There may be errors in spelling, grammar, and accuracy in this machine-generated transcript

Roger Harris: Hello everyone. It's another federal tax update podcast. This is Roger Harris and Annie is my co-host as always. Annie, how are you today?

Annie Schwab: Doing pretty good. Anxiously awaiting the 15th of October, but I think I'm probably not the only one.

Roger Harris: Yes. As we sit here and record this, we've got a couple more weeks, and we're also potentially [00:00:30] 48 hours away from a government shutdown. And seeing what impact, if any, that will make on us. And but like always, probably a last minute solution seems to always come up. But maybe this one's different. We'll see. But by the time you're listening to this, you'll know the results.

Annie Schwab: Yep.

Roger Harris: We'll see how it goes. Um, we have a very special guest. We're really honored to have Blake Oliver with us today. And most of you probably know Blake from the fact that he has [00:01:00] his own podcast. Um, the accountancy podcast, which I think is one of the most popular in our industry. Uh, he formed earmark, and he's also a CPA. And we're going to talk something that we all know about and need to know more about and wonder about. And that's AI and Blake. Welcome.

Blake Oliver: Well, Roger, thank you. Thank you so much for having me. Uh, I have to do a small, quick correction. Uh, my my CPA status is currently suspended.

Annie Schwab: Oh, no. Oh.

Blake Oliver: I forgot to get all [00:01:30] of my, uh, CPE. And you might wonder, how is that possible? Because I'm the founder of earmark.

Roger Harris: Which.

Blake Oliver: Federal tax updates is on. And everybody who knows earmark knows you can get CPE for listening to podcasts. And I host a podcast every week for CPE. So that's like 150 hours of instructor credit. So how did I not get my credits? Well, in Arizona you have to get 16 hours of your 80 hours as in person.

Annie Schwab: Cpe oh.

Roger Harris: Oh. [00:02:00]

Blake Oliver: I forgot about that. So I need to go figure out how to get some in person. Cpe to to get my.

Roger Harris: They really mean you've got to go sit in a classroom and you can't even do it, you know, live, virtual or something. You've got to physically go to some building and sit there.

Blake Oliver: I can do webinars.

Roger Harris: So okay.

Annie Schwab: Gotcha.

Blake Oliver: So I have to go sit on some webinars. But like, my schedule's been crazy and I lost track of it just like everybody does. Right. So this is why this is why we exist. So this is why that's.

Roger Harris: How you are so relatable [00:02:30] to everybody in the audience. We've all been down that road. I mean, we've all done that and they changed rules and somehow they don't tell us about them.

Blake Oliver: Oh, and thank you for reminding me about the October 15th deadline. I was going to miss that too. Oh, no, said something. So thank you. I'm. I guess there's a reason that I'm not a practicing CPA anymore, because I'm not very good with deadlines. Well, you know, that's tough. So yeah.

Roger Harris: We got plenty of those.

Annie Schwab: Yeah. But I'm so glad that you're here with us today. Ai is something that we. Roger [00:03:00] and I attended several of the forums this year, the IRS forums, and it was a hot topic there, but it seems like everywhere you go somebody is talking about the newest version or this new tool or app or, you know, something that can write a letter or answer emails. You know all the things that it can do. I mean, I suppose one day our cars will be driving themselves and, you know, the robots will be serving us food at the diner. But here we are today, just kind of, in my opinion, maybe not [00:03:30] in yours. I know you've had more experience in the AI realm, but it's still kind of new to me and a little can be overwhelming and slightly intimidating. And, um, but but also, I'm so curious about it, and I think, I think many people are curious to see, well, what can it do for me and how can it help me? Um, so I'm going to turn it over to you to get us started on our topics for today.

Blake Oliver: Yeah. Happy to. And and, Annie, I would say that you are not alone. And I don't I [00:04:00] don't blame you for feeling the way you do, because it's kind of janky right now, to be honest. Like, it works sometimes, but it doesn't work a lot of the time. That's what I mean by that. And I sort of compare it to what we see in our lives. So I live in Scottsdale, Arizona, for instance, and Phoenix and Scottsdale are a big testing ground for autonomous driving vehicles. So Google has the Waymo unit here.

Annie Schwab: In.

Blake Oliver: Scottsdale, and [00:04:30] it runs all up and down the city. There's still parts that you can't go to. But I could from my apartment. I could hail a a Waymo and take a driverless taxi all the way to the airport. And that is something that wasn't around a few years ago. And it just. Have you ever taken one of these?

Annie Schwab: No I haven't.

Roger Harris: I'm still a little freaked out about it.

Annie Schwab: Yeah.

Blake Oliver: Yeah. It's one of those things where you get in and there's no driver, and at first you are like, what am I doing? And then it drives for a while and you're like, you realize, [00:05:00] wow, this this is this is a really good driver. This car is very safe. Uh, it follows all the rules. And then you look at the statistics and you realize these are safer than human drivers on average. Like, significantly like the number of accidents and whatnot. So, like, there's this psychological barrier. I think, um, there's also limitations to what they can do. They're just different than human drivers. And you just have to get used to that. You're going to take a longer route, right? It's going to it's not going to sneak [00:05:30] into traffic for you. It's going to it's going to go around the block and stuff like that. But that's.

Roger Harris: I don't want it to.

Blake Oliver: You don't want it to. Um, but it's also been like extremely, extremely slow. We've been talking about self-driving cars for decades now, and they're still not quite there yet. But every month, every year they get a little bit better and the map is going to expand. And eventually they'll go everywhere, all over Phoenix and on the freeways, too. How soon will that happen? Hard to say. It's always slower [00:06:00] than anyone in tech will want to estimate, but it will get. We will get there. And I think of this moment in that context with AI like we are, we are so close to getting it to really be able to do stuff for us in our firms, but it's not quite there yet in all cases, like in most cases. So that's frustrating.

Roger Harris: Yeah. Well, and this journey we've been through this before where we're hesitant about something, we wonder. And then [00:06:30] you look back five years and you go, well yeah this is all working. I mean, this is this is just the start, you know, like I can't see me getting in a car without a driver. But probably five years from now, every car I get in will probably be that way.

Blake Oliver: You might not own a car. Um, yeah. So I think of it also in terms of the web, when the you know, which we can remember very well, I you know, I remember the first time I ever ordered online, it was I think it was Amazon, it was books, [00:07:00] and that was all you could order. And that was a big deal for us. It was a huge leap of faith. And you'd wait a week and then stuff would finally show up, maybe more, and think about where we are now, where we can just literally order anything, anything you could buy in a store, you can order to your door. And that was this. It happened fast, but also so slowly that you didn't even notice it happening.

Roger Harris: And you never saw it happening. You never saw it coming.

Blake Oliver: Exactly. So I feel like we're [00:07:30] at that moment right now. Ai is sort of like the web when the web was brand new and we're at the point where, yeah, you can now order something online, but is it easier than going to the store? No, it's not right. Not all the time. Right. It's going to get there. So, um, it's great because it's a big opportunity, just like with the web. The accounting firms that got on there earlier and started using it with their clients and started hosting their QuickBooks files [00:08:00] or hosting their tax software, had an advantage and could establish multiple offices and, you know, be more mobile. And then we went to cloud. It's the same sort of phase in kind of approach. It's not going to be magically all at once were there. And so that's why it's important to be experimenting always and to to stay in tune with the new developments, even if you're not yet using it day to day in your practice. Although, as we'll talk about [00:08:30] today, I do think there are ways to start using it.

Roger Harris: Yeah, and we might be using it and just not even knowing it in some instances that the results we're getting are AI driven. And it's like anything that we don't understand, we're afraid of it because we don't understand it. And as we get comfortable and understand it more, we'll will be more willing to venture into it. But again, we're probably all there and don't even know it. And we're definitely on the journey whether we know it or not.

Annie Schwab: Sure, sure.

Blake Oliver: And I'm glad you brought up that point, Roger, about understanding [00:09:00] it, because I believe that this technology is in many ways there's there's similarities to what has happened before, but it's also fundamentally different. And there's a psychological like switch that we need to have to effectively use AI. And that is related to understanding how it works and how it mirrors human intelligence, but also how it's different. And, you know, I grew [00:09:30] up reading sci fi. So like, you know, here I am reading reading iRobot by Asimov. Um, checking that out again, all the stuff, like when I was a kid that I fantasized about is now coming true when it comes to artificial intelligence. And so I'm really trying to understand it. I'm not a technical guy, I don't know. Ai from a technical perspective, I couldn't tell you how a how an LLM is built in detail. [00:10:00] The same way I couldn't tell you how a car engine works, although in theory I have it right. But I think it's important to understand a little bit of how it works and the way I describe it, the way I've come to understand it as a layperson is that these these models, these large language models are basically giant. You can think of them as giant, multi-dimensional spreadsheets. That's how I describe them to accountants. So your typical spreadsheet has an X and a y axis.

Roger Harris: Right. [00:10:30]

Blake Oliver: Um, and then a bunch of values in the cells. And that's how they are related to each other. Now imagine that the spreadsheet has a z axis as well. And that that z axis that goes through the screen is is potentially infinite. So what you have is this giant, um, it can be represented in, in multiple dimensions, in three dimensions. But it actually like there's not only a z axis, there's like an a axis and a [00:11:00] b axis and a z axis. There's more dimensions than we can actually visualize in three dimensional space. And it's n dimensional meaning there can be like infinite dimensions. And so so it's hard to you can't actually visualize it. Um, but that is what the that is what is created when a model is trained and it, it looks a lot when you visualize it, it looks like a brain, like a neural network. Right. All these connections and, and and when um, and basically that's [00:11:30] how that, that model is how we store ideas, like that's how the human brain works, is like we take in sensory input and we convert that sensory input somehow. We don't know exactly how into the concept of a podcast like that. We are talking to each other right now is something that our brains comprehend, and these models actually comprehend it too. What's different about them versus the human brain [00:12:00] is that our models have the ability to update in real time. So as I talk to you and as you talk to me, my model of the universe, of the world, the way I understand it works is changing. These computer models, these artificial intelligence models, they're fixed once they're trained. So that that is the fundamental difference. They do [00:12:30] not move through time the way they learn.

Annie Schwab: They learn new things. I feel like.

Blake Oliver: Well, so they can when they're in their training phase. Face.

Annie Schwab: Yeah.

Blake Oliver: That's how they learn. But then once, once they are trained, the only thing they can do is they can be they can be fine tuned a little bit and they can, uh, acquire knowledge, but they don't actually the model doesn't update. And that's why if you start a new conversation, the [00:13:00] model doesn't remember the previous conversation unless that's been added to its knowledge. But it's knowledge that would be just like, let's say you and I had a conversation, Annie. And you, you took you wrote down that conversation in in a notebook, but you would have no memory of it unless you went and read it in the notebook. It's like they have amnesia.

Roger Harris: Yeah.

Blake Oliver: So um, it's like that. I don't know if you ever saw that movie Memento by Christopher Nolan.

Annie Schwab: Oh, I think I have, actually, [00:13:30] yeah.

Blake Oliver: Classic film. You know, a man who has amnesia. And so he writes notes to himself to try to solve a mystery. And I sort of think like model. The large language models are kind of like that. Now you might wonder, like, what does any of this have to do with?

Annie Schwab: You should see the visual that I have in my head right now. This this thing.

Blake Oliver: So what it means, what it means is that, um, it's like imagine, imagine that a large language model in your firm is sort of like an intern, a smart, very, very smart intern, like [00:14:00] Harvard trained, has all the knowledge, uh, read all the books, but forgets everything that you say, right? Forgets what you taught them and has to be given extremely, extremely detailed instructions on how to do everything because they have no experience in the real world. They've just lived in a library their whole life. So that's why it's really frustrating, actually, to work with AI, because it lacks the context that humans have, and we just assume [00:14:30] so much when we talk to each other because we have a shared experience of the world that these models don't because they were just trained on text from books. And. On the internet, which is a representation of the world. But it's filtered and it's missing a lot of like the stuff people don't write about.

Roger Harris: So there's a lot of information and a lot of knowledge there. And we've I guess one of the things we have to learn is how to properly tap into that knowledge to [00:15:00] get what we want back. Yes, because it can't it can't interpret what we're saying and understand what we're saying or have, like you said, the benefits of an earlier discussion. So we've got to kind of bring it up to speed and focus it. So it really zeroes in on the knowledge that it has.

Blake Oliver: And that's what effective prompt writing is. So it's it's giving the model when you ask a question, giving it enough context so that it can give you the right answer. And [00:15:30] that's really challenging because in order to do that, you have to kind of work through it yourself. You have to think, if someone was asking me this question, what context would I need to be able to answer it? And that's why that's why the most experienced people in your firm executives, partners, the most experienced people actually have the easiest time using AI. And that's because they are familiar with [00:16:00] this concept of delegating. And when you delegate to people, you have to give them context. You know, you have to give them, um, you have to tell them all the background they need to know. You have to make sure they understand the background of the project you're asking them, or the task you're asking them to do. Um, in the case of a client, right, you need to give them a lot of information about the client. Um, you need to tell them what is the objective. Like, what do I want [00:16:30] you to do? And how do we measure success? And you need to tell them what? What are your what are you allowed to do? What decisions can you make on your own? And what do you need to come back to me to ask me about? You need to you need to be able to say, um, here's a process that we typically follow.

Blake Oliver: Define that so that this very smart intern can follow that process. And you need to give them examples of work that's been done before so they can follow it. So that's a lot. You [00:17:00] can't just write a question. You have to write the question and give all of that. Um so learning how to do that is is kind of a challenge. And actually I feel like it's very much related to how, uh, many experts, many domain experts in tax struggle to grow firms because nobody ever taught them how to manage people. Like, we didn't learn that in school. And it wasn't something that our accounting firm that we worked at ever developed in us. And it's a whole [00:17:30] skill.

Roger Harris: Yeah. No, I think what we normally do in this profession is instead of training people or knowing how to work with people, we just do it ourselves, right? And that's why we're overworked.

Blake Oliver: Yes.

Roger Harris: Because it's easy for me to do it than to know how to hire people and teach them how to do it and train them and supervise. It's just I'll do it.

Blake Oliver: And AI kind of is the same way in the sense that you have to teach it what to do. Um, and it takes a lot of time. So if you have a task you want to automate in your firm, [00:18:00] it's going to take a long time compared to doing the task. It's going to take many, many times longer. In order to set up what I call a you'd usually use like a custom GPT, right? A set of standard instructions to accomplish a task, to document all that. Right. That takes a long time, the same way as documenting it for your team does. So AI is not a magic bullet that's going to solve your capacity problems or get you out of the weeds because it's you still have to do the same work that you would be doing [00:18:30] if you were having a person do it. In many ways, the benefit is that once you get a really good set of instructions and you build that custom GPT, which we'll talk about, then it's very reliable. If you're using it for the right tasks, it's more reliable than a human. So that's where the benefit can be for certain things, not everything.

Roger Harris: And we need to get into those use tasks that we can do it. But you hear so many people. So what I'm hearing you saying is a lot of people think that AI is going to be what [00:19:00] puts them out of business, that somehow it's going to to replace what we do. What I'm hearing you saying, and hopefully we're all correct in this, is that this is a tool to actually make us more efficient and better. It's not going to replace us. So for people who are hesitant to take it on because they say, why should I learn it, it's going to put me out of business. I don't even know it because I'm going to have to go find something else to do. That's not how we have to think about AI. It is a tool that we need to learn how to use, and we'll talk about some of those use cases.

Blake Oliver: It will [00:19:30] put some people out of business, but I'll qualify that right. It's AI is really good at doing tasks, specific defined tasks, and we can talk about some of them. Examples of some of them might help to illustrate this. It is not good at managing a process right. That's what people do. And so if your job is mainly task based, where I don't know, you're you're taking documents from clients and you're putting the numbers from the documents into tax forms. That's a task. [00:20:00]

Roger Harris: That's a task.

Blake Oliver: That's getting automated right now with AI. Sure.

Roger Harris: And that's a good thing.

Blake Oliver: That's and that's great. As long as you're creating value by doing something other than that.

Roger Harris: Right?

Annie Schwab: Sure.

Blake Oliver: Um, so.

Annie Schwab: So there's, but there's so many of them, like, I feel like I hear someone's like, well, have you tried this one? Have you tried that one? Which one do you like better? Which one are you using? Which one's worth the price? Which one? You know is. So how do you know which ones? Well, I guess which [00:20:30] ones you can actually rely on, because I'm sure they're all based where they're pulling their answers from the sources that they're getting their answers from may not be the same across the board.

Blake Oliver: It's interesting, actually, so that that the major models seem to have all somehow coalesced around like the same body of knowledge.

Blake Oliver: So.

Blake Oliver: So like and part of it is because they were cheating and they were like training their models on each other's models. So like [00:21:00] like. It's.

Annie Schwab: Sounds kind of smart to me, but okay.

Blake Oliver: Yeah, yeah. Um, I mean, they're all stealing everybody's intellectual property anyway, so who's going to complain, right? Right. So, um, but like OpenAI with ChatGPT, I've been using that for years now. Um, the latest model, number five, GPT five, is just spectacular. And that's the starting point I would recommend for anyone is just start using that because it works really well out of the box. You don't need to know a lot and [00:21:30] you'll get a lot of value. I used to use clod and that's by anthropic. That was my favorite until the latest GPT five model. Clod is better still at creative writing and business writing. It just sounds more natural for some reason, whatever text it was trained on. Um, you know, maybe, maybe ChatGPT was trained more on just whatever's on the internet, and maybe clod was curated better to be trained on, like, business writing and creative writing, like good writing. [00:22:00] So the writing is better. But, um, it's just not as like, functional. They haven't been focusing as much on the, uh, like user experience side of things, like the way you use it. And OpenAI has really leaned into the consumer type features because that's their most of their user base is just individuals and small teams, whereas clod is sort of pivoted towards software developers and focusing on helping them code. So these are like different use cases. And [00:22:30] yeah. And so like if you go with Claude, I mean most of us are not developers. We're not going to benefit from that type of functionality. So I'm leaning towards ChatGPT these days. Um, there's also specialized models or tools. And um, so specifically for research, I'm a big fan of perplexity. And what's neat about perplexity is that it's like a, like a search engine layered on top of [00:23:00] AI models, and you can use different ones. So you can try your research with GPT five or with Claude, or with their own proprietary model, and see what the difference is, which is kind of fun to experiment like, ask the same question and try different models.

Annie Schwab: Yeah. So I've done that. Some tried like literally type in the exact same question and see, you know, which one comes back. Even if I knew the answer, just kind of see what what comes [00:23:30] back. And I guess, are they all safe? Like, I know there's some you can, like upload tax documents with people's names and, you know, like I'm assuming all of the main ones have gone through some really crazy security background type thing.

Blake Oliver: Yeah. So, so like ChatGPT OpenAI. Right. Like they've got that deep partnership with Microsoft. They've gone through all the security audits. Now the key is the key is you got to be using a team plan or an enterprise plan like a business [00:24:00] account.

Annie Schwab: The free plan doesn't have all the security.

Blake Oliver: If you're using the free version, then they can take that data and use it to train the model, which you definitely don't want client data in there.

Roger Harris: Yeah.

Blake Oliver: No, no. So it's use the team plan. Enterprise plan. You know, review the terms of service. Um, uh, to make sure that your, your data is protected. But if you do that, it's really the same to me as hosting something in the cloud.

Annie Schwab: Okay.

Blake Oliver: Like that's how I feel about it. Um, and we're all going to get there eventually. And yeah, I always [00:24:30] say without without risk, there is no reward. So you have to be willing to take a little bit of risk in order to benefit from this new tech. Just the same way, when we move to cloud, we took some risks. And you just have to mitigate that with proper insurance. Um.

Roger Harris: You gotta take the same precautions you do on all kinds of other ways, data that we have and handle and use. It's I mean, it's this is just another place where we've got to take it seriously and not just know it'll never happen to me or what do I. I mean, [00:25:00] that's sadly too many people's attitude. Yeah.

Blake Oliver: Yeah. And there's specialized tools too that are being built specifically for accounting that address those concerns even like more strictly. And, um, you just you're just going to pay more for it, right? So you've got tools like Bluejay, for instance. Right. Which are orders of magnitude more expensive than using ChatGPT, but promise better security and promise, you know, more specialized.

Annie Schwab: So what is the range of these things cost?

Blake Oliver: Well, I [00:25:30] you know, I don't know the pricing. I don't know all of Blue Jays pricing plans.

Annie Schwab: But yeah, but just in general thousands.

Blake Oliver: It can be. Yeah. So you can go anywhere from like like with ChatGPT team. Like let's say you have a small team like five people, right? You're going to be paying 150 bucks a month, I think. So it could be anywhere from 150 bucks to like 1500 a month, depending on, you know, what, what you go.

Annie Schwab: With. And I'm sure specialized ones that are, you know, specific industry based or specific.

Blake Oliver: Yeah. [00:26:00]

Annie Schwab: Yeah. Interesting.

Blake Oliver: But but you know, it's like, um, you have to weigh that against how much time are we going to save? Or better yet, how much better, like tax planning can we do? Can we uplevel our staff so that they can actually do some of this tax planning that only the partners can do. Or maybe the managers. Right. Like if I'd had this stuff when I was a manager in public accounting, I could have been doing partner level work. So you got to think about like that [00:26:30] for your firm. Like, is that worth $1,000 a month to like, level up my partners so they can act like directors?

Annie Schwab: Um, that doesn't make sense.

Roger Harris: Yeah, it's just a cost of doing business. If you want to really be professional in what you do. And I guess one of the challenges, I mean, this stuff is changing so fast. I mean, and people are I know we get bombarded. Bluejay is one of the people that come to us. They say, do this. How in the world does the average practitioner out there, is there any like these [00:27:00] five things you should look for? These questions you should ask. I mean, how do they. Because they're going to get something probably every day from somebody telling them that this is a better version of what you didn't understand before?

Blake Oliver: Um, well, as with all tech, I say ask around in your professional communities to see who's using it and if they're getting value out of it. Never trust the marketing message or the sales people all on their own. Trust but verify.

Roger Harris: Yeah, right.

Blake Oliver: Um, and also, like, just pay [00:27:30] for the pay for tattoo or Claude or perplexity and test it against what you're thinking about using and just see how much better it is. It's it's not that hard to, like, take exactly the same prompt and copy it into the other one and just see what you get and just do that on a like run. Run them in parallel. It's sort of the same way, like when you're testing a new general ledger application, you're going to run it in parallel and see how it does. Like, you just have to we're at the stage where we're just going to be trying a lot of stuff.

Roger Harris: And [00:28:00] so start with the more affordable things. Learn, get, get your knowledge base up. Understand the basics before you jump into the.

Blake Oliver: Exactly.

Roger Harris: Thousands of dollars, so you have a little better way of judging why this is supposedly better. Exactly. You don't want to start there.

Blake Oliver: You're buying a car, right? Like you got money to buy. Like. And you want to buy like, a Porsche, right? Well, you don't go, like, if you don't know anything about cars, you don't just go buy, like, a 911, right? You you like try.

Roger Harris: Well, some people do.

Blake Oliver: You shouldn't.

Roger Harris: You know, they got plenty of money. They don't care.

Blake Oliver: Yeah. [00:28:30] You know, like, work your way up, right? Like, if you're buying, if you're if you're tasting wine and you're not, like, very sophisticated, don't go buy $100, $200 bottles like it's a waste of money, right? Yeah. Start with a $30 bottle. And that's that's what ChatGPT is, 30 bucks a month, so.

Roger Harris: Yeah. There you go. Go get some basic knowledge. Yeah. You buy that $200 bottle of wine. You don't know if that's better than the $30 bottle of wine or not.

Blake Oliver: Exactly.

Roger Harris: You don't know what you're comparing it to. We've gotta all have comparisons and baselines and things like that. Um.

Blake Oliver: The [00:29:00] key is to do a little bit every day. That's that's it's like everything in life is like, make sure you've got, like, even just like ten, 20 minutes of using this stuff. Like have it open on a monitor and just be thinking, what could I what could I try?

Annie Schwab: Yeah. Interesting. So are there some like definitely don't do's like or things to not even like let yourself go down the rabbit hole of or.

Blake Oliver: Well [00:29:30] so the don't do my number one don't do is don't take the human out of the loop.

Annie Schwab: Don't take the human out of the loop. Okay.

Blake Oliver: So that means.

Roger Harris: Like.

Blake Oliver: Don't let AI write your LinkedIn posts without editing them.

Annie Schwab: Or verifying the or verifying or proofing and all the things.

Blake Oliver: There's all these tools now that will do. They claim to do like automated email prospecting and reply to your emails for you and like automate all this stuff. And what will happen is it will look good at first and then [00:30:00] it'll just it won't sound like you. It won't be authentic. People can tell when you're using AI, and so be really careful about damaging your own brand by letting AI speak for you.

Annie Schwab: Seems pretty logical. I can see how that is an important. Yeah.

Roger Harris: It goes back to you've got to hone in on that skill of how to give it what it needs to do its job. And if you take yourself out of that. [00:30:30]

Annie Schwab: Yeah. I mean, I mean, garbage in, garbage out, right. If you if you're not asking it in the right format or the right way with the right terminology and asking it in a way that it understands, you're not going to get what you want out of it.

Blake Oliver: Mhm.

Roger Harris: And something you said earlier trust but verify. I know we've been testing out, you know in the research and the tax research. And you give it what you think is the right amount of information. But you know the answer is wrong that you got, it's not 100% [00:31:00] wrong. But you know, it's not complete or not there. So we've got to be better at giving it the information, but it's going to be a while before we can just say it, take whatever it spits out and go. That's it.

Blake Oliver: Yeah, and I don't I don't know if that's even like the best use for it because. These, these like a lot of these complex tax situations. Um, the issue is that there's a lot of conflicting information. And I don't just mean misinformation on [00:31:30] social media. I mean that there's conflicting information in the legislation, right? And in the regulations and in the guidance. Right. We have three, three bodies of law, essentially. Am I right on that? I mean, I'm not a tax guy, but.

Roger Harris: We live in a gray world. It's not black and.

Annie Schwab: White, right?

Blake Oliver: Right. Okay. So you've got you've got the tax code, you've got court cases, you've got regs. And then isn't there even like another category of stuff that's not regs. That's just like.

Roger Harris: You can have FAQs.

Blake Oliver: Yes. [00:32:00] Right. So so that's that's where AI struggles. Right. Because you ask a question and it's going to go out and it's going to like find the first answer from whatever source, and maybe they find an FAQ and it gives you the answer from the FAQ.

Annie Schwab: They won't break the source as like a primary or a like it won't pull from, let's say, before it took an article from Accounting Today or, you know, professional journal or something like that. It would take a primary source from, you know, the IRS.

Blake Oliver: So [00:32:30] like these models are often trained to be fast, not to necessarily be thorough. So if you want it to be thorough, you have to tell it to write. You have to like again remember this is a smart intern here. So like if you ask a tax question of a smart intern, that intern might go do a Google search and give you the first thing they find, which happens to be like an Accounting Today article with an answer. And it's not totally lined up with the facts of this case. And so the details [00:33:00] in tax are so important. And you can't. One shot the answer and one shot means just do one prompt and get the answer.

Annie Schwab: Yeah.

Blake Oliver: So you have to know how to ask all the smaller questions.

Annie Schwab: So is that why you're saying they're not going to replace us? Because it needs that verification, that it needs the human verification that all the facts have been applied?

Blake Oliver: Yeah. So it's it's like that, that that [00:33:30] question that we ask it or that, that like here's a situation, give me the answer is actually like a hundred smaller questions that we are doing in our brains as part of that bigger question. And the AI cannot yet do that. Now, there's certain, um, models that are very expensive that claim to be able to do this kind of work. They can do like scientific, they can answer very complex scientific questions that are and break it down into multiple steps and [00:34:00] get to the answer, um, you know, you have to pay hundreds of dollars a month and they're not always accurate. Still and there. And I guarantee you that like the companies like OpenAI are like losing money on this because they're they're having to do like this chain of thought prompting inside that's hidden, that is just costing lots and lots of money. So it's like not actually it's so inefficient that it's not going to be cost effective necessarily. Like to [00:34:30] actually have an AI that can do tax analysis like you do. Annie. It might cost more than you. Seriously. Like that's why Microsoft wants to go buy a nuclear power plant.

Roger Harris: Yeah.

Blake Oliver: So it's funny, um, like, the human brain is a miraculous thing that we can essentially do prompts every second, and we can intake terabytes of information through all of our, all of our senses and, [00:35:00] and make sense of it. Our brain is just a a magical thing. And, um, you know, these models are like, doing, like, a tiny little fraction of that for a very high price. Yeah. Yeah.

Roger Harris: You know, what's interesting is because when you talk to people in our industry and he said, we've gone around to IRS forms, they all jumped to taxes being the first place. They think it's going to be the magical solution or whatever. Yeah. And yet you were very quick to say that's not where I would start. Well, because.

Blake Oliver: When people think of tax, they think [00:35:30] of the administrative stuff.

Annie Schwab: They just think of putting numbers in a on a form and letting it calculate the end result.

Blake Oliver: I mean, and then when they think of accounting, they think of the bean counters with the green shades and the calculators. And what they forget is that, like once you get past that, it's really complicated. It's like, that's why like the best, uh, the like the people who are like, really good at those insanely complicated, you know, board games that take hours and hours [00:36:00] and hours and hours to play with all the pieces, the hundreds of pieces. It's always like. It's always people that I know that are like tax people or have something to do with, like regulation, like like some sort of insane, insanely complex, like, you know, filings you have to do with the SEC, like, it's it's it's it is. These are some of the most complex systems, the tax code, the regulations. Oh, yeah. There's some of the most complex systems that exist in society. And yeah, it's it's not easy. [00:36:30] And so it is funny to me that like, yeah taxes the thing. But that's because everyone's thinking about like personal tax.

Annie Schwab: Right.

Blake Oliver: Just take a W-2 and put it into a form and file it like what TurboTax does. I think TurboTax is already done. There's going to come a AI is going to do what TurboTax does and do it better. It's just a matter of time.

Roger Harris: Right? Because it's basically just taking numbers and putting it in the right spot. [00:37:00]

Blake Oliver: Exactly.

Roger Harris: And making sure that in the calculations are done.

Blake Oliver: It's so deterministic.

Roger Harris: Right.

Blake Oliver: And but like yeah. If, if you're like dealing with like I don't know one of these you know foreign like a foreign tax situation with like multiple entities nested all together or like real estate or, you know, any of these, um, nonprofit accounting with fund tracking across multiple years, like, all this stuff is like it's hard.

Roger Harris: Yeah. And all the areas where it's not just again, I'm oversimplifying it, but [00:37:30] it sounds like AI is great in a world of black and white. But most of what we have to do in our brain in taxes is great. So we gotta take a little bit of red. I mean, a little bit of black, a little bit of white and figure it out. We're not there yet with AI.

Blake Oliver: And there's a risk assessment right as well.

Roger Harris: Yeah.

Blake Oliver: And yeah, the gray area, the. That's a great way to put it. Roger I'm going to start I'm going to use that because there's so much gray area in accounting and tax and in everything we do. Like there's a lot of.

Annie Schwab: It never gets easier. [00:38:00] It just gets more complicated. Right. Look at this new this new, enormous piece of legislation. Um, and I've been using some of the AI tools for it, and it's. I feel like it's it's learning what Oba has in it. Yeah. Like, my my results are getting better as a as more time goes by as it relates to it.

Blake Oliver: One of the problems I had researching it, using using AI to do it was like there was so much that changed [00:38:30] from the draft to the final legislation that, um, I got something wrong on our podcast.

Annie Schwab: I did oh, I got I've had several searches that were that I know are were blatantly wrong.

Blake Oliver: And it was it was a legit source, but it was outdated. So it was like I was referencing a Tax Foundation blog post that had been changed when the legislation finally passed and the the model hadn't, couldn't figure out like the date on that. And so it gave me the outdated information. I went with it. And [00:39:00] so it just shows you that like even simple things like that, like facts and time. Um, it can struggle with. So.

Roger Harris: So what are some good things again, thinking of people in our profession. If it's you know, tax is not the panacea. That's not where you're going to go and see your life change overnight. So is it something that you would hire that intern to do? Should you think of it that way, if you would hire an intern to do it? Ai might be a better place [00:39:30] to get it done.

Blake Oliver: All the administrative minutia that gets in the way of actually getting the numbers into the form, or it.

Roger Harris: Takes a lot of time.

Blake Oliver: Yeah, I mean, that's that's honestly like just getting the information we need from clients in a timely way so that we can actually do the analysis like that, to me is a great application of AI. There are tools that are being built. I don't have any recommendations at the moment, but there's tools that will just like go through your document [00:40:00] storage and like look at a folder of client documents they've dumped in and just like, organize them, rename them and put them into your tech software.

Annie Schwab: Oh.

Blake Oliver: Right. That that sounds good. Yeah. Like just here client. Here's a link and go and and just upload your stuff and you know, now my, my AI admin is going to put it all in and tell me what I don't have and tell you what you need to keep sending me and, you know, handle that.

Annie Schwab: Yeah.

Roger Harris: Um, yeah.

Blake Oliver: That's a great.

Annie Schwab: One.

Roger Harris: Yeah.

Blake Oliver: Yeah. [00:40:30] Um, I like one of the demos I did for, um, for Paget. Uh, when I did my session presentation on this this summer was responding to IRS notices.

Roger Harris: Right.

Blake Oliver: So, like, handling correspondence, um, understanding what this notice is and what we need to do, and just even, like, error checking. So, like, um, I had to respond to a notice, and I, I, I used AI to help me do it. And then once I had done it, I used AI to check to make sure I had everything that I needed to make sure it wouldn't get rejected. Like [00:41:00] that's that quality control can be really cool. Um. Obviously the research is the big thing. Everyone's talking about doing research with it. Um, I'm trying to think like in terms of of tax. I think there's, I think, yeah, quality control is actually something that AI could be really helpful with. So if you're comfortable, like uploading a draft return along with like client source docs [00:41:30] into one of these tools, you could have a go and like do the like give it a if you have a checklist in your firm for like reviewing a return, I think AI would be very good at following that checklist and identifying anything that's missing or needs to be reviewed.

Roger Harris: Which could be a huge savings in a tax firm where you're using a human to check another human's Humans work, which means you're not. Neither one of you are doing tax returns. You're checking tax returns that have already been done. Just let AI be your checker, and [00:42:00] you can reserve the human for actually the interaction with customers and the thought processes to do the.

Blake Oliver: Exactly. Yeah, that reviewer is very expensive. Um, and then you're probably reviewing it as the partner in the end anyway.

Annie Schwab: So yeah.

Blake Oliver: You know, it's interesting is there's an example of this in the medical field. So in Britain one of the earliest studies was radiology using AI in radiology, looking at, um, the scans for cancer.

Annie Schwab: Interpreting the scans.

Blake Oliver: Interpreting [00:42:30] them because that's something like radiologists get trained for years to be able to look at blobs on these scans and figure out is that cancer or is that not? And it's like an art and a science, and it takes a lot of experience. And so, uh, AI researchers took thousands and thousands of scans. And then the related Diagnosis and trained a model specifically on how to recognize cancer. And the model is actually as [00:43:00] good or as or better than the human. And so it hasn't replaced the radiologists there, but now they're using it to verify the radiologists diagnosis.

Annie Schwab: It's like getting a second opinion or something.

Blake Oliver: It's an instant second opinion.

Roger Harris: Right.

Blake Oliver: And and so it's reducing the false positives and it's reducing the false negatives and it's saving lives. And we could do that with tax returns real easy I think.

Roger Harris: Yeah. Well [00:43:30] if we're going to save many lives we might we might not kill as many people by making the mistake.

Blake Oliver: Maybe maybe we save some states, save some estates, save some, uh, save some tax bills.

Roger Harris: Yeah. No. We might. I think we've scared people to death, but I don't know that we've ever. Well, I don't know. Maybe we've brought a few people back to life when they got a refund. They weren't counting.

Blake Oliver: That's right. You never know, right?

Roger Harris: No, you never know. You never know what happens after they leave your office. Yeah, but that's that's a good example of how the, the the routine [00:44:00] tasks, the checking. So like right, right. Now think of it as a really good, um, intern, really smart intern. And the things that if they were doing that you could trust and count on and reserving your time for the things that are uniquely something that your skills are needed for.

Blake Oliver: The other way to think about it is that it's like a, you know, personal research assistant. Yeah, and [00:44:30] not just for answering questions, but more for giving it a situation and asking, is there anything I'm not thinking of? Is there anything I'm missing? Like, give it your analysis and say, here's the situation and here's what I Believe poke holes in my argument. B b the tax court and challenge me. What could I be doing better? What am I not thinking of? And that's where it really [00:45:00] it really succeeds because we only have a very limited knowledge. There's only so much that we can read. And you know, in the tax world, people can only really be experts in like part of it. That's how big it is, right? You can't do it all. And so, um, the these AI tools can like help us to know what we don't know, which is often the thing that gets us right.

Annie Schwab: Something you just didn't think about. Right, exactly.

Blake Oliver: So I use it all the time for that.

Roger Harris: That's such a big addition. To what? Our typical [00:45:30] research. I mean, I can go back far enough to remember when we used to have to build a room to put all the books in for research. Right. Then we we moved on to online research, but it could never challenge us like you're talking about. I mean, we could basically go look up something to help us come up with an answer, but it could never challenge our answer.

Blake Oliver: That's what it can do now. Yeah, yeah. So not about giving us the answer. It's about challenging our answer.

Roger Harris: Yeah, yeah. Which is a completely [00:46:00] added dimension to what we have done or typically do, for whom or how we utilize research products in our history.

Blake Oliver: And that's why it takes a different mindset. And and curiosity is so valuable if if you have a curious person in your firm and you give them this tool. I mean, they can be I can learn anything I want to know. I can learn anything so fast. Um, but but I but it has it takes somebody who wants to ask those questions. [00:46:30]

Roger Harris: Right?

Blake Oliver: Yeah.

Roger Harris: So you think it's here to stay, huh?

Blake Oliver: It's going to get better and better. It's. You know, it's fun being able to, like, talk to it. Um, that that I find is, like, a lot easier to. It's like I really recommend trying the voice mode. You can do that. Yeah. Because like, the problem is here's the challenge is like it's easy to type one question. But again, if you don't give it all the context, the answer is going to suck. So to give it a lot of context, I mean, [00:47:00] I'm writing paragraphs for a good prompt, and I find that I just hit that little microphone icon in my browser on my phone and I just say, I just like, say what I need. And this takes a mindset shift, too, because we've all been trained over the years to like, do a Google search a certain way. And remember when we had to use keywords and syntax and so we. Yeah, and it was only until recently that we were able to actually start asking natural language questions like a full sentence. And so, like all of us who grew up in the era of Google have to retrain our brains [00:47:30] to say, it's okay for me to just talk to this thing like it's a person.

Annie Schwab: Mhm.

Blake Oliver: Um, and we're, we're seeing that the younger generation is able to do that faster because they never, like, learned. They're learning how to search on AI first instead of Google Now. So they're like getting this. And you can see it because they're like becoming friends with AI and falling in love with AI and like, having psychotic episodes because of AI. Um, that's a whole other like, realm. That's just fascinating. Like, I feel like [00:48:00] it doesn't happen to people 40 and up because we we we don't think of tech that way, but we need to we need to start thinking about it that way. Not to that extent, but to some extent.

Roger Harris: Yeah. Can you talk to the car? Can you can you talk to the driver and say.

Annie Schwab: I changed my mind. I want to go here.

Blake Oliver: Yeah. Yeah. In the Waymo. Yeah. You can, you can, you can ask it to like you can press a button and talk to the AI. And by the way there's [00:48:30] a human that takes over if there's a problem.

Roger Harris: Oh okay.

Blake Oliver: Like they'll intervene. There's actually like a, um I believe they have like a whole, like a bunch of drivers that are all in, like, um, have controls and, and they can switch between all the way modes. So like if there's an emergency, a human will take over and pilot it remotely.

Annie Schwab: So crazy to me I am not ready for I am not ready for that. I can't even imagine being like, okay, here kids go to school and just get into this car with no driver. [00:49:00] And, you know.

Blake Oliver: It's, um, it's interesting. I didn't even think of this as a man. But like a lot of the women that I talked to say they like the Waymo's because they.

Roger Harris: Yeah, I read an article.

Blake Oliver: Safer.

Roger Harris: Yeah. They feel actually feel safer. Yeah.

Blake Oliver: Um, and, um, I could see that, like.

Annie Schwab: Meaning safer that you're not going to wreck or, like, safer that you're not like, your driver's gonna, like, steal your purse.

Blake Oliver: Yeah. That. Yeah. Like, they don't they feel safer with the the AI than with a male driver. Who.

Roger Harris: Yeah. Because think about it. You're getting in a car with a strange [00:49:30] man you've never met before. Uh, and you trust it because you got him from an app called Uber or Lyft, right?

Annie Schwab: Right.

Roger Harris: Uh, Versus a car with no human in it. That'll get you the same place.

Blake Oliver: So I think I'd rather put my kid in a car with an AI. Honestly. Um, yeah.

Annie Schwab: So we'll see. We'll see. I don't think I'm there yet. Right now they can just ride their bike. Okay. Just ride your bike.

Blake Oliver: I'd rather him do that than drive a car himself.

Annie Schwab: Yeah, well, that's true too. Mine are not driving age, so I haven't [00:50:00] even started that fear.

Blake Oliver: And these days, the kid, the kids all want e-bikes, which are basically like motorcycles.

Annie Schwab: Yeah, I.

Blake Oliver: Know that's that's horribly dangerous too, so I don't know. I don't know, I'll sign him up for Waymo. I'll pay for that.

Roger Harris: Well, you wait till we get driverless school busses. What's the parents panic when.

Annie Schwab: Uh, I don't.

Roger Harris: Know, the bus pulls up with no human, uh, you know, but it's going to happen. I mean, I'm sitting here laughing about it.

Blake Oliver: And driverless trucks on the on the on the interstate. Oh. [00:50:30]

Annie Schwab: Like that big 18 wheelers.

Blake Oliver: Yeah. So the one of the companies that's closest to getting regulatory approval for that is based in Phoenix. And so that'll actually I mean, you know, think about how think about how irritating it is driving across the country. And you've got these like truck drivers and most of them are great, but then some of them, just like it seems like their mission is to block traffic in the fast lane, right?

Roger Harris: Yeah.

Blake Oliver: So that'll be over. And we'll actually all be able to pass them and they'll be, you know, properly distanced. And nobody will care because they don't have to take a lunch [00:51:00] break and get there on time or whatever.

Roger Harris: Yeah, yeah. Are there are they drive past whatever their designated hours they can drive, but they're so close that they just keep going. No. But you know, it's it's it's part of life that everything changes. I mean, when you first hear it, you go, I think I heard and I'm assuming this is AI. They had a robot Olympics somewhere. I think it was in China or something.

Annie Schwab: Oh, wow.

Blake Oliver: Yeah. The, um, it's fun watching those videos of, like, the Tesla robots and the Boston Dynamics robots. [00:51:30] They're getting so good and so close to being able to mimic human motion. It's really just like the the fine motor skills, the fingers that they're working on now. But once they get the fingers, then we will be able to have fully autonomous robotic manufacturing facilities. And why is that exciting? Well, it's because and this all ties into tariffs and everything that's going on. Um, the the biggest cost in manufacturing in this country is labor. We just don't have enough [00:52:00] people. It's very expensive. Uh, and compared to other places. And if we have fully autonomous factories, it means that we'll be able to basically import raw materials and energy or hopefully produce our own energy. We need more nuclear. But if we can get and we have lots of raw materials in this country, right. North America is extremely rich and wealthy and raw materials. So if we can start extracting them in an environmentally responsible way and we have robotic manufacturing facilities we can make, that's how we could actually make everything [00:52:30] here. Right. Or a lot of it. And basically that will drive and.

Roger Harris: The affordably.

Blake Oliver: And affordably. It'll drive the cost of goods down to close to zero. And so that's why I'm not worried about, like, people losing their jobs and not being able to live. Because yes, people will lose jobs and will have to make new ones, but the cost of living will continue. This trend that has continued over 100 years or more to where I mean, we live like kings and queens compared to just 100, 200 years ago. You know, [00:53:00] 100 years ago, hardly anyone in this country had ever had a hot shower, you know? And now we all take it for granted every morning. So it's like we just we have to, like, look back at the scale of human history and think just how fast things actually have moved. But it's funny because when you're in it, it feels really slow.

Annie Schwab: Slow, right. Exactly. I can see that. It's because you're you're you're anxious for the end result to get to the, the best AI or the best tool or whatever it is. But. Well, this has been [00:53:30] interesting. I definitely am going to have to go check out this Waymo thing.

Blake Oliver: Yeah. Come to come to Phoenix. Come to Scottsdale. You can take one from the airport. That's the best way to experience it.

Annie Schwab: Okay. I'll put that on my bucket list.

Roger Harris: I don't think here in Athens, Georgia, we have any Waymo's riding around, but I would. I wish one could take me to the Atlanta airport for sure.

Annie Schwab: Yeah, that would be nice.

Blake Oliver: Well, you're never getting to the Atlanta airport anyway. That's what I hear from everybody who lives there.

Roger Harris: So, yeah, you can't get there any faster than I can, because it's got to [00:54:00] deal with the same traffic that I do and the construction and all those other things. Maybe they'll build one that can just when it gets in traffic, just lift off the ground and fly over it.

Blake Oliver: So they're working on those two. Those uh, yeah. Those air taxis. That's going to be great I can't wait.

Roger Harris: Really I don't know.

Annie Schwab: Really weird.

Roger Harris: Different world, different world, but a good one. I mean, this is all good stuff and it's all things that, you know, again, a few years from now, we're going to look back and act like what was the big deal.

Blake Oliver: Yeah. I mean, just like we have everything in the cloud now, right? Yeah. [00:54:30] It's just normal. That's where we start from. And remember how like how I know you must remember, Roger, when people started first doing that, that was like a big deal.

Roger Harris: Oh yeah.

Blake Oliver: Yeah.

Roger Harris: Like, where's it going? Where's my stuff going? It's up in the cloud. What cloud. Yeah. It was.

Blake Oliver: Yeah. And now we just take it for granted. I mean yeah, my podcast used to be called the Cloud Accounting Podcast because that's how Big Cloud was. And we had to change the name. We had to get rid of the cloud part because it just became the default. Right. And that happened within like five years.

Annie Schwab: Mhm.

Roger Harris: Well so [00:55:00] and that's what I mean if we're going to stay in this industry and stay in this business, we want to be maybe not the lead dog, but we sure don't want to be the tail end. No.

Blake Oliver: You want to be in that early majority right. Yeah.

Roger Harris: And you'll have a better firm a more effective firm. You'll be better at what you do. So it's not going to put us out of business once we learn how to adapt it and use it properly.

Blake Oliver: And the good news is that we are just tipping past that point of the like, Vanguard. Right? So the people who are on that bleeding edge, right? [00:55:30] Like, that's like me. I've been playing around with this stuff for a few years now. Now we're starting to get to the maturity point where the the early majority is going to start adopting it, and that's going to be like a ten year process.

Roger Harris: Right?

Blake Oliver: So you have the if you're listening, you haven't done anything yet. You're not too late. You are still like in the first year, you're like in the 90%. Right?

Annie Schwab: You're still but get moving on it. Right. Yes. You're playing with it. Yeah.

Blake Oliver: Yes.

Roger Harris: It's going good. It's going to actually make your life better. And [00:56:00] I'm all for that.

Annie Schwab: Yeah, yeah for sure.

Roger Harris: All right, Blake, we gotta get you back and talk more, because probably we could get you back in 30 days and something. Something radically different.

Blake Oliver: It's always changing all the time. So thank you for having me on. And yeah, it's been it's always fun chatting with you and getting your perspectives as well.

Roger Harris: Yeah. Yeah. And we can get you back and talk about a lot of other things. Interesting industry, interesting times. Technology is part of it, but there's a whole lot of other crazy things going on.

Annie Schwab: Yeah.

Roger Harris: Well. [00:56:30] Hang in.

Annie Schwab: Yep. Well, this was great. Thank you so much. Um, Roger, any last remarks?

Roger Harris: No. Just, uh, thank you, Blake, for joining us. I thank all of you for listening. Um, we're going to continue to monitor the end of this year. I know there's a lot of questions about the OBB or whatever they call it, and government shutdowns and IRS staffing and technology. It's an interesting time to be in this business, and we're just going to continue to monitor it and try to stay one day ahead of [00:57:00] the craziness.

Annie Schwab: Yes, we'll bring you whatever we find out when we find it out. So tune. Tune back in, um, for our next.

Roger Harris: Don't forget to listen to Blake's podcast. Um, yes.

Blake Oliver: Listen to the accounting podcast wherever you find podcasts, YouTube, Spotify and.

Roger Harris: Found this one. You can find his because we're on his platform.

Blake Oliver: So earmark actually, that's where everyone should go is earmark app. Get the free earmark app for your Apple or Android device. Sign up for free. Earn free continuing professional education credit for federal [00:57:30] tax updates, The Accounting Podcast and many other fine podcasts. What better way to get your continuing education out of the way?

Roger Harris: We're not near as boring as sitting in a class for an hour.

Blake Oliver: Do not do your CPE between Christmas and New Years this year, right? That's not what it's for.

Roger Harris: That's right.

Blake Oliver: I know you may. You may not want to spend time with your family, but you can do that in other ways. So get your CPE done and then avoid them out on the golf course if that's what you want to do. Yes.

Roger Harris: All right. Well thank you Blake. [00:58:00] Thanks, Andy. Thanks, everyone for listening. We appreciate it. Uh, we hope you'll come back soon for another federal tax update podcast and catch Blake's podcast all on earmark. So thanks everybody. Have a great day. We'll be back soon.