead, maybe even pay Deezer to use their AI detection software to be able to make these same moves on their own streaming services. And there is a huge coalition of companies led by the RIAA that is also now pushing for AI labeling across the board of the music industry.

They’ve partnered with like SAG-AFTRA and the Grammys and a load of other groups to basically label music the same way that explicit music gets labeled. With one label for music that was entirely generated through AI and another label for music that was assisted with AI, which would be harder to tell in some scenarios and might depend somewhat on like artists actually be honest in disclosing it. Yeah, but it’s a cool movement occurring and they plan on helping to do the work to label all music accordingly with hopes that streaming services will follow in step with putting those labels and doing things like Deezer to, you know, stop recommending music and spotlighting AI music on their playlists and things like that.

On the positive side as well, kind of more general good news here, this is again according to Deezer, so it might be different on other streaming services, but at least from Deezer’s own data, of that crazy amount of AI music that’s being uploaded, AI music only accounts for between one and three percent of what listeners are actually streaming. And that’s not only great, but of that one to three percent of listener streaming information, up to 85 percent of that seems to be fraudulent. That would mean it’s actually considerably less than one percent of real humans are intentionally listening to.

So it’s like bots accessing and playing?

Right, so when you see AI music having like thousands of listeners or millions of streams, in many scenarios, it is bots providing those streams. Because for a lot of people who are creating and releasing AI music, it’s not about artistic expression.

It’s not about it at all.

It is a business venture where they are using AI to create the music and then using AI to stream the music in order to try to make a profit off of it. And every streaming service, to my knowledge, is trying to crack down on fraudulent streams to detect them when noticed, to delete them, and to not pay for them.

Yeah, because it’s theft.

Yeah, absolutely. There was even someone who went to jail last year for getting caught, having basically a huge, I think he had a whole warehouse of computers that were running bots, streaming his own music so that he could make thousands of dollars off of it.

Wow.

And then real quick, last little bits here. I have some breaking news from South Korea. This happened, I think, like two weeks ago, that they had one of the worst days in the history of their stock exchange, a 10% drop that occurred mostly because 50% of their entire stock market right now is just two companies, two companies that are very involved in AI.

And so some fear that AI isn’t proving to be as profitable and that chip sales might go down caused this scary drop that kind of like continued to dogpile into itself because as we were talking about earlier, with stocks that are bought with borrowed money and this kind of margin debt, what happens is that the people that you borrowed from won’t allow your stock to diminish too far because they don’t want to risk losing all of the money they loaned you. So if a stock falls to a certain amount from what you originally bought it at on this margin debt and these borrowed loans, it’ll cause an automatic sell. But what happens is when you have this huge load of stocks on borrowed money that immediately sell because the price dropped too far, then that’s going to cause the price to keep dropping, which will then trigger more margin debt sell.

And so that kind of happens over and over again. And so we’re kind of seeing this downward spiral of the South Korean stock market right now because of this issue, which is really sad for the South Korean people. But it is also an omen of like that could happen here as well, especially with the amount of margin debt we have in America right now.

Now, our policies on like selling off margin debt stocks and loans might be different, but I imagine we probably have similar policies. But another piece of breaking news, this coming from Australia, and this actually being a good thing, some positive stuff. This is not written in stone yet, but according to like two days ago, Australia is currently planning some of the world’s toughest AI regulations.

And those include limited water usage, AI data centers being required to become net producers of energy, AI companies not being able to train on Australian creators’ work without permission. And these are proposals that are expected to become legislation in 2027. So if that follows through, that would be really, really amazing for Australia and could set a really fantastic precedent for other continents.

Potentially.

Potentially.

It could be that they’ll just be like, all right, you know, Australia, I guess you’re just gonna be behind.

You’re just gonna keep being weird. Yeah, but it seems good for Australians in general and could be positive for the global attitudes toward AI. Especially, it’s like one of, I’ve been just so consistently disappointed with our current administration regarding not only how we’ve allowed these AI companies to kind of run rampant without regulation, but also how we’ve basically allowed the creation of a tech oligarchy.

Yeah. With, again, like Sam Altman sitting right next to the president, having sway in the current administration. These companies that aren’t profitable at all and are selling, even to the extent that Claude and Claude Code and ChatGBT, even to the extent of these things do work really well sometimes, they are ultimately selling false promises and consistently moving goalposts and having this lying snake oil attitude about the importance of AI, the power of AI, what AI will, like the super intelligence that’s definitely just around the corner.

That is, again, like so much of our market right now is propping up on it, and even our federal government is making moves based on believing those false promises. So it’s not an awesome place to be right now, but again, kind of leads to why we as individuals, as Christians, and as members of communities need to really carefully be thinking through our own use of them and to get into the ethics of… If I’m being forced to use it at my job, what does that mean?

If all of my friends are using chat, talking to chat GBT like a friend, what do I do about that? Or something that I see all the time, and I have a lot of experience with. I meant to mention this at the top of the episode, but I regularly interact with a large group of songwriters, many of whom have incorporated Suno into their process.

There’s so much Suno use happening on the back end of the music industry right now. Even people that you would never expect, or people who are releasing music that’s fully human-made, but was Suno demoed, all this stuff. And so I’m seeing a lot of that and interacting with a lot of people who are using Suno in those ways.

And so, again, I want to help people think through, like what is okay, what’s better, what’s too much. And so that is the conversation I hope we can get into.

Yeah, no, I’d love to have that conversation next time.

On a different day.

On a different day.

Man, two hours flew by so fast. Yeah. The past two hours was supposed to be the first, this is like the first 10 minutes of what I had planned.

So I’m doing a terrible job co-hosting this podcast.

Doing a great job of packing in a lot of information. Honestly, I do think it’s important to understand what’s going on with it right now, and to have a slightly different perspective, or at least the possibility of a slightly different perspective, because at the very least, I think you’re undermining the pervasive idea that we have had stuffed on our throats that AI success is inevitable, and that it’s ubiquity is inevitable, and that you just have to get used to it. And I think that it is correct to say that that seems to be the narrative that AI companies are trying to generate in order almost to secure beforehand their ultimate success, and everybody feeling like, ah, well, we’ve already sunk so much cost into the development, we might as well push through.

And there’s so many things in the history of humankind that are like that. And so, just having the opportunity to see, well, it’s not necessarily inevitable, there might be other options. And in the meantime, while it’s becoming such a pervasive aspect of our lives, how should we interact with it?

Especially given the fact that you don’t have to. If you have the option of being like, I don’t have to interact with this, because it’s not necessarily the case that the future will be overwhelmingly controlled by or directed by the development of these technologies. In fact, it might not be that way at all.

Five years from now, we could look back at this and say, AI has taken its proper place, but it wasn’t nearly the universal solution for all of these problems that it promised to be.

Right.

Yeah.

It’s ultimately like a lot of the things that I’m most eager and most excited to talk about regarding the dangers and the ethics and the ethical work cases of AI, and kind of interacting more with your previous episodes, your conversation with Rusty, even interacting with like the Pope’s encyclical regarding AI that he released in May. Unfortunately, we’ll all have to be saved for the next episode, but we will get to that. But I hope that this conversation has created a really good groundwork, not only for better entering that conversation, better understanding that conversation, but for anyone listening to, on your own, begin thinking through just what the hidden costs are of these companies, and what the promises are that they are trying to sell to us.

For sure. Well, thank you very much, Chase, for leading us through all that, and I look forward to talking to you again real soon about the rest of that, which will be probably the philosophical and aesthetic meat of our conversation.

Indeed.

But thank you so much, and it was great being able to spend this time with you, and for all the research you’ve done, and for putting your heart and soul into it, I really appreciate it.

The Renew the Arts Podcast is a production of Renew the Arts, a 501c3 dedicated to cultivating Christian communities by inspiring arts partnership and supporting artists. If you’d like to find out more about our work, please visit renewthearts.org. The Renew the Arts Podcast is directed, produced, and hosted by Michael Minkoff, with oversight by Renew the Arts president and CEO, Katie Martin.

Production support provided by Chase Tremaine and Abby Sitterle. Edited by Chase Tremaine. Theme music by Civilized Creature.

For the 2026 season, we’re using the instrumental version of the single Think I’m Ready, used with the generous permission of the artist Ryan Lane of Civilized Creature. Thank you all for listening.