Recent bank failures and pressure on liquidity have shined a light on how much digital has changed the banking industry. Alex Habet and Dallas Wells discuss the anomalies of the past few weeks and what regional banks can do to be proactive in this volatile environment.
Listen
Watch
Subscribe
Transcript
Alex Habet
Hi, and welcome to the Purposeful Banker, the leading commercial banking podcast, brought to you by Q2 PrecisionLender, where we discuss the big topics on the minds of today's best bankers. I'm your host, Alex Habet.
So, it's mid-March 2023 as of the time of this recording. We're about two weeks out from the news around the failures of Silicon Valley Bank, Signature Bank, Silver Gate Bank, and things are still swirling. We're not here, and we're not coming onto this show today in this episode, to necessarily add to the noise. It was very important to us not to do that. But given that this is a banking podcast, after all, we felt it was important to react to what's going on in the news.
Now, it certainly seems like everyone out there in the media space is all of a sudden an expert. And even here on this show lately, it started to feel a little bit more like we're covering banking 101 type topics, just recalling how we never ... or at least not so deliberately used to ... focus on things like the differences between interest-bearing versus non-interest-bearing deposits or time deposits, etc.
But things are going into overdrive now. And on the heels of these failures, it's important that people are understanding what the even more basic explanations are of what's going on. The banking system, by and large, is a marvel of society, but none of it is possible without trust as its foundation. And from what I've been reading out there in the news today, and leadership teams across all the banking franchises out there, it's all hands on deck about projecting trust these days.
So, we wanted to come back here and offer some perspective that we hope is helpful, to at least cut through some of the noise and offer some things to just think about. It's a little bit of a dizzying time these days, but here to help make sense of it all, we wanted to welcome back, of course, Dallas Wells. Dallas, welcome back home to the Purposeful Banker. We're thrilled to have you here because you're at least one of my favorite people to talk to around this topic.
So Dallas, welcome back. How's it going?
Dallas Wells
Thanks, Alex. Going well. It's been a wild few weeks, but glad to be here to talk through it.
Alex Habet
So look, as we're sitting here listening to the experts, or in some cases enduring the echo chamber a little bit more, I don't know about you, but you can't help but wonder why we potentially, or why we feel like we fell victim to the mismanagement of banking basics. We had early indicators of trouble, and we even talked about it on this show, specific to the institutions that are now dominating the news, but it was more of a "Hey, what's a deposit strategy around this?" We weren't necessarily thinking about how they're managing their balance sheet around it. We assumed all of that stuff was under control, and it wasn't a thing even just a few months ago. But how did we miss this as industry watchers and players? How did we go through this black swan event and it felt so shocking all of a sudden?
Dallas Wells
Yeah, when you just talk about some of the issues that banks are facing, it's things like long duration bonds, and we're not talking about subprime mortgages, we're not talking about CDOs. We're not talking about derivatives. We're talking about treasuries, and we're talking about agency mortgage-backed securities, that because of the pace and the degree of rate increases, are creating these unrealized losses. And let's be clear, the loan book has the same kind of thing. They just don't get marked to market like the securities portfolio does. But it sounds like the basic banking business. You take in deposits and you go out and you buy safe securities. You take little to no credit risk on that bond portfolio, and all should be well. And I think that's largely why this risk snuck up on everybody is, this feels like, you said the word banking 101, that's kind of what this felt like.
But what got us was not this exotic credit risk this time. And it wasn't a new kind of securities, it wasn't a new kind of banking. It wasn't even really new business models. It was just good old-fashioned concentration risk. It was good old-fashioned interest rate risk, that came around for the first time in a good long while. And it feels like a lot of bankers and a lot of industry watchers got caught by surprise by maybe how fast things changed. And we all got caught, oh wait, this stuff snuck up on us and we should have known better. So, it was really interesting the way this landed at our doorstep. This is stuff that you learn at intro to banking stuff, that shouldn't take down the kinds of institutions that it did. And yet here we are.
Alex Habet
So, I want to tug on that one thread that you mentioned around concentration risk first, because I think even in previous episodes with you, we've talked about one of the best ways for certain institutions to compete against bigger players, is to specialize or to concentrate a little bit on a segment. So, does that still make sense? I think the answer is yes. I don't want to oversimplify it, but I think the answer is yes. However, does it only make sense if you plan on staying small? If you're getting to that medium size institution, does it become more dangerous? Should specialization or concentration give way to different priorities at that point?
Because as you know, SVB had the lion's share of the tech startup business out there. And guess what? They earned that. From the VCs to the tech startups themselves, they have countless stories about how SVB has been pivotal or instrumental in the success of those franchises, and even admits the failure of the bank. I'm sure you saw those letters of the VC pledging to continue to bank with them if they were to survive somehow. So, they specialized and they succeeded. They captured a huge market share top in its class, but they committed that sin of being too concentrated at the end of the day, and they became too vulnerable amongst other things. So, how does that whole strategy evolve?
Dallas Wells
Yeah, so you're absolutely right. We've talked a lot about how, especially community and regional institutions, they shouldn't try to compete by being all things to all people. And that differentiation used to be just geographic, where the bank of whatever county, and we're close so bank with us. That was the old-fashioned approach. And now, the differentiation needed to be specialization. And you're right, Silicon Valley Bank nailed that part of it. And so, I don't think the issue is necessarily that their deposits were concentrated. It's just that that was a risk that should have been managed and mitigated.
So, with a business strategy like that comes a particular risk. So, if you look at the charts of SVB's deposit growth. And heck, just look at it since the beginning of COVID. So, if you go from where they were in '19 to where they were at the time of failure, you're talking about a three to four X deposit growth. So, that strategy was working. As the tech world, the tech boom exploded, they were as you said, they nailed that market and they were capturing a lion's share of that growth. But then, they concentrated in a second way, which is they went out and ended up with a concentration of long duration bonds with a ton of interest rate risk. And so, when some of those deposits that had ramped up so quickly started to bleed back off because of rate changes, because of the slowdown in the tech market, because of the slowdown in tech funding, they could not shrink down the asset side of the balance sheet as easily. They had to realize losses as they were doing that.
So, it was the concentration on both sides. It's a fine strategy on the deposit side if you stay liquid, if you know that those deposits can come and go quickly, especially when they increase that fast, that you have to maintain some flexibility. Or if you're going to take that risk, you've got to hedge it. So much of this was unhedged. And so, the specialization concentration strategies are OK. It's just that there's always a trade-off for those. There's always a specialized risk that comes with that, that you just have to be aware of and you have to manage through. And they had done that very well for years and years. And just this unique set of circumstances caught them in an unexpected way.
So, we've talked about a few other examples of institutions that have specialized in an industry. I think they all probably are taking a step back and evaluating that. And not necessarily saying, "Is the strategy wrong?" But what about the strategy could serve up that same black swan combination of risks for us that we just need to be aware of, and we've got to have an exit plan, a plan B and C of how to step back from this if that toxic combination of things happens.
Alex Habet
So, you're not bearish on the strategy, at least the overall mission of the bank per se, but it's definitely necessary to take a second look on how that risk is managed in a much more cohesive way, is what you're ...
Dallas Wells
Yeah. And look, the Monday morning quarterback view, the hindsight view is really easy for all of us to take. But rewinding it backward, if they had looked at that three to four X deposit growth and said, "Gosh, what if some of this just goes back out just as fast as it showed up?" So, they stretched for an extra, I don't know, 50 to 70 basis points of yield by going out on the yield curve. What if they'd kept a meaningful chunk of that shorter, and made less in earnings because of it? But when that toxic combination of things happened, they would get to live to fight another day. So again, that's easy for us to say looking backward, but that's the trade-off that if you have a potentially volatile source of funding, that is concentrated like that, your asset decisions just have to be different.
Alex Habet
Yeah, that's excellent. So, let's shift a little bit around some other aspects of what happened. And probably the one that dominated almost to an equal level of the banking 101 type issues that we just talked about. Look, as someone who leads the product teams here at Q2, this is something you just live and breathe every day. The first banking crisis in this age, the last one that happened, at least at a large level, happened before everyone had smartphones. When we're studying what happened here, or what is still happening right now, how do you think ... And this is just a hot take on Dallas, because it's impossible to tell right now, and I don't want to put you in that spot. But I want to ask you just, what do you predict the postmortem will look like to the technological impact to this? Whether it's things like the tools that the banks give to their clients that now make it easier to move money, or things like social media, of course, and all that? So, how do you think you're going to do in that postmortem?
Dallas Wells
Well, let's do a quick restatement of how this went down, because I think that's really interesting. This was Silicon Valley and a few of the follow-on institutions that have hit some rough patches since then, that were next up in the crosshairs. But the run on Silicon Valley happened live transparently in real time on Twitter. It was the craziest thing I've ever seen. So, used to a bank run would be, again, a very localized event, but it would be people lining up at branches. And the word of that spreading, takes a little while and then it takes a little while to move the money out. And it was like a multiday process.
This started midweek, and by Friday morning, the FDIC was in control. It happened really fast. So, if you picture what's happening to an individual, say tech company that does their banking at Silicon Valley Bank, they're probably also on Twitter, a fair amount of the time. They see on Twitter, some of the noise, they see some of the venture capitalists out there saying, "Hey, this looks bad. We're advising our portfolio companies to move their money." You can flip over from Twitter to your online banking app and transfer that money out in about 30 seconds. And that's what happened, at a large scale.
And we've had this conversation inside of Q2, some of those tools that folks are using to move that money out are things that we build. And it was an interesting weekend for us being prepared for what might happen Monday morning, to be ready to handle all kinds of volumes of stuff just like that. So the question is, is the technology a problem? You can't put that genie back in the bottle of giving folks quick, easy, instant, transparent access to their funds.
But instead, again, going back to what we talked about a little earlier, of the double concentration risk. In SVB's defense, we've never seen a bank run happen like that. So, all of the liquidity models that they were running, I'm sure, because all institutions run them, there's some assumptions you make about how many deposits are really at risk and at what time would they bleed down, and what would your reaction be? That everybody does dry runs of this. Or they're tabletop exercises of how to deal with a liquidity issue. Those models are all wrong, I think is what this showed us. The technology has changed things, and that box is not going to be closed at this point. The box is open. And in fact, it's going the other direction. So, FedNow, real-time payments, and the rest of the world is actually a few steps ahead of the U.S. on this stuff.
So, it's easier than ever, and will continue to get easier to move money between institutions. And all the float, all the dead time, all the time you might have to react is shrinking down from what used to be days to seconds. So, I think it just goes back to the models need to change and the risk evaluation needs to be adjusted. And we also are working on very outdated models in terms of rate sensitivity, just because on the deposit side, we've been in a zero-rate environment since 2007.
So, rates went up a couple times since then, but not on the deposit side because nobody needed deposits along the way. And so, all of a sudden there were assumptions in there about, well, gosh, we've got all these non-interest-bearing accounts. Our models tell us from decades ago, those are not rate sensitive. Rates can move up and down. Those are operating accounts, they're fine. What we found out is that money was parked there because the difference between zero and 0.5% for somebody that's got 10 grand in their account, it compares. You're talking about a few pennies. If all of a sudden that difference is between zero and 5%, OK, now I'll move it. Now there's a reason for me to move a portion of that balance. And not only is there a reason for me to, but again, I can pull it up on my app, I can shop around for what my bank has, I can shop around on bank rate and see what else is out there. I can open an account online and all that can be done in just a few minutes and boom, I'm yielding something much better.
So, all those things combined just means that maybe your deposits are not as solid as you think, and they're not as solid as your models are telling you. So, I believe SVB did their homework. I believe that Signature Bank did their homework. I believe that the other regional banks ... I've seen some of those models myself, that these institutions run, it's just that those assumptions are maybe stale and the technology's moved a lot faster than we've adjusted for and accounted for.
Alex Habet
That's an interesting perspective. I didn't quite contemplate that. So, definitely think that's worthy of a look. And this is coming from a complete newbie in that part of the world, but is there no such thing as a circuit breaker when it comes to money movement? If there's all of a sudden too much moving, is there such a mechanism or is that just unfeasible with money movement?
Dallas Wells
Look, you're legally and contractually obligated to give access to the money, and it's the fractional reserve banking system at work. And that's the part of some of the ignorance of how banking works, that you saw happen on social media following this. It's like, my gosh, what were they thinking investing this in bonds, instead of just keeping it on hand in cash, or keeping it overnight at the Fed? It's like, what? That's not how the system works at all.
So, there's not necessarily circuit breakers, but I think that goes back to the risk management of being aware of anomalies happening and some early indicators happening in your deposit portfolio. Because to be clear, this run was the death nail, but the deposits had been bleeding off for a while. And it was a combination of rate sensitivity and just some changes that were inherent to that technology market for SVB, and some crypto stuff for Signature that caused the deposits to have been declining for a little while. And they'd sold down the liquid assets that didn't have big losses on them to meet those withdrawals. And now it was like, OK, the next setup we've got to eat some losses on, and it gets a little harder.
So, I think that's the risk management part where you've got to see that coming and again, have the plan or the hedging already in place to be able to have access to safe and sound funding or assets that can be liquidated. And look, that's not free. It costs earnings for that to be in place, and that is really why you've seen the broader set of bank stocks out there in the universe come down so much. It's not that everybody's worried about, they're all going to fail. It's that, OK, they've all got to be more cautious, and this rate impact is real, and we haven't seen a rate increase at this pace since the '80s. And so there's going to be earnings implications to that. And so, the entire sector has some earnings headwinds, and I think that's really what it's about.
Alex Habet
So, let me ask you this then, because another part of the talk around what's going on, is whether or not what happened constitutes a bailout. At least when the FDIC came in and basically guaranteed the deposits, even the uninsured deposits. Now, you and I understand the differences here, whether this is a bailout. It's not bailing out the institution; however, people are contorting themselves these days to try to explain the differences. And I'll admit that the optics don't look good to those who don't study this space very often. So, do you think the noise around whether this is a bailout is going to grow, or is this going to be a "It'll eventually get contained and we'll move on to more important debates?"
Dallas Wells
Yeah. This one's a really sensitive area. It seems like the Occupy Wall Street stuff was two lifetimes ago, but that was the reaction to the last time we had a banking crisis. And look, I don't think it's a stretch to say that the reaction of the federal government and the Federal Reserve was successful through the financial crisis. The system was on the brink, it was saved. TARP was actually profitable for the U.S. taxpayer. It worked. But there's no doubt that that qualified as a bailout, and some folks were very upset about that.
So, this time around, I think what you see is regulators, government officials being cautious. They don't want to wait too long to intervene. We don't want another Lehman type episode. So, they looked at not necessarily just the losses that could potentially happen to SVB's depositors that were above the insurance threshold, but the fear that that might stoke across the rest of the industry. So that's what they were really trying to contain, was they didn't want this to spread and for other institutions to all of a sudden basically for them to be losing any of their large depositors to the too-big-to-fail institutions, and we're talking about the top four, that would be the clear place of safety.
So, was it a bailout? Look, shareholders got wiped out. Debt holders are going to take their lumps as it all gets unwound. And I think what they made clear is, they don't want depositors to feel like they have to make a constant credit evaluation of their financial institution. So, that's on one side.
The other side is, there was some basic treasury management, risk management stuff that did not happen. When you look at the size of some of the accounts that were held at SVB, again, with no hedging on their part either, that's not very wise to have a 100% of your liquid reserves in one institution, and fully at risk, because there's plenty of tools to avoid that. So, I don't think it counts as a bailout.
Now, your question of, does this noise go away? I don't know that that's the environment we live in now. And I think you see the bank lobbying groups being very conscious of that. So, there were pockets of them that have been asking for increases to the insurance limit and things like that. But largely what you've heard from the industry is, "Hey, this is fine. We don't need an industry bailout. We don't need changes to the FDIC coverage. Our institution's fine. We haven't seen broad-based withdrawals across the system. This is an institution-specific risk, and we're not worried about it."
So, I think it did have the whiff of bailout, and I think the industry was very quick to try to distance itself from that and say, first of all, look, the insurance fund is not taxpayer money. It's taxpayer backed, but the industry actually pays for any losses incurred by the FDIC. If the insurance fund gets low, the rest of the industry's supposed to replenish it. So, the taxpayer's ultimately on the hook, but only after the industry can't fund it anymore, and we've got much bigger problems then. So, that's why they're cautious about that standing in there with a handout. So, we'll see how it goes from here, but for now, the industry seems to be trying to step back from that and saying, "We want no part of that or any insinuation that that's what's happening."
Alex Habet
Yeah. It's difficult to tell if it's just going to exacerbate things or not, but certainly appreciate your perspective there. Look, so for the rest of this episode, I wanted to shift gears and talk about a completely different topic, which I think if this stuff in the news wasn't happening, would be front and center of what we would be talking about instead. And to do that, I want to introduce the topic by playing you a clip, Dallas. And I just want to get your reaction in real time as I do that. So, here it goes.
Hi, and welcome to the Purposeful Banker, the leading commercial banking podcast, brought to you by Q2 PrecisionLender, where we discuss the big topics on the minds of today's best bankers. I'm your host, Alex Habet.
Alex Habet
Now, I don't know if you could tell, but that wasn't me. What that was, was a really incredible text to speech platform that I discovered. Where not only can it just generate these incredibly natural-sounding voices, but it also allows you to take your own clip, upload it, and then it'll replicate your own voice. So, I took an old episode recording from the show, put it in there, typed in the intro to the show, and that's what it produced. So, this is just a way to demonstrate how far things have come. It's a little tangential example of some of these AI tools that are now out there.
But to someone who doesn't know me very well, they probably wouldn't be able to tell that wasn't me anymore. And now, you're seeing that apply to things like image generators and document generation, and of course chatGPT, which pretty much seems like it could almost solve life's problems, but it's also equally terrifying just based on just even how creepy this stuff could get. And there's news coming out every day. Carl Ryden is on fire on Slack about this.
I saw the article recently that Microsoft researchers are even now claiming that GPT-4 is showing sparks of AGI, or artificial general intelligence, meaning it can pick up, it can learn a new task, learn from its mistakes, reapply again, and then solve problems just like humans do. It's mind-blowing that we're at this level. And so, it's impossible now to tell what the disruption is going to look like, because it's going to disrupt every industry almost at the same time, whether health care, banking, retail, whatever it is.
So again, as someone in your position, I'm sure you've been tracking this very closely. Well, where do you think the most profound potential is? And in your thoughts, how do you think the adoption of some of these disrupting technologies is going to compare, in the banking industry, for example, versus others? Is it going to be slower to adopt in that industry? Is it going to be faster? But you know what? Ultimately it doesn't matter. I just want to get your take on that too, because it's huge.
Dallas Wells
So, there's a lot of parts of it that are inherently terrifying. The fact that that wasn't you, that's scary that your voice can be out there saying who knows what. And your only answer is, "I didn't say that. That wasn't me." But it sounds just like you. And there's video versions of the same thing. That in itself is scary.
I think the other part that, from my conversations with bankers, the scary part about chatGPT is the absolute pace of the innovation. So, the incremental gains from chatGPT-3 to 4, and then on top of that, how fast people are able to build with it. So, how they can just put this as a layer on top of all the other things that they're doing. You can go spin up an AWS instance, and connect some stuff to chatGPT, and feed it a dataset or some text that you wanted to learn from. In two hours you've got a new product you can sell. The speed of innovation is a little scary.
Now, how does that fit in the banking world? A lot of the answers that these generative AIs are giving, they're wrong. And so, the error rates have to come down before, I think, bankers get really comfortable with it, but we'll get there. That's inevitable that those are solvable problems. And so, I think some of the early signs of what's coming and what ... Heck, I don't think I'm sharing anything too early. Some of the things we're working on is the concept of the co-pilot kind of through everything you do. So, I don't think I fear, or we fear, or bankers fear that generative AI starts taking jobs. It's just that what the humans can delegate to the bots keeps changing and keeps increasing. And the complexity of what you can delegate to the AI keeps increasing. And that leaves more and more time for the humans to do what they are uniquely suited to do.
And that's really exciting. If we can get out of the mundane, repetitive just filling out of TPS reports, which by the way is still in every bank we work with, a substantial chunk of your personnel costs is people looking up information and keying it from one system to another, or shuffling paper, or doing the digital version of that. A lot of that stuff starts to get automated. And I think what you also see is the humans get really good at using the tools that are at their fingertips. And right now, they're not very good at it. So a typical employee might have, I don't know, 60 SaaS tools or software tools that they're using for their job. And so, they're in and out of them. They know the basics of each one, but there's deep functionality in each one of those that they're not really using. So, there's a handful of power users that use them, but most people just are doing the basics.
These generative AIs are the things that can look over your shoulder. You can train them on the documentation. So, we have inside of Q2, thousands of pages of product documentation, really deep technical way in the weed stuff that no human has the time to read or to remember, but the bots can do that. That's what they're really good at. And you can just speak normal human sounding language to them and ask questions, and they can go peruse those thousands of pages of documentation and come back with, here's how to do it, and here's the best answer, and here's the most efficient way. That's super exciting. And it's not something that, oh, we'll be there in 10 years. These are models that we can just plop down on top of the stuff that we have.
So, we've got some skunk work stuff going that it works. So, will we roll it out to customers? Well, we've got to do some air testing. We've got to get everybody comfortable with it. But it's pretty exciting just to tiptoe into it by saying, "Hey, we can get your user manual. And instead of having that just be a thing that we ship you with every release, and we spend a whole bunch of money writing it, and then nobody reads it, how about instead, we feed the co-pilot who can look over your shoulder, which by the way, is Andi inside of PrecisionLender." We've been doing that since 2017, and I'll come back to that in a second. But it's not a new concept, it's just that the language models have gotten infinitely better. And so, now we've crossed the threshold there to where you actually get usable interactions and real support from those bots, that's not just another tool to learn. You can just talk to it like a coworker, and it'll go do what you ask it to.
Alex Habet
To use your example with Andi, we've been doing that for a long time. But as someone who spent some time a few years back trying to get Andi to do this stuff, and we achieved that it still required several steps, parsing information, monitoring how people are interacting with it, tweaking this and that. Now, you just upload a PDF document, and you get just the same level of output now, and that's incredible. Continue though. Continue. I cut you off there.
Dallas Wells
No, it's good. So, I'll give you an example that's not ... Well, it applies to Andi, but I'll use one broader for those who don't know Andi and use Andi. So Amazon Alexa had a promise that was similar. Hey, just tell Alexa what you want, and Alexa will go do it. If you've tried to do that, it's actually frustrating. So, you can do a few basic things like set a timer and check the weather, but anything very complicated that you ask Alexa, you have to know a specialized command. And there's a whole bunch of those. There's thousands of them, but you have to phrase it exactly right for Amazon to be able to retrieve that and do what you want done. Well, you can just put a chat GPT layer on top that has been trained on what are the thousands of Alexa commands that are out there, and what's the exact right phrasing that's going to send back to the Amazon servers, what you're actually wanting done, and then go through the steps to do it.
And so, you just ask in plain English, "I want to check my bank account balance." "Which bank?" And it's Bank of America. Then we just spin up the right Alexa commands to do that. And that happens in the background. We can do a similar thing with Andi now. Andi is a similar thing of super powerful, but we have to trigger that with some specialized commands, that we've tried to be pretty good with. But if you don't word it exactly right, we've tried to guess at what the variations might be. But if you're slightly off, Andi's going to be like, "I don't know what you're talking about."
So this just, again, adds to the sophistication of that language model, and it's a learning model. So, it's going to learn banker-ese, for lack of a better way to put it, and understand pretty quickly what bankers are actually asking Andi for, or ultimately what end users might be asking Andi or some version like that for. So, that's what's exciting is, you just get the interaction. Instead of having to have a translator between the technology and the human, that was some specialist who really understood the tool really well. I think you get to where all the users don't need that translator because that's built in. And then, when you get to the fact these things can actually write code now too, it's just bonkers.
Alex Habet
Yeah. Who knows what's going to happen with all this stuff. But that point you make about not having someone in the middle who can help the banker phrase the prompt the right way, is a bigger deal than most people would think, primarily because a lot of this adoption hinders on what your first impression is like. So, if you've never tried a chatbot, whether it's Andi or anyone else, you try it for the first time, if you don't get what you want back or what you're expecting back, you already are pretty much ruling it out as something useful to your life. And so, this eliminates essentially that first impression risk that you have with this tech. And then, immediately that should reshape how adoption looks like for these tools. I know it has for me. Have you abandoned Google yet? I know Google just came out with their version, but I tweeted out in February that I can't believe I am contemplating switching to Bing. I was contemplating it at the time. I haven't been to Google now in weeks.
Dallas Wells
The short version of it is, you search for something in Google and it's like, here's some places where you might find an answer. And the iteration of that now is, here's the answer. You don't have to wade through the links and the garbage stuff where somebody's gained the SEO, and the thing that was published 12 years ago that still somehow gets on the first page. That stuff's sorted through for you. Now again, there's work to be done on how it gets cited, and how it chooses the right thing, but that's all just incremental progress that, like we said earlier, is happening by the day. So, as it's being used now so broadly, the models just get better and better, and pretty incredible that ... And I think you see the fear from Google of they've sat on this magical algorithm that turned them into a trillion-dollar company, and they've tried a bunch of other things, but that's still the business. And now all of a sudden, that is potentially being disrupted. What an amazing thing to happen.
Alex Habet
And what I also discover in my experimentation with this, is that it's creating a new skill, at least for me. So now it's becoming, how can I prompt this thing to get what I'm looking for? So, it starts with you ask for something generic, you're going to get partially there. And I think that's what most people's experience is like now with these tools. But I discovered the power of modifying your prompt. So you could say, "Rewrite this with three paragraphs instead of two paragraphs, and add more detail about this topic." When you start to perfect how you prompt these things, the sky's the limit, dude. Seriously.
Dallas Wells
Yeah. So again, people fear the, "Oh my gosh, paralegals are going to be put out of work because chatGBT can do that job." But this prompt engineer concept, that's a real thing all of a sudden. It is now today. And there's like, imagine what it's going to be five years from now where somebody who's just really skilled at using these tools, and they can do the work of what used to be a hundred paralegals, or fill in the job title. That's exciting. And where this can take us is pretty amazing.
Alex Habet
So look, I know that you'll be keeping a close eye. I certainly will. And I look forward to working on some of these skunk work projects alongside with you and the team, as we enter a new era in the, I guess it's the information age now. I forget what the new term is, but ...
Dallas Wells
I lose track. I'm an old guy trying to keep up, so ...
Alex Habet
Well, Dallas, appreciate your time and coming on the show today. We talked about two consequential topics, which they're going to dominate this year. So, I expect you'll be back here to give an update on some of these things. But for the time being, The Purposeful Banker audience always appreciates an appearance on the show, and wish you and the team best of luck. Godspeed as you guys keep innovating in this space and serving our clients every day.
Dallas Wells
Yeah. Appreciate it, Alex. Thanks for having me.
Alex Habet
Alright. Well, that's it for this week's episode of The Purposeful Banker. If you want to catch more episodes, please subscribe to the show wherever you like to listen to the podcast, including Apple, Spotify, Stitcher, and iHeartRADIO. And if you have a minute to spare, let us know what you think in the comments. You can also head over to q2.com to learn more about the company behind the content. Until next time, this is Alex Habet, and you've been listening to The Purposeful Banker.
