Podcast Episode 1: Daydreams of Multiplayer Agents and (Maybe) the End of Duplicate Records | Ep 223

Description

In this episode of Rock Cast, the team talks about the next unlock for AI Agents, especially after a long 3-month period of limited progress in the industry around harnesses. Instead of working with personal agents in a private chat, imagine collaborating with your team by sharing a live agent workspace. This is the transition from single-player agents to multiplayer agents, and it will dramatically change the way your team works.Then, they pull back the curtain on a problem every Rock church experiences: duplicate records. They invite listeners to brainstorm early product thinking and daydream about a unique solution: requiring verification for phone numbers and emails when adding a record. Jon and Emily are careful not to promise any guarantees, but allow listeners a peek into the crucible process, that is, the refinement and testing of all ideas.Visit the show notes to download all the resources talked about in this episode. Don't forget to join the Rock Cast Rocket Chat Channel to see what other churches are saying about this episode.

Transcribed Content

Welcome back to Rockcast. On today's episode, we're gonna talk about the next unlock for AIUs, and that's moving from single player agents to multiplayer agents. And then we're gonna dive into a segment on something that probably all churches, all Rock churches experience, and that's how to eliminate duplicate records. So, John, do you wanna kick us off with some AI updates? Sure. Yeah. Things are moving a lot in AI as always. Obviously, the two big things I think if you're paying attention to the podcast is AI safety. But that's just in my mind, I don't really wanna talk about too much about that because I feel it's turning into a very political thing. And I feel there's not a lot of facts flying around, but there's just a lot of emotions. And I just don't think that's super helpful. I would say though, with it everything in our culture, always try to infer , what is the intent behind someone's communication methods? Are they trying to be helpful or are they trying to control? And I think when it comes to AI, control is going to be the big thing. AI is to dominate where our culture and society and economies go for the next ten years. So it's obvious that people are going to want to influence it in different ways. But on the happier side, there's some really cool new other things happening. Yesterday, which depending on when you're releasing this podcast, a new model came out that's very, very exciting called JEV. We don't often get new models, but what's exciting about this is it's a whole new classification of types of models. So, it's not your traditional LLM, which we're very familiar with. It's basically what they're calling a judgment model. And so, it can make decisions. And what you do is you give it questions and it responds back not with text, responds back to with answer. So there's a few ways you can do this. You can use it to, , perhaps do classification of emails. Is this person happy or sad? And you just give it the body of the text and you just tell it what you want the answers, the possible answers to be. It can do sentiment analysis. But they're calling it a judgment model because it helps basically judge a question. You can kind of think about it as , if you're writing an if statement, inside of Excel or code, you've always had to be very precise in there. If this value is greater than two, then do this. But where it says value equal to two, imagine just putting in your thoughts in there, and then it will decide if it's true or false. Or it can classify it. You can say, is it A, B, C, D? And so, it's going be a major unlock because it's 20 to 200 times faster than an LLM. Wow. Yeah. Woah. And 40 to 400 times cheaper. So this thing is lightning fast. It can process these questions in literally a blink of an eye. So when it comes to judging fraud, it's going be huge. So it's not going to replace LLMs, but it's going to replace some of the jobs that LLMs do today. One kind of nerdy example would be when we use RockIQ and you give it that question and it's going to go call the skills and tools within Rock, well, first thing the orchestrator, that logic that's going to figure out what to do, has to do is look at all the tool calls they could make and say, Which tool should I call to go do what you want? That would be a perfect job for this new JEV. And it's going be way faster. Right now, it has to send the whole body to an LLM and saying, Hey, what do you think we should do? And it's going to use its LLM capabilities to return back a whole bunch of text that will be formatted, but that's really hard for the LLM. JEV, that's a perfect job for it. And it's super fast, super cheap. It's going be a huge unlock. Again, won't replace LLMs, but we'll replace some of the things LLMs do. I said, it just came out. I just got access to it last night and so I was playing with it. It's really cool. It's not something that the average person is going to go to use JEV. It's something that developers and AI platform creators will use to help make the tools that we build better. So it's not you're going to go to AskJev or Chat with Jev. You don't chat with it. You send it structured questions and it sends you back structured answers. But super exciting. And again, it's because we're seeing a whole new type of model come out. And I think the future will probably be LLMs will still be a very big thing for us, but there'll be these new types of models that are being innovative. That's really exciting. I think there's a lot of potential for the Rock development, Rock platform, sounds . For sure. Classification of prayer requests in terms of sediment or category or the big one I always worry about is trying to find self harm requests so we can be responding to them immediately. And I know some people I've seen out there in the general community worrying about AI in the use case of prayer. And I think that we do need to be careful, but the classification of that is pretty low risk. It's pretty harmless. And it's not that we're saying we're judging a prayer and what response kind it's going to get. We're just trying to do stuff that helps us keep an aggregate view of what is on people's minds. But also, we will prioritize a request that has self harm. Feel that That should be happening. And the faster we can make that determination, the better. Right. And I think these tools, what they do is allow us to pray more. back in the day when you I lived in the time where at the church, we would type all the prayer requests in that came off cards. And then that would go to a very select few people on the prayer team. Whereas now we can get immediate prayer requests because we can look at it, make sure that there's no scandalous content or even if it's unintentional. Some people don't think about the fact that they just outed somebody's divorce or Personal situation. Yeah. And then so AI can, the goal would be that we pray faster and more often because we can take care of sensitive things that unfortunately need to be done. Wow. Very exciting. A lot of potential right there. Yeah. So, next topic is moving into more of a conceptual thought. And that's around if you think about everything we're doing with agents today, it's amazing, it's great, but it's what the industry is calling single player mode. So if you think back to the '80s and you went to the arcade, most games had single player and some had multiplayer. So we're really much in the single player. We play with our agent, we create our agent, our agent does things, which is good. I think there are some things in our work life that match that need. But I get very frustrated because I feel the harness has to become multiplayer so that we can all work with it. And that could be in the setup and configuration, but it could also be in the interaction. There's ways we can share these skills, but that's still not multiplayer. That's me giving you my game and you play single player now too. You hand it to me, now it's in my environment and I can do the same thing you're doing. Correct. And I think most of the work that we do in our organizations is truly multiplayer. Even if it's something that, Oh, that person checks the main inbox for emails. Yeah, they do until they go on vacation. Or until something else happens. Or I want to put in some knowledge into that workflow and I can't. I have to email them and then they can put it into their bot. So I play with a lot of these frameworks, GrokBot, Buzz, all of these things. And they're all neat. I them. But I always get frustrated because I'm , cool, but how do I help somebody else with this? So the industry is really talking about the need for multiplayer agents. The good news is they're talking about it. The bad news is there really isn't anything that does this, which is shocking to me a bit. I think there's stuff coming. Okay. You're starting to see a few people talking in some interesting ways that make me think that they get it. And so hopefully in the next few weeks, I mean, it's the greatness of this area, things move fast. We'll have some stuff. But I think the reason why we're talking about today is you need to be thinking through that. There's a really great mean, it's legitimately called a manifesto for the need for multiplayer agents, and I can get that for the show notes. And this guy wrote it. And as I'm reading, I'm , Yes, these are the things I'm still frustrated with, too. And the part I love the most is at the end, he's , And there's nothing that does this. I'm , Okay, good. I'm not just missing it. It's not just me. I thought for a minute I was clueless and there's all these solutions and there really aren't. But he wrote it because he was frustrated. And it's a really well done manifesto of it. And again, I think soon we'll start to see some things come out. I saw another person posting about the same frustration, and I thought he had an interesting take is that here we are probably nine to ten months into agentic AI, and we've come so far. But yet here we are so frustrated that we don't have this thing because in the last three months, the harnesses really haven't improved that much. And we're all getting anxiety over this and frustration over a lack of three months of innovation. I think that really shows the pace of what's out there and how fast it's coming, that the fact that many people are getting frustrated because there hasn't been innovation in three months. So that was kind of , Oh, yeah. In the Internet generation, which I kind of grew up in, the creation of the Internet, it was years between new versions and Major innovations. Yeah. And then there's a dry spell there for five years where HTML didn't really change much at all. So, it's just a wild world we live in. And it's kind of fun. There's definitely a lot happening and for our expectations to be moving so quickly along with it just kind of goes to show that this is very exciting. It's totally new. Yeah. I think right now, because the tools don't exist, we should be thinking about what we would do with these tools and how we would organize with these tools because I think there is stuff we can do today. Yeah. So, I mean, let's apply that to the Church context. Where do you think would be a first or natural way to apply this inside a Church staff? So I think, first, an evaluation of all the tools that you use. You have your Office product. You have, obviously, Rock. You might have an HR system. You might have I mean, just probably a project management tool, I'm guessing most churches use. I mean, I think Rock is really great for project management of connection requests, but I probably wouldn't use it for sermon series planning. There is a cool project management plugin for Rock that a lot of people use and . But I would say that's almost it's bringing different new capabilities to Rock, which is good. But it's kind of an alternative to an Asana, a Monday, which again is good. The cool thing is it has access to all your people data. So if any of those things have a need for that people data, that's awesome. But go through all your tools. And then I would be evaluating two things. I think every tool needs to have MCP access, so you can talk with the tool with AI. But I also think it has to have webhook capabilities. Okay, that's interesting. So, a lot of times, I would say MCPs are reactive. , I need to get something, I will go get something. But they're not proactive. , an email just came in. I will take action on this. Instead, an MCP will , once an hour, I'll wake up and go check if you see if there's a new email, which in a lot of use cases is fine. But a lot of times I want things to be a little bit more quick. So, I think making sure your systems and you understand your systems, can they do webhooks? Can they do MCP? So, if you take Asana, for example, it can do webhooks. It's not easy to set up, but it can do it. And so, in Rock, you can do webhooks. You kind of have to create them yourself a little bit with automation engine. That's something we probably will be looking at soon to make easier. So I think these multiplayer harnesses are probably not you probably don't Rock Rock to be your multi agent harness or multiplayer harness. I mean, we could change our mind on that. But I think it's to me, it reminds me of , you don't need an LLM from Spark Development Network. that's a utility. You don't need an email service from Spark Development Network. So, we think Rock should be the center of your people data. Anything that touches a person, that should be in Rock. But there's a lot of things outside that. So, think as you think about the harnesses, be thinking about your tools, how would you wire them? So for instance, I've been thinking maybe the front desk email box. Cool. So now I can have AI go through that, potentially start classifying, writing responses that a human could look at and approve or modify. But then what happens is, Okay, but where's all the knowledge that needs to go into that? Where's all the tribal knowledge? Where are you going to put that? How are you going to get that? So I've even had, using the MCP, I've had it go into our inbox, look at every email, classify it as , what is the type of email is this? And then our responses, there's the tribal knowledge. So, we start sucking our responses back, our human generated responses back to maybe an issue with a plugin. Oh, well, that's in the documentation. Or we updated the documentation to make it more clear. Well, it can track all that so they can start answering those for us. And again, I probably for now would not have it auto answer, but can instead of someone sitting in the inbox looking and responding, maybe they go into their Asana and says, Hey, there's an email. They're basically asking about problem from here. Here's my planned response. Do you that? Yes. And then it sends. So trying to think through, trying to brainstorm those things. I think, I guess, the tool doesn't exist, but don't wait for the tool to exist. It's coming. Enough people have a pain point and innovation one will catch 100%. The problem will be we'll have too many of them soon, and we'll have to wait and see which one shakes out. If you had a guess, how long do you think it'll take? That vacation? By end of year. Think by the end of will be multiple that will be battling over because we all have our own preference. And I think we're starting to see some very interesting companies wake up and I think they've been in stealth mode a little bit. And so I imagine there'll be some interesting announcements soon by some of the products you the R community uses today. Not Spark not Spark though. Not trying to hint at something. So is the term multiplayer been around for a while? Or have I just been living around? It's been around for a long time in AI world, which is about two months. Got it. All of our timeframes have just been really compressed. It's a legacy term. It's been around for more than a week. Again, it was a big wake up call reading that post and saying, we're all frustrated by this. I'm , yeah, because it hasn't moved in three months. I'm , Oh, right. I've learned how impatient I've become. To me, my biggest impatience, frustration to the level 10 is that you couldn't send Teams messages from an MCP. I finally got so frustrated, I found a solution and implemented it myself, only to have the next week, Claude, Croc, all now can do it. I'm Is that the way it goes? Yeah, but I wait three months. I'm , Okay, I'm done with this. There's got to be a way. I find a really cool way to implement it, which meant standing up Docker containers and everything. And then the next week it's Moot point. Yeah. So, I guess we just need to be patient. But it reminds me of the likes when Rock features come out, we always tell you, start playing with it on pre alpha, start reading the documentation. Don't wait until the upgrade to actually get started. I would say the same thing with here. Be thinking about multiplayer harnesses and what that could do with your organization, with your Rock data, with your email communication systems, your project management? , how would you dream for that? So you're prepared when they actually do arrive. Yeah. I'm really looking forward to all of that. Yeah. Next unlock. Yeah. That's right. For our listeners, I wanna know that I wanna hear if this is also a pinpoint that you've been experiencing and if you're looking forward to it. Do you have anything else to add before we move on to our next topic? I don't think so. Okay. So, Emily, can you lead us into, the problem that all churches experience and that's with duplicate person records? Yes. Duplicate records happen. Everyone is frustrated by them, but they happen, and they happen for a good reason in a lot of cases. So before we get into this next segment, I wanna just make a a statement here that we should keep all in mind. A lot of times our listeners ask us for, hey. Can we get a peek behind the scenes on things that the team is thinking about from the product perspective, but hasn't yet been , it's not being coded. It's not a feature that's coming immediately. And a lot of times, are things we can't talk about. They're just not formed enough to be able to discuss. But we're gonna bring an idea today that has a little bit more formation and, is something we're thinking about a lot. We're gonna talk about it here, but we're not promising that we have something that's on the , that's coming in a near future. It's kind of on the road map. It's kind of a brainstorm, but it's not, , ready to go. Yeah. So it could change. I would classify this as a daydream. This probably won't even happen. But it's getting thought. But we think about it. I think people will be surprised how much thinking goes on behind the scenes in meetings that we have hour long meetings on what ifs. Yes. But I'm gun shy too, because I mentioned a forward thinking idea once on the stage at RX in Chicago. And I still hear about it. I'm , one day this could be what we could do. And I was , when's that coming out? It's , well, that was a I think part of the maybe part of the way of thinking about it is most of those ideas or brainstorms or daydreams don't come out in a way that people hear. There are so many of them under the surface that any one that you do hear about is just as likely not to happen as the next because there are so many. Yeah. But we are gonna talk about one today. And just as long as everyone realizes, we're not promising anything, but there's a lot of thought that's going into this for some good reason. So John, do you want to talk a little bit about what that looks ? Yeah. This is not coming out. This is just us in a We're just chatting. Just talking about current issues. Thinking about church data. Yep. And then for our churches, is this something that we want their input on or if we think it's a viable solution? How should our listeners be taking this? I think I would just say, Hey, join us behind the curtain on chatting on things that probably won't happen, but it's interesting to think about. And here's what I would say when we talk about these things, we're usually talking we could do this, and then we don't do that. But then later, it informs a different way with a different pattern. So Yes. I would just say, hey, you guys said you wanted to hear more about what goes on behind the scenes. We're letting we're gonna try this. We're gonna try talking about it. Yes. And the other thing to keep in mind is we've shared here on the podcast that a lot of times the ideas that we fact, our last podcast, go through even if they get the forward momentum to become something in the future, they go through a refining process of a lot of different people's input and a lot of important considerations, and they come out looking very different on the other side. So even if this had momentum at some point, it probably wouldn't look exactly the way that we're maybe going to talk about it today. Call it the crucible. We put it through crucible the impurities kind of float to the top, we skim those off, do that a few times. But I just know, I just don't want to regret Saying these things. Saying these things and stirring pots and I already have enough things to worry about. So no holding feet to fire. This is just a little listening session from behind the curtain. Yeah, okay. As long as everybody's on the same page. Yes. Okay, that's enough lead in. Let's do it. Okay, well, I'm still nervous, but we'll do that. So, a common frustration, you said, Emily, is duplicate records. Why do we have duplicates? Well, we're trying to do some complex things in Rock. And I think most systems will fall in one of these two camps, but we fall in both. We're trying to make things really easy for people to do where they don't have to be logged in. We just want to take their information and trust this information. So think about a lot of event registrations. You don't have to be logged in. We're just taking unstructured names, basically, and then trying to treat them as real people and matching it to data that we already have. Or if you're trying to sign up for a new Here class, we don't want them to go through an interrogation process. So we're trying to be very friendly. At the same time, we have some reasonably sensitive data in the database. I would call it sensitive, but at the same time, we're not a bank. We don't hold medical records. But there's some sensitive things in there that we have to protect also. in the levels of military secrets, we're not top secret, but we're secret. Although I can hear all the security people saying, Yeah, but I mean, literally, we don't have we're not storing credit cards in our database. Or social security numbers. Hopefully. If you have social security numbers in your database, please expunge them immediately. Thanks. So holding those two tensions that we want to be kind and friendly, and at the same time we have these sensitive information, usually systems with sensitive information do not become kind and friendly. That's true. So, we're kind of skirting both of those and trying to do our best. And those things are in tension with each other. So, one day, and this is really days, we've spent weeks talking about this. We were saying, Well, what would it be to have no duplicates? zero duplicates. I hear the echo of database administrators everywhere clapping. Yeah, it's a bit of a dream. But you have to start with something from a first principles approach. If we want first principles, no duplicates, what would we do? If we were starting over the fresh plate, what are we doing? Well, I don't think we could ever get to peer zero, but we could drastically reduce this. And we feel the right way to do this and the way other systems would do this and other systems are doing this, not church management, but secular systems would do this, was to have a few rules. One is only allow a mobile phone number to be on a single record. You cannot have it on two records. This is pretty common, actually, in a lot of secular systems. 100%. Almost every secular system would not allow you to put a phone, a mobile phone number on two records. And it makes sense because most people have their own mobile phone. I don't know too many people who are literally sharing the device. Hey, can I make a call now? I'm sure there's an edge case out there, but those are edge of an edge. So, also, only allow an email address to be on a single record. That way we know that they're there. And on both of these, we would say, whenever we get that information or we need to tie to that information, we need to immediately have them prove access to either the email address or the phone number. Verification. Which we're very used to now. That's a common thing now that you have to prove access to that. Because of the email, you could say, Well, lot of Christian couples will share an email. I mean, that's an edge case, but it's not an edge of an edge. We might say, Well, we could have a feature that could say, Hey, email me through my wife or my husband. So it's still on one record. But hey, if you need to get ahold of me, go ahead and use her email address. That's interesting. There actually are couples where even they have separate email addresses, but they might prefer that option because one keeps a calendar and one has no idea what they're doing tomorrow. Those people often marry each other. Have no idea what we're doing this weekend. I'm actually happy. So that's one thing. But I think the big thing is every time you need to link a record or you need to know if that's a person, you have to test it. We would probably say that this might create two classifications of records in verified records and unverified records. Unverified records, we might still allow in certain use cases, yeah, you can put that information in, but it's going into kind of think of the cesspool. we're not going to trust those. Don't even call them people records. They probably are people records, but don't consider them people records. They're in the cesspool until they can be verified. Makes sense. And don't care if they're duplicates in there. It's just , that's a garbage pit. But we're going to try to get anybody who's in a small group, anybody who's giving, anybody who's significantly active in the church should be a verified record. That's our goal. And for the most part, we're going to try not to create the garbage records, but there are a few use cases where maybe we might need to, but we're going to try to create these are verified records. Again, I'm saying this as if we're doing this, but we're in the daydream world. Good point. We're still in the daydream. We need that Wayne's world. We're in a parallel universe here. And any unauthenticated person would have to verify again one of those two identifiers. That makes sense. And it's something that we're all used to. Yeah. I mean, now, I'm sure there's lot of people saying, Yeah, but we want to be the friendly people too. Well, here's what I would say. I think, yes, that might have been the old world of thinking. But in the new world of thinking of security, I think we're actually doing them a disservice. if I go to the church for the first time, and I'm trying to register my kids, now we need to make this user experience super polished and very easy. So not today's user experience. There would So have to be some I picture if I'm going to a church for the first time, I want to check my kids in, the registration system might be , hopefully they're entering their own information. I think it's faster and I actually feel better about that. But I'm going to have to verify my account right then. And I would probably proactively say to Hey, we really care about your kids. We really care about your kids. We really care about your data. Therefore, we have these protections in place. I don't think that's going to seem unnatural. And someone might say, Well, I don't know if I trust you, the Church, to give you all this information. Well, you've given me your kid. I want that to be The most secure? Right. Yeah. So, now if our kids are now out of the house, if I'm going to a church as a couple with no kids, now I have a little bit more control to say, When do I trust the Church? That's true. And I can be unknown. But I think it's really important then in Church communications that we say, Hey, if you remain unknown, we can't serve you. We can't be there for you. So, the expectations of what we want to do for you drop because we literally, you want to be a ghost, we can't serve you. Right. It's very difficult to personalize things for someone if we don't know them. Yeah. So, are some of the things I think would actually improve security. That's interesting. Because duplicates, we create them because we don't know what to do. We have two records, they're kind of the same. We don't know which one to trust. We probably know what to do. We don't know what to trust. So we put it in the duplicate bin. And what happens? A human who has no more context than we do decides what That's to true. Mean, in some rare cases, I'm , Oh, I know Ted. And Oh, yeah, he just did that. I know he signed up for camp. But that's really rare, especially in the churches that are using Rock. So what we're doing is we're just pushing that clueless merge from us to a human to do, which I think is good because we don't take the risk. But honestly, the people who are doing the merges probably have no more context than an AI system or us to do it. So I sometimes feel we don't do it for security reasons, but it's no more secure when the human does it. Yeah, that's a great point. That's a great point. So if I think through current church data and structure, and this alternate reality came to be, there would be a little bit of cleanup work related to accounts because there I believe there are quite a few accounts that have the same email address and in some cases the same mobile phone number. And people put their information on their kids' accounts too. Yeah, which again is a really bad idea. To put your mobile phone on a kid. Now, a lot of people , Well, we know it's not there. Because there's certain things we probably shouldn't be doing. We should be texting kids under a certain age. But they're , Woah, I know it's not the kid, it's the adult. Well, you don't know that for sure. You assume that. But again, we could put features in that says, Communicate to this person through that person. So, when I need to text, when I need to get out to all the kids, I can say, Hey, now you can do this today, right? You can say parents of and then text the parents that feature exists. That does exist. But we could even extend that to say, Hey, if you need to get ahold of Ted Decker, email Cindy. And you can still send the email with his name on it and personalize to him. I think when Cindy gets it, she's gonna go, okay, I'll let Ted know. That makes sense. Guess I'll tell Ted to go get take me off his send to me. That's probably more likely. That's for them to figure that out. Yeah. We have couples that are paying for. That's great. So I guess then the big question is, is that additional friction that would come into place at event registration, at child check-in, and in some of those areas worth the additional security. Right? That's the crux of it. Yeah. Think it does scary aspect is probably the biggest thing. I think it does let us administer to people better because when you have duplicates, it's difficult. I think there's a lot of churches out there who are doing a good job on their duplicates, and they have volunteer teams who are doing that really well. I know there's churches who pay people to do it, and I'm always an advocate for that. Hey, that's a great volunteer task. If you think the difference between trust is that you pay somebody and now they're more trustworthy, I mean, I think, , there's trainings you can do and documents that they can sign, and I don't think paying someone makes them more trustworthy. I think that's, , good stewardship would be to get a volunteer team, train them and get them to sign confidentiality agreements. So, I mean, I think it helps the system. My fear is that if we go to roll it out, obviously there's somebody who's not going to want it. So now we have to play in both worlds, which is incredibly difficult. Think that's become the answer to everything in Rock, we'll just make a setting. Yeah. It's we got more settings Too many probably. Than a 1970s aircraft. And you go to the cockpit and there's switches everywhere. So I think that's a cop out in many cases, make a setting. That said, we still have people who want us to be able to have a person have multiple email addresses, one person have multiple email addresses. That's true. We do hear that sometimes. And we did that in a former system. Literally, that former system allowed you to have as many email addresses as you wanted. And that causes chaos. Because then people are , Well, if it's between this time and this time, email me at work because maybe my band's schedule got changed and don't email me at home because I'm not going to get it till after the band was scheduled. It's just there's no way. Or email me about these topics at this address and these topics about the other. I mean, literally, that's the feature request we would get. And when we went to do it again, we're , the one thing we're not gonna do? What you're not gonna do is give them two email addresses. Oh my goodness. If you think back to the conference when we had the comedian. That's right. Some people will catch that. Were , you're not gonna do is give them multiple email addresses. So that was the first thing we were , we're never doing that again. No. A comms director is just gonna say, well, email all their email addresses. So pretty soon, every inbox they have is gonna be getting the exact same message every time. And what a pain to have to go delete it in other places. Eventually, you just stop reading things. Yeah, but that's literally today still a feature request. Yeah. But again, would go back to what secular system allows you to have multiple addresses. I'm sure there's a couple. I don't know what they are. But I'm sure they all regret that fact too. They're living in the first world you just mentioned. The regrettable one. Yeah, once we had it in there, mean, there's no way we could have taken that feature away. Which I fear a little bit is this. If we wanted to change to this brainstorm alternative universe, the rollout would be very difficult. I think that would be a challenge. But if we did, I mean, one of the features would be text every person and make it a verified account. And the text message would probably say something , In an effort to increase our security and because we care about your data, we want to make your account a verified account. And that means I actually think as a church attendee, I'd be , cool. Thanks for caring about my confidential personal information. Yeah. And then we've talked about on the podcast, how security in the AI age is going to change drastically. And I just want to tell everybody, we don't talk a lot about it because it's security. It has already. The number of security issues that are turned into us is exponentially higher. Does that mean that Rock has exponentially more issues? No. It's just that there are more people looking for those and they have way more powerful tools to help with that. Before, you had to read a ton of source code and you had to nitpick and now you can type in some pretty broad attack vectors and say, Look for these types of things. And it's a thousand times easier to do that. Now we're responding to those quickly and doing that. And also we're getting some that are not as bad as reported. And those are not common, but there are some that's , Hey, this is a big deal. It's , Well, actually, it's really not. And in fact, if we did that, I think most churches would be upset that we took that stance. And this is another good opportunity to remind people, keep your Rock instance up to date. Absolutely. Because we are being very responsive to security issues if they're reported, and those are included in our releases that we put out. But if you're slow to adopt them, then all of our responsiveness on that is to what end. Yeah. And a reminder, we can only keep Rock completely secured to major versions. That's right. So I would say if you're not on 19 right now, you're going to be in trouble because 20 is going to be coming. So right now, two that we support are nineteen and eighteen, but 18 is about to get off that list. So I would be moving to 19. Yeah. If you have a cadence that says we only update once a year, you're going to not be under a secured version during portions of your year. Yeah. That's important. Yeah. So, but I think this if we did do this, it would drastically help duplicates, security, I think make the system the way it should be. That's a really interesting topic to think about. And thanks for sharing it here on the podcast. But now let's make the closing reminder, right? this is something that's just kind of in in the thought space. Yeah. So I'd say, , go into the podcast rocket chat and, know, share opinions. But I think sometimes at this stage, loosely held are good. I think it'd be good or I could see that or I think a challenge might be, instead of sometimes what some community members might be is , that can never be Very definitive statements. If you want to brainstorm, that's a shutdown. Sometimes we go back to some of the rules of improv when we do this. And so there's a rule of improv, which is , you can never say no. You can say, Yes, and, but you can't say, No. Would again, these things are not on the roadmap. Now, would I to get them on the roadmap? Maybe. But I've tested it on a couple of people. And I was actually surprised at how receptive they were to it. I think there was one church who was , I don't think we could do that. In fact, could we get multiple email addresses? literally, I was , no. Oh, boy, put that in reverse. Again, because I think we've kind of in some cases want to put our round peg through the square hole. And we can't, It just doesn't fit. You can't have both. So pick your poison. That's it. And every feature that exists today came from some idea at some time and went through that crucible and came out differently, and other ideas were left on the cutting room floor. So this is just a peek back way into that original soupy process. Yeah, and some of ways we get to this is I mentioned, we went from our first system to the next system. We said we would never do that again. But we've had some decent levels of brainstorming to think , well, what are the other things? If we were to start over, what would we do differently? And this was one of the top ones. Fascinating. Well, we're definitely interested in hearing people's opinions in the podcast channel. But, , I was reading a quote just this morning that was talking about how opinions are the most commonly shared form of human knowledge, but not always the most valuable unless they're applied correctly. So, , this is the chance to have loosely held opinions that can help idea shaping and it's a fun thing to be a part of. But we don't often share things this far behind the curtain simply because there's not enough shape to it yet to make promises. I think it goes back to just experience. In the past when we've done this, we tend to think that made our life harder, not easier. And so we're trying to be even more transparent. Although I'd say there's not another system that I know is as transparent as we are, that has a community chat, that has conferences where you can share all kinds of ideas, where you can literally talk directly with us. We want that. That's why we did this. But that's also been in , if we ever did this again, much transparency would we where would we set that transparency mark? Because it makes our lives harder because we really care to keep people happy and keep But things I do sometimes daydream what it feels to be a sauna and be , I don't really care. , I think that's what most vendors are , I don't really care. We'll listen maybe outside when you don't think we're listening and we'll put that into the product guidance. And as an Asana user, I don't feel I need to be heard or should be heard. It's when they don't have a feature, I'm , Well, I guess Asana doesn't have that. Well, do have some forums where people suggest ideas and things, but the Asana team doesn't usually interact with them much or at all. They do, I think, take some ideas from there because you can see some changes in their platform from time to time. But it is a totally different dynamic than what we have. You basically have someone in community relations team saying, I will take that and make it And it's not going anywhere. Churches shouldn't miss the heart behind this, that we care for them and we want to give the best possible solutions. And then leave us your thoughts in the ROTCast channel. Hold them loosely. And welcome to the crucible process. You've officially been looped in. Now that since you mentioned conference, we do have some fun announcements. The first one is that the Hyatt Place is halfway full. Wow. Which is very crazy. And that's our second hotel or third? That's our last hotel. It's Number 3. Woah. Hurry It's your room. So if you've been waiting, now's the time to get it before rooms completely sell out. And then our last announcement is that our vision day attendance has doubled since last year, which is so fun. So there'll be even more ministry leaders and decision makers and pastors at this event. So if you haven't yet, bring it to your leadership. Invite them. Get them signed up. Book your hotels, and we'll see you in the fall. Thanks for tuning in to this episode.