navarre.aiClarisTalk AI · the FileMaker AI podcast
It's all about Personalization
Matt and Cris talk about the key factors that make AI great in FileMaker, and that personalization is the next big one. How much does your database know about you? Does it have memory of what reports you've run? Your job title and main responsibility? Your KPIs? Shoe size? 📌 Show Notes & References Key Clips & Talks Steve Jobs at WWDC 1997 – “Start with the customer experience” https://www.youtube
Apple PodcastsSpotifyYouTubeAll episodes
Transcript
Transcribed by machine (whisper, large model), 14,779 words. Names and product terms are sometimes misheard; the audio is the record.
Claris Talk AI. It's Claris Talk AI. Chris Hippolite, welcome, sir. Hey, Matt. Always good to see you. It's been too long. It has been. We've got to do this more frequently because it's always fun. Yeah, you're so busy, though. You're, Mr. All-Over-The-Place, doing all sorts of training and speaking and whatnot. I know. I've got a lot happening, which is great, right? What's coming up? Yeah. So, let's see. The FM Disc user group, it's online. Yes. Based out of Los Angeles. I think it's the largest one in the world, in the country, in a way like that. They always have a lot of people. Free. Join it on meetup.com. I'm going to be interviewed on there, September 12th, about semi-retirement.
Oh. That's my favorite thing about you. So, kind of like the... It'll be... Some of the stuff... Some of the psychological stuff, some of it's like the control of your day, talking about how amazing parts of it are, and also the parts that are challenging. And just kind of, you know, thinking about what we want to do. Some people are going to retire at 40, if they can. Some people are going to retire at their 70 or whatever. Oh, my God. Or come in and out of it, right? Because I've taken a couple of sabbaticals over the decades. Yeah, because, you know, fireman developers are rich, man. A million just... Yeah, apparently. I mean, what am I doing here? Dude, geez.
Yeah, I mean, some filemaker developers are commanding those $200 million salaries. Oh, no, wait. That's... Yeah, what is happening? I mean, so that's dominated the headlines, hasn't it? Yeah, for a while, for sure. But, yeah, so what do you... So, tell me, are you doing any of the conferences? I know this is sort of conference season in Europe, where you are based, so... I've got two. So, classes I'm teaching. So, I'm teaching a foundation class online, September 8th, for a foundation. Foundations class, and then an intermediate one, September 22nd. Each of those is a two-week, like a 10-hour interactive thing. And then expert one, an AI-related, mostly AI-focused, is October 20th to 29th.
A lot of interest in that so far? I mean, I hope so. I really hope there's a lot of thirst out there. Cool. Yeah, so these ones, you just go to my website, Navarra Training, or send me an email, mattatnavarra.training, to get information about pricing and whatnot. The foundations one's on sale. Ooh. So, let's see. So, those are big. And then there's two conferences coming up. Naples, FileMaker Week, that's just FileMakerWeek.com. It used to be Rome, now it's FileMaker Week. Yeah, it used to be Rome. This year, there's like a huge Catholic Church centennial thing, or something like that. So, there's way more tourists than usual this year, October, in Rome. So, they chose to do it in Naples instead.
I'm doing a session there. Maybe I'd love to nerd out with you about it a little bit. Basically, on ways to use JSON to improve... To improve AI results. Well, in general, it has been my experience that AI loves JSON. Our language models love JSON. It just loves structure. As a matter of fact, even in prompt template design, using tag structures to delineate different parts of your prompt template make it easily more consumable. And the same is true for data in JSON, or JSON format, whichever. You know, whatever you fancy. So, yeah, no, I think there's a lot of there there. I think, I know you've done some really good work with it, too. I've been impressed with some of the videos I've seen that you've done.
Yeah, I've been continuing to work on it. I think it's been great. So, there's going to be some cool free utilities for that one, and I'm excited. So, the Naples Conference, the Tours of Naples, they do an amazing job running a conference from Tours. And they have a really good theme. I love Evelyn, who runs it. So, the 6th and 7th in Naples are going to be like walk-ins. You know, walking tours of the city. Then the 8th is going to be an Italian language conference day. Because so many people, there's a really big FileMaker community in Italy. And then the 9th and 10th are going to be, I guess, all English or English/Italian. I'll wear my AI glasses there to understand what everybody's saying.
I'll be doing something like that, too. And then the other conferences engage you, and that's the largest European conference. There's going to be 300 people this year. That's in Antwerp, November 9th to 11th. That's on the bucket list. Yeah. And then I'm going to probably hang out in Amsterdam and go to some other places in Europe, in Northern Europe, for a week or so after. For that one, I'm doing a panel discussion, which is always awesome. Get a couple of people from Claris and grill them. I love doing discussions. Zero prep. My favorite. No, no, I actually do a lot of prep. Oh, sorry. They're much better if you can prep. But you can do them with zero prep.
Zero prep. But if you do some pre-interviews and have some ideas and come up with -- and have people sort of agree on what you're going to talk about but don't talk about it yet, I think you get a really good -- Like we do with the podcast. Yeah, yeah, yeah, kind of. But my goal with panel discussions is to get some disagreement and interest, right? So you kind of -- Like the podcast. Yeah. Except one of these days, Chris, I'm going to be right about something. You're always right about stuff. And I'm doing a three-hour -- Three-hour AI class there as well. So AI training, man. It's just -- Yeah, I'm glad you're doing it. I hear from people at Claris that you're, you know, leader in the clubhouse there.
I'm sure with the latest 2025 release, there's a lot of additional thirst for some of the new things that were released to kind of help people keep along on their journey. I forgot to mention, I'm doing free classes for Claris, too, you can find on the community page. I'm teaching some AI intro classes, like 45-minute ones. That's right. I remember when we were talking about that. And also, like, a fun one to do is the first time to FileMaker, like a class for people who haven't really seen FileMaker before and you see how much you can do and, like, build your first custom app in FileMaker, which is fun. And actually, as crazy as it sounds, I give people a little sample file.
You can do some AI in that, right? Like one of the new features that constructs a find request from a prompt. Okay. And you can say something like -- Yeah. Find all the companies who have -- Oh, yeah, you're talking about the 2025 release. Yeah, the 2025 release where you can -- it's actually really kind of a brilliant feature to be able to take a text prompt, look at the data on the current layout, and turn it into a regular old find request and run it. Love it. So it takes advantage of that. And it's just insanely easy to use. And what's cool about it is you can take the script that does it, call it from another layout. And it will just work, right? It's, like, it doesn't even care about context.
You can put whatever fields in the layout. And so the new calculations come along to say, hey, what's on this layout? And how do I interpret it? And then, you know, the AI will figure out what the fields are. So if you say something like, you know, company or contact date, it will look at something and go, oh, this is actually -- Semantically. Yeah, yeah, yeah. But if you say something like find companies in California, and it's actually -- It's not magic, but it's pretty good. Interesting. So, yeah, that's fun. Love doing that. I don't have any plugs, dude. Like, actually, the only thing noteworthy is just yesterday I found out I got -- well, so OpenAI has these dev days that's just, like, a big announcement of new features and stuff that they do every year.
And kind of a big deal. It's a very newsworthy event. And I guess I had -- I don't know if you've heard of it, but it's called OpenAI. And it's, you know, it's a -- I don't know if you've heard of it, but it's a -- I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it, but I don't know if you've heard of it.
It's literally there, it's like just attend in person and see the new stuff, you know, in person and to me, I think it's cool. I like to check out what's going on there at their facility and, you know, just kind of being in the San Francisco AI vibe and at the epicenter of what hopefully is some pretty newsworthy stuff. So I'll be there reporting live and gathering all sorts of info for the podcast. How's that? That's on October 6. While you're out at one of your conferences, I'll be sitting in the audience, soaking stuff up live and looking at Sam Altman. Tell us what is going to change the world next. Speaking of news, I wanted to, I wanted to share something with you.
We haven't even had a podcast since GPT-5 came out. Oh my God, you're right. If you watch their videos, it's called GPT-5. GPT-5. Yeah, I mean, I guess I don't feel as bad about, yeah, so the Dev Day should be GPT-5 related stuff. It'll be new stuff. Who knows what they're going to do. But yeah, a little underwhelming of a GPT-5 launch. Really? I use the model. I use it as sort of a model as a judge for other coding that, you know, different models. I know they were a little apologetic with the launch. It wasn't super spectacular. The AI community has met it with sort of a meh, you know, even it didn't like grab all the top benchmarks or anything like that.
So it was kind of interesting. And the combination of sort of the meh GPT-5 launch and then the release of this paper, this paper that I'm showing for those that are just listening, it was created by MIT. It was created by MIT Media Lab's NANDA initiative, which has got a big focus on AI. It's called the Gen AI Divide State of AI in Business in 2025. I consume these things all the time. I want to see, I love the studies. I'd love to see where things are going. And this one grabbed all the headlines. And it grabbed all the headlines because, well, you know, media loves doom when it comes to AI. So they found this one of the entire like 30 pages. They found like literally if you Google Gen AI Divide State of AI in Business in 2025, go ahead and do that, folks that are listening. 100% of the headlines that you're going to see under the Google News are going to say MIT report says that 95% of AI pilots fail.
And that is not actually what the actual premise is. It says that there's a certain type of AI pilot, meaning, you know, a company has decided they want to explore AI. And they put some money behind it and they do some experiments and they do some stuff. Well, do some of them go into production or not? And they said 95 of them that have a certain characteristic do not. Well, a pilot is supposed to fail. It's not supposed to. Well, I know that's a little. I mean, it's not. It's a yeah. Well, anyway, I mean, it's a pilot. It's it's like, hey, let's come up with, you know, 15 different things. Yeah, I mean, we will actually when we the first thing we do when we engage in AI customers, we say, what experiments?
Can we do like we, you know, AI is a bit of a black box. So how what can we do with your data and your users and your workflow to show you that AI could be useful? Let's just like take a tiny little thing, a couple of documents, a couple of pieces of data. Let's see if it works. I mean, there really isn't a fail there, right? It's just we're going to learn something. And we certainly we could learn that what we're trying to do is not even possible. But here's the really fascinating thing about what this study actually captured, even though no one read beyond 95 percent and then the word fail. The reality of it was, is it said that any the pilots that fail are the ones where they're just connecting like directly to AI.
Like there's there's just like asking questions and expect, you know, we could you and I have used this term, you know, kind of tongue in cheek, zero shotting. We used it in sort of a pejorative sense. And people in AI who are talking about who say zero shotting there, they're they are speaking pejoratively about just expecting you give AI like one command and it just like does the perfect output for you and then you just ship it. That is you. No one should be under the perception that that's what AI is for. You know, I really the the what they really pointed to was there was actually an emergence of this concept that's called context engineering. You and I have talked about prompt engineering.
Everybody's talked about prompt engineering. But earlier this year, this idea of context, context engineering. And when you look into context engineering, it's really about, you know, doing techniques like red. Where you're grabbing data, you know, chunked up documents or data from a data source or something. And then you're sending that along with the user's questions, so to speak. There's also other things like introducing memory into these techniques. Context engineering is really just a sophisticated, like smart way to actually do AI programming, whether it's a genetic or it's just LLM workflows, really focusing on the additional data and tools that you're making available to your integration are really what make them useful.
This particular study points to those techniques as having a considerably higher success rate. Well, not just considerably higher than 5%, but like very successful rate. And I urge people to actually throw this this paper into ChatGPT or something and talk to it or whatever. So one of the things that was also pointed to in there is personalization or AI applications that learn as they're being used. And so I thought that would be a really interesting topic for us to talk about here is personalization. I'm pretty obsessed with it in the I'm very obsessed with personalization in the context of AI. I'm happy to explain what that is and get your thoughts on it.
And I really think it actually is really important within the FileMaker community because and I'll try to frame that up. So first off, we're all in the FileMaker community, not just because we like software or creating business software, but it's a very special reason why we're in the FileMaker community. Everybody that's listening is because we. They believe in custom, like that's our religion right now. Now, save for a small handful of very successful verticals out there that don't do anything custom. They just do the one thing or one could argue that their first iteration was custom and they found a product market fit there and then decided to iterate on it.
From what I know about verticals, and I know a lot of those guys, they actually there is a lot of custom with it, right? So they'll give you like the core, but then either unlock it, some of them actually allow the whole thing to be custom. Customized and so they can't really upgrade them because they like here's the code that's the huge code base and we'll do take this whatever direction you want or the way I did it back in the day with pre one was here's the core software and it's like a customization module and you could do whatever the hell you want with that. And there's that's there's there's hooks in and out of the main system to go to it.
So if you want some other report, it'll hook out to the custom file where you can control your layouts and then if you click a button to go to a detail, it hooks back into the anyway. But yeah, yeah, yeah. So I agree with you. Custom is custom is the thing. Yeah, yeah. And I think in the way I always like, well, and that's why I became a FileMaker guy, I was a user, I was using like the act member act with the exclamation point at the end there, and it didn't have any customization. And so I moved to semantic and so I moved to FileMaker, which was all customization. And so for me, I always felt like, you know, custom versus SAS to me is just the lack of compromise.
It just means like, I actually want to get the thing that that's purpose built for me and my organization. My individual users. But that's what's kind of fascinating is the tool set that we've had up until this this AI moment has allowed us to build customized apps for companies or organizations, right? We couldn't really ever take it to the user. Now, granted, there are some techniques, you know, like customized landing pages or something like that. But at the end of the day, FileMaker applications are purely deterministic. They are someone decided you go there, you do this, you do that, you do. And there may be there's some conditions and some branching going on and maybe some level of personalization as far as that branching is concerned.
But the reality of it is it is a deterministic application. And we've in the most deterministic applications are like SAS systems, like verticals that are built for one reason. And every user, every subscriber gets essentially the same experience. And I don't and I and the thing that got me into this and I don't know what your thoughts are on this is that I believe that every organization should have their own custom software like that. That is the optimal. The optimal scenario is that a purpose built just for that organization, for its users, I think, is what the goal is. I would assume you agree because we're kind of in the same space. Well, yes or no.
I mean, it depends on the kind of business that you're running. If you're running a subway, then no, you shouldn't have that. Sure. But, you know, if you're running any kind of a chain business that sort of comes with a program, then then no. But for like so many other businesses that are that are unique in whatever way. Absolutely, a custom program is going to allow you to outcompete your there's a couple of things I want to first of all, lack of what was that phrase you use? It was so great. Lack of compromise, lack of compromise. Oh, yeah, I guess. Yeah. Off the shelf is a lot of compromise. And yeah, but I mean, lack of compromise is just such a great phrase.
I like it. Yeah, I dig that. I mean, just because I said it doesn't mean I dig it. But yeah, that that's it. I, you know, as a business owner as well, too. And you can probably agree with that. I'm not saying I don't want to do this. I want something that's for me, right? And not just because I'm selfish or whatever, but because I feel like I do things uniquely. Like, how many times have you heard your FileMaker customers say, well, you know, we're we definitely do things different here at XYZ Corp. And you kind of think in the back of your mind, like, that's literally what every single customer says. So, like, you know, of course you do. Right. And it's good that some organizations embrace that.
And I've always thought that, like, that's that's why we're all in FileMaker. But there's there's one more level to it. I imagine that customization is where you build apps. There's a lot of companies, maybe workflows, but personalization, this concept that I'm pretty obsessed with these days is where you're building individualized experiences for users, for each one of those users within the organization. So just like a user's job role is different. That's really what I'm talking about here today is especially customized for the user experiences. Let's take a step back and like to try to put that into something that that is maybe something that people can understand is just using the chat GPT.
Right. So if you use chat GPT and you're having a conversation and about something, you can tell it, remember this thing. And then it's building a little it's building personalization for you. Yes. And it knows that when you ask for something and knows that about you, you know. Yeah. And actually, so that that is that exists in chat GPT. You know, other ideas are doing the same. You know, Gronk does that claw. Now, everybody has the personalization feature. It's very high level. I mean, that's how different. Okay. What's your vision? Well, first, if we look at so so there's two there's two things. There's the personalization, meaning identifying the user and bringing that into the context of their interactions with language models.
And then there's personalized interfaces. So if we look at the the example that that you just gave with the chat GPT, that's more of the here's me. I'm Chris and I want you to know as much about me as possible so that you can give me more relevant results. Now, why is that? Why is that such a big deal? Well, I like to think of, you know, the other thing that everybody focuses on is hallucinations. Right. And and you've probably heard me talk about these in the past. And I say, like, if you're using, you know, AI systems as question answer machines for business, you're just doing it wrong. And what I really what I'm really saying that there's there's some nuance there.
So really what you should eat, I don't actually believe that language models are if they give you the wrong answer or the answer that you weren't looking for. I'm not saying that they just made something up. I'm saying. You get the wrong answer. I actually think that's on the user like, you know, that the Hitchhiker's Guide to the Galaxy where they're like, you know, they find the all knowing being and they go up to it and they go, OK, what's your question? You know what? And they go, oh, meaning of life and life, the universe and everything. Yeah. And the thing says the answer is forty two. And then they go, what? And then its response that no one seems to parrot as much as the forty two answer is it goes, well, you ask the wrong question.
Right. And I believe strongly that that's what I experience is with language. All smart. It is. The all knowing being and they get smarter and smarter and smarter and they have tons of intelligence. So if you're not getting the answer that you want, I'm not saying hallucination, but you're just not getting a relevant answer. It's because you're not telling it enough about who you are, what your goals are. It doesn't know enough about your data. It really doesn't own all that. So what the the answer that people think is the answer to that is to create like extremely like expansive like prompts. And I believe that if we're developing AI systems or integrated systems, that we should not.
Our users should not have to be prompt engineers. It just it's like anything else. They should just have to ask a question or express themselves. So the way that we actually do that, the way we put our finger on the scale and this has been my experience for years now is that the more you know about the user and you can build that in in the back in the background is like a user persona or like a statement about each individual user. And the more you know about the data that they're interacting with, you can you can describe that as schema or a system prompt or something like that for a data set. System prompts to MCP servers, for example. And that is a great way for the language model to be educated on, not just here's a bunch of data and a bunch of fields figured out on your own, but saying, here's a bunch of data.
Focus on these. And if a user ever asks a question about this, they're probably they're going to find the answer over here and so on and so on. Like that's actually development work that people that people like us can do, especially domain experts that are really close to the data and actually know the users and understand business rules within organizations. We can actually add a lot of value to this. So in the systems that we create, we define user personas. Here's everything you need to know about Matt. This is likely what he's going to be trying to do. It's your business persona, unlike ChachiBT, where it might know you and the name of your dog and the fact that you live in Greece and you like blue or whatever.
You're a nice plunge guy or not or whatever. It's strictly your business role. Like this is me when I'm working. Now, the interesting thing about the example with ChachiBT is it's updating like a little text, like bullet point stuff that it's observed. It's a little text that it's observed about you and there are and I believe the systems of the future, the systems that we're building now and what I think everyone eventually gravitate towards is having them listen to those exchanges that users are having in that hot path or in the in the comfort in the conversation that they're having with the LMS. And you might say something like, oh, yeah, you know, I mean, my role or, you know, I need to know this information because or whatever.
So the idea is that the system would like identify that. Using language models and store it away in a hard storage memory architecture that it then can then go find that information semantically and bring it into future conversations. So sort of the dream is imagine like you're working with an LLM integrated system and it says to you, well, I remember a conversation that we had where you told me about blah, blah, blah. So as a result of that, I can tell you that the answer to this question is such and such. Right. Or maybe it just sort of assumes that and builds an answer. So I mean, there's part of it that I can really see some utility and there's part of it that I just don't get right.
It's like a data entry screen that everybody in the organization shares has to have certain information in certain order. Not deterministic and it has to be the same. Data entry is 100 percent deterministic. Right. But your dashboard example gets me excited thinking about your description. Well, that's UI. So and no, no, no, it's no, follow me. So, well, I actually I have a soapbox I stand on a dashboard has to be three things meaningful, visual and actionable. Hmm. Mm hmm. And if it's if it's not all three, it's not. It's not a dashboard, right? So I would say the same thing for a layout. Oh, yeah, I give you that. Yeah, that's true. Let me let me watch your whistle now that you mentioned that.
Let's listen to one quick thing here. I think you're going to love this. This is perfectly topical. You're going to love it. Ready for this. Here is here is AI luminary Eric Schmidt. If you there's a lot of guys out there that when they talk, you should listen. Eric Schmidt is one of those guys you can you can rabbit hole on him. He used to be Google. Yeah. Google AI. DeepMind. He's basically a luminary that all the VCs listen to. Let's listen to him on a podcast. It's called Moonshots with one of one of the big AI investors. And this is something that really caught my attention because this is my position on this. So take a listen. I think user interfaces are largely going to go away because if you think about it, the agents speak English or other languages, you can talk to them.
You can say what you want. Do you want can be generated? So I can say generate me a set of buttons that allows me to solve this problem. And it's generated for you. Why do I have to be stuck? And what is called the WIMP interface? Windows icons, menus and pull down that was invented in Xerox Park 50 years ago. Why am I still stuck in that paradigm? I just want it to work. I think user interfaces are largely your thoughts. And by the way, we can dig into the WIMP interface, too. I love that. Yeah, yeah, yeah. I mean, there's some truth to it, but then. The find by example feature in FileMaker, that's 50 years old, not quite. Is awesome. The idea was great. It's still great.
It's it's easy for users to understand and use. It gives them a lot of power. And I don't know that we're quite there yet. And I don't know that I think you actually get results that would be bad if you had different users being able to create their own user interface for data entry. That doesn't conform to standards, which is kind of what he's sort of suggesting. No, he's he. Yeah, I mean, data entry is deterministic. There's certain data that needs to be collected. I don't even. That's the main job of almost every FileMaker system. Oh, no, I don't think so. I mean, your dashboard is not data entry at all. Well, yeah, but I mean, the main job of a FileMaker system is to is to structure data and allow entry of data in an organized way so that you can use it for whatever you want to do.
It's data in, data out. I would say that is a important part of its responsibility. But there are some users that don't enter data at all in our systems. And I don't think they're that that's that small of a community. I think that's the exception. Now, I don't I can't think of a customer like that offhand. But now I'm not talking about organizations, dude. I'm talking about individual users. So get out of the way of like organizations. I'm talking about I'm just talking about why an organization would use FileMaker. It's custom so that they can track what they need and they don't have to compromise. They can track what they need and not track what they don't.
That's a that's a that's a key point of, I don't know, virtually every FileMaker. You're talking about the value of customs. So yeah. So now when you get to the point of individual users and this is really where the personalization thing comes in, this is literally here's my vision. Here's my like dream that came up when you're describing it. You can tell me if this is kind of what you're thinking or not. OK, so different users, when they log on, if the system is personalized, then on their dashboard, I hate the phrase KPI, but it would basically when they log on, do an AI, you know, a smart query. And then they can say, OK, I'm going to use this. I'm going to use this.
I'm going to use this. I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this.
And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. And then they can say, OK, I'm going to use this. Yeah, I think so too. And always when we have conversations, my brain starts kind of going like, what would the steps be to get there? How's the rubber going to hit the road? For example, I've been thinking a lot about one aspect of that, which is just reports. And historically, for the whole of our careers, we've been building reports using list layouts with sub-summaries and find commands.
And then-- And then kind of going steps beyond that to build it with like a virtual list or something that does multiple of those and then build a PDF from data from different contexts. But it's all basically kind of the same. Reports are probably the most deterministic thing we build. Yes, I mean, yes. Deterministic in terms of what it is. But of course, the data changes all the time. Correct. By the way, I always think of deterministic versus probabilistic. But I'm not sure if that's really the other end of that particular spectrum you're talking about. Yeah, it is. Yes. It definitely fits here. I think when we're talking about UI, it would be deterministic versus adaptive, I guess.
Yeah, that's fine. Makes sense. But it's the same principle, you know? Yeah. So what I've been-- Well, how many times has a customer said to you-- and FileMaker's not the most flexible reporting tool, right? Like, you have to pick your columns. And sure, I get it. There's some tricks. And there's certainly virtual lists and ways that you can mess around with sub-summary reports and sorting to reveal. And hide certain layout objects. Great. But I know everybody listening, and even yourself, has had the experience where you build a report for a user and they go, yeah, I get that the rows can change and the data can change based on foundSet. But can my columns change too?
Like, can I just have this thing not be there if I want to do this or that and the other thing? And we reluctantly say, well, not really. Like, I can build you a different report that has that and maybe take it to you conditionally. So I feel like it solves for that problem. What we're talking about here is it just presents the data in the format that makes the most sense for what the user is trying to do at that moment is kind of where my head's at there. And also, I think reporting might be the first. If anything goes away, it's probably reporting when users have the ability to just query data and have conversations with it instead. Same data, same user, same goals, but why not just ask?
Well, and I think that... most of the time the report exists because they want to get the same or similar data on a recurring basis. Sure, but I'm saying you can still achieve that knowledge by asking and getting responses. Yes. So actually, kind of the things that I've been building is rather than doing all of your... my work as a FileMaker developer, building list layouts and doing all that, instead, I create prompts, system prompts and user prompts, store the prompt in a table, in a record, and then when the user wants to run a particular report, they call up that record, which already has the known working thing, you know? And so the report says, you know, you are a report that looks up invoice data and summarizes it by this.
When you format it, you always use U.S. dollars and you use two decimal places. And that's the programming you're doing. And you get all that information, right? So all the programming I'm doing is that rather than... That's where I'm going to. So if I have it like working and I know it's right and I know it doesn't hallucinate, it gets the right data, then I can say, okay, well, I'm going to give that as a system prompt. Then the user can say, oh, well, how do you want to tweak that? You know, what are the parameters you want? And the user can say, you know, compare Q1, you know, for the last four years on the top whatever products, right? That's kind of a weird sales example.
Anyway, you can get... That's where we can go. And as long as the reports are reliable and you can give it enough data there's definitely issues around that right now, even with all the new features. And FileMaker you're talking about? Yeah, yeah, yeah. Like having FileMaker analyze a gigantic amount of data with AI is a problem. It is a challenge anyway. I know no one from Claris is listening, but yeah, I don't think FileMaker is quite there yet for just like whatever data sets. Like it's got some limitations and within those limitations it's useful. You're right. It's not greenfield opportunity stuff. I mean, so like, you know, the ones that we talked about, I talked about earlier where you can do a type of prompt and it will do a regular FileMaker search, that can work on millions of records.
It can also create a SQL query that can also work on millions of records. But if you want AI to use its AI engine and have it intelligently connect the information and understand the meaning of it and really get to the nuance of it, like for example, you know, which of the 10 products that have come out in the last six months are most likely to have the highest sales in customers in this space or something like that. It has to have really deep understanding of the data, which means you need to give it, you know, you need to give the AI engine 10,000 invoices. And you can't do that in an API call. You know, not only, not the execute or the, you know, insert from your own, also not in the new ones, it has a limit of tokens.
And I've done some play with it. And you can send, I did a huge one where I sent like, I don't know, 500 customers with a huge amount of data underneath it. And it took like five minutes to come back with the response. Yeah. Not optimal. Not optimal. And it also was a limit. I couldn't do everything right. You know, you could do things like RAG. You can do, you know, other types of, and those are the things that I'm trying to resolve now to try to make as easy as I can for, for customers to use. And I'm still working on it. Look at this nice chart you've got here. Yeah. So this chart that for those that are just listening, what I just pulled up was, well, there was a, there was a lot of chatter not too long ago about a concept called context engineering.
I think I mentioned it earlier. And, and, it, it followed prompt engineering. And I thought, oh, wow, there's a new technique. Let me go look it up. And I, you know, click on all the different links and I see, oh, what they mean by context engineering is just the right way to integrate with language models. And you just define this. So context engineering is actually sort of an amalgamation of a different, a bunch of tech, different techniques where appropriate. Like you said, RAG, right? That's one of them. Prompt engineering, which you also said in here, that's one of them. Structured outputs. So, you know, you've defined a use case with structured outputs in your, in your report example.
And then the only thing that like, maybe wasn't so obvious that we weren't talking about is you want to introduce statefulness, right? So like having history that's available, that is the thing that can help you learn about users and, and, you know, analyzing these logs to, to be able to, to understand the users and their, and their goals better. And then just simply memory to actually facilitate that, right? So there, this big movement of context engineering, to me, I'm just delighted to learn about it and, and just go, oh, that's just how I've been doing it this whole time. And, and if anyone's just getting started and you want to kind of accelerate your, your, your journey, dive right into context, just, you know, search on ChatGPT, use study mode, by the way, which is that cool new stochastic learning thing and have it guide you through context engineering and let that be the basis for how you learn AI, because it's really what we're talking about here.
And that will then facilitate a more, you to be able to create a more personalized and I, you know, if I replace the word personalized with relevant experiences, that's really what the goal is. It's not just personalization for the sake of personalization. It's like, I want this experience to be as relevant for me as possible. And I, and we have the tools now to actually make that happen. So it's, it's a bit of a future leaning thing, certainly, but that is where I expect us to be sooner than later. And so how do we prepare for that is as folks in the file maker industry, lean in on your domain knowledge, lean in on what you know about data, lean in on being able to express how one user is different from another or how this data sources is should be relevant to those users or relevant to language model.
Make those your skills, because arguably your customers probably can't do that as well as definitely can't do it as well as you can. And be great at that. Master that part of it. And then that way, when we move into the, even if we move into the, you know, a completely personalized, completely personalized or ephemeral UI era. Yeah. Someone has to drive the coding or the programming is that this contextual, you know, knowledge that you're actually building in. And I think we're all in a great, great spot for that. Yeah. I mean, it's, I've always thought, and not only in this domain, but in a lot of other places that when we get better tools, it allows us to focus on the higher value parts of our expertise.
Right. So as file maker developers, we've worked so hard to get this domain knowledge. We've had, we have all these skills that are now not as useful because we can just, we can call them up. Right. So if we want to write a, you know, a Jason function, I did one really recently to, to calculate the distance from two points from latitude and longitude to calculate the distance in kilometers. Uh, because I've got one of the databases I'm working on is really cool. It has a bunch of different points and it wants to, it wants to get relevant data from another table. That's the closest to this particular point to see what else is going on around the area. Well, I didn't want to write that calculation.
I certainly could have done it, but cause it involves, you know, a sign cosine curvature of the earth, blah, blah, blah. It's not, it's basically, it's giving you the information as the crow flies. It's not a super simple function. Um, it involves, you're throwing four pieces of data at it, right? Latin long source, Latin long destination, whatever. Um, uh, where was I going with all that? So basically great coding example. It's a, it's a, it's a coding example, but basically it allows me to, uh, kind of know what I want and then not have to do parts of it. Right. I have a calculator, so I don't have to multiply 375 by 1931 in my head or on paper. Like I, you know, that used to have to do a long time ago.
Now I have that part is now automated and I don't have to do it anymore. I still could, but I'd use a better tool, which allows me to spend my time on higher level pursuits. Yes. And so our job as developers for what we do is that plus our job as developers is to create systems so that our clients can be using more of their high level pursuits. Uh, and because they also have this domain knowledge. This domain knowledge and expertise in, yeah, we know you're, you're, you're nailing it. I, I think, um, you know, it's kind of the perspective I can lend there is that, uh, I've got two separate businesses. One, I solutions, which, um, is a file maker, development consulting, uh, company, uh, FBA.
And we just build file maker applications. Uh, a hundred percent of the people that contact us are already using file maker. Um, they already know what they need. And so we really have to have two skills. A lot of people just jumped to the second skill, which is develop the thing and deploy it for them. But really the first skill is to be able to capture really what it is that they need and be able to translate that into what needs to be built. That's like a special talent. Totally. And, and so that doesn't go away. Uh, so if you look at what happens in the programming world, uh, where it's already happening, let's face it is where you just say, this is what I need.
Um, you know, the Claude code experience, as opposed to the cursor experiences, like, here's what I'm trying to achieve. Here's the requirement specification, build the application based on this. You have to be good at that. You have to be really good at that part. And that a lot of people call orchestration. I like that terminology. So when you're not doing the actual coding, you're orchestrating the creation of code. And that's where domain knowledge in that context becomes so critically important. So for, for those that have refined that scale over the decades, um, then, you know, frankly, you're really good at that first thing. And arguably maybe just as good, if not better at that than you are at the second thing, which is the programming stuff.
So if programming goes away, it seems like the floor falls out and it's like, oh my God, what am I going to do now? But then what I'm saying is that the first thing becomes even more important. And that's something that we're all good at. And so that's my hopeful note. Exactly. And I don't know when AI is going to take that over. I think that's, you know, I, I think that's, I'm not exactly sure how that's going to work, right? Because the customer, like if a customer asks for something, how are they going to know, uh, what, you know, what the best way is to do it? I don't know. I'm still, I'm really having a hard time wrapping my head around that. The gap is actually in that the gap is not the tech.
The tech is a hundred percent there today. The, the, these models are absolutely capable of taking direction and, and creating a production worthy deployment, or the deployment worthy bug free version of the, that experience. That's part one. So then, um, the real question is how do the customers even know that there's AI that can do that? How do they go find these tools? Do the tools exist? Are they prevalent within the marketplace? Do people know to go interact with those? Probably not. We're, we're probably at least a couple of years away from that. So I think there's really two things to be checking, you know, the heartbeat of, and one of them is, are our AI models capable?
A hundred percent? Yes. We're there already. There's literally nothing to work, wait for. If you've got your head in the sand on this stuff, just here to report it's there already. But, and then some, like, I mean, even my best developers, like the guys who are just like total or gals who are just rockstar coders, they don't even write their own code anymore. They actually, the smartest ones don't even write their own, their own code. And just recently like GPT, which ultimately ended up being GPT five, one, a like international coding competition, which is like, oh, this is like the best humans. It's sort of like the big blue moment. So it's there. But the question is, do all our customers know it's there?
Do all our customers believe it? Do all our customers who read the headlines about a very promising, you know, document that talks about capabilities of AI, are they just seeing the headline that says 95% of AI projects? All right, well, AI is not there yet. No, it's there just like, how do we connect those dots? And so therein lies an opportunity, Matt, for us to be the Sherpa, to actually, you know, be the stewards of this experience and actually say, yes, well, let me take you there. Just, it's just the same thing as any other consulting is, is that like any other kind of consulting, like my wife, um, well, before my most recent endeavors, like she, people would be like, well, what are your husband doing?
She's like, I don't know something with computers, but he just can't fix my iPhone or my printer. But otherwise I don't know what he does. And then I started just saying consulting. And then she's like, she's got this like rap, like, well, what is consulting? And I defined it as, well, I don't know who hired me. Like basically that's, you know, or more, right. But that's the idea. So now the 10% more could be how, where those tools are, what the capabilities are, what we know is possible. And then, then, then what we learn is their domain knowledge. And then now we're a value to the tools as well. So I don't know. That's, that's, and I'm trying, this is a hopeful message. I believe this is a hopeful message for the folks that listen to us, talk about this stuff.
I'm there with you too. I'm definitely like, uh, you know, uh, having a positive view of it, mainly because I'm choosing to use it for positive things in my life. I'm choosing it. I'm choosing to use it to, you know, to help me, uh, for things, things like ice cream. I've been making ice cream. We got a really sweet ice cream maker. Um, uh, and you know, and coming, it's, it's feels sort of weird to, you know, have a long conversation with chat GPT and end up with a really good, uh, saffron, honey ice cream, which is good. Um, um, um, well, the tools are there. There, the tools are there. And then of course, you know, learning, coding, doing API integration, whatever new thing is coming at me, failing at better ways to teach stuff, organizing, you know, help with marketing, blah, blah, blah.
Well, here's a good, a great example. I'll reveal. I, I talked about this on the last podcast that I did this presentation on AI and education. It was the first time I've ever done this, a bunch of like sort of disparate thoughts and references. So for the first time ever, I used AI to create this presentation. What I did is I went into Gemini 2.5 because I had a million token window because, um, uh, I talk a lot about, you remember the brave new words book, that Sal Khan thing. I literally uploaded the entire book to this conversation. And then I talk a lot about guide on the side versus Sage on the stage. And there's these two papers that talk about that. I uploaded the entire papers, not just snippets.
And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 1984.
And then I talk about this, uh, Benjamin Bloom study from 1984. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54.
And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh, Benjamin Bloom study from 54. And then I talk about this, uh,