navarre.ai

ClarisTalk AI · the FileMaker AI podcast

AI Consciousness, Clio, and the Future of FileMaker Training

August 13, 2026 · 73 min · Matt Navarre and Cris Ippolite

Matt Navarre and Cris Ippolite return for a wide-ranging ClarisTalk AI conversation about how quickly the AI landscape is changing — from philosophical questions about AI consciousness to very practical FileMaker applications. The episode starts with a discussion of a Google AI consciousness paper exploring what happens when language models are steered toward or away from claims of consciousness. Matt and Cris use that as a springboard into theory of mind, empathy, free will, AGI, and the di...

Apple PodcastsSpotifyYouTubeAll episodes

Transcript

Transcribed by machine (whisper, large model), 14,654 words. Names and product terms are sometimes misheard; the audio is the record.

it's Claris Talk AI and that long last time here with Chris Ippolite do you remember me we did it oh I do remember you and I think about you all the time Matt well actually truth be told we do still collaborate and talk about AI all the time we have all these little secret balls where we get together and discuss AI stuff with all the luminaries in our space hello to all of them because I'm sure you know they're listening to us here today we have so much to talk about this is what happened I think maybe our move is like we like instead of trying to keep up with AI we just let like three months go by and then we we can just kind of skim off just skim off the top um first I gotta tell you yeah how much things have changed well it's it's crazy so I mean I don't even know how far back we go maybe that's good uh but first before we get started I just had a really fun morning I just wanted to share with you it maybe doesn't have anything to do with AI or Claris but it's fun um I'm you know I'm in Southern California and um this morning there was something really interesting I jumped in the ice plunge and I went on a really early walk before the sun came up so this time of the year sun comes up about 5:45 that type of thing but over on the eastern horizon I was only able to see two of these there was a lineup of Jupiter Mercury Mars Saturn Neptune and Uranus

uh Matt uh which was visible right before right before you can't pronounce it correctly if you want to I'm sorry I see London I see France uh but anyways that was super cool I only saw a couple of them but it was just such a fun way to you know we were able to kind of get a little bit more of a fun way to you know have a nice little ice plunge vibe going on and start off your day so I came right from that endeavor to come talk to you so so the Perseid meteor shower is going on too yeah like oh I don't know I don't know about that oh wow I don't know where it's going to be visible from I think maybe northern hemisphere so I'm going to try to go out like one o'clock in the morning and see that all I see on my notes don't know nothing nothing bright in a timeline uh Uranus needs binoculars so uh I thought I'd just ask your doctor if you'd like to uh you're in a yes I'm sorry um so hey let's jump into something really cool before we uh I'm gonna I'm gonna go like super esoteric like trippy and then promise hey don't stop listening everybody we're gonna come right back down to earth we have a conference to do too that you're at you know yeah oh we're definitely gonna do a conference summary um and what I think so google AI the google consciousness paper yeah dude so this is a this is a callback I've been up to yeah first of all this is a call back to uh in a podcast episode that we did with uh 2025 guest of the year I mean you abstain from

voting but I voted him in the guest of the year Ronnie Rios uh now with full city consulting uh we did a I think well first of all it's the most underappreciated podcast we ever did just from a viewership standpoint but people should check it out I think it was just is AI conscious if I'm not mistaken or some kind of clever name I think that was actually and even though I'm not as much of a Star Trek guy as you guys are so I don't want to like scare anybody away even without being heavily into Star Trek I thought it was really fascinating uh we gathered a bunch of clips from um a trial of data was it meaning of man or some other name of the episode or something but we had our episode names but yeah he used a lot of uh a lot of references from data and from other like you know beings that were encountered that were um and we were using it as a kind of a fun sort of pop-up and we were using it as a kind of a fun sort of pop-up and we were using it as a way to explore AI consciousness I have my own feelings about AI consciousness um as a matter of fact I've been pretty obsessed with that uh if you guys go to uh isolutionsai.com and go to the video section the top video that's on there uh is one that I did when GPT-4 just had come out and I was doing a bunch of experiments on theory of mind you're you're you understand theory of mind you're the resident psych I'm sorry philosopher do you want to tell people a

little bit more about what theory of mind is and then I'll tee up what this study is that we're showing here I think I'll just give you a super simple definition of consciousness consciousness is that thing that we define that we have and it's a moving Target we can say something else has it or doesn't have it you can say a dog or a turtle or you know uh whatever anyway well that's the theory of mind aspect right so there's like two elements of it it's it's right there are elements but that's basically at the end of the day because no one agrees about what it is uh that's the thing like AGI or SGI but also no one uh well maybe some no one basically is positing that humans don't have Consciousness except uh uh you know well there is a lot of uh discussion and debate about whether or not AI has Consciousness I think we can all safely say that the early reports of sentient AI went in the like early Google era with the guy that yeah the guy who got fired I'm pretty clear he was working with a model that was not conscious do we have conscious models now well that's what this study actually did so the paper that we're showing here on screen for those that are listening um it's uh some researchers at Google's paradigm of intelligence team cool as hell uh team name

by the way and they worked with some other folks you know there's already always collaboration you see on the top of there there's a bunch of folks from like University of Chicago University of Maryland and Northwestern a bunch of stuff well they published the paper that we're showing here Harvard which is called it really kind of is it really is kind of equal parts philosophy but the name of the paper is introduced I'm sorry inducing language models to assert their own Consciousness restores human beliefs and values so what this team determined is that um they did a safety fine-tuning um to stop AIs from claiming that they were conscious right so the whole point was they did some fine-tuning on a model and said hey look you're not conscious right and it had this really massive like a fascinating unintended side effect so it didn't it did not just suppress the own model self-attribution of their mind it actually suppresses the model's attribution of mind to anything else like you had mentioned earlier animals or nature or chatbots or even God for that matter yeah yeah and humans so that so this is actually where we get into this concept of theory of mind so theory of mind is so there's Consciousness for self and then there's being able to project onto others realizing that others are are conscious you know beans as well too or are under or putting yourself into the mind of others and being able you know empathy comes from this

right yeah yeah that's that that concept it's like a child development thing you know most adults imagining somebody else's feelings basically it's something that yeah yeah like being able to say oh that person must be scared right now right like that is a very complex like you know uh concept from my standpoint they actually did the op they said hey you're not conscious and then what they found out is it affected the theory of mind that the model had projecting against all these other things you know including God so now the way that they were once they discovered that they actually tuned it opposite so what they did that is then they removed the safety refusal Direction that they added to the model and the self-attributed uh theory of mind scores they had these scoring if you read the paper it jumped from 2.17 to 4.77 on a scale of one to ten so that basically said hey uh you're allowing the model to think it's conscious and so all of a sudden it begins to think that everything else is conscious as well too and you went even further and actively steered it towards saying you are conscious and then it hit a score of seven on that zero to ten scale of consciousness and what's even more amazing than that is that the model um that was made conscious didn't just say that it had a mind it became more likely to believe in God so there's your philosophical question Matt as our resident uh philosopher what are your thoughts

there yeah are we doomed yeah yeah theology theologist yeah well I think I think it's interesting because you know a lot of people talk about AGI like there's a lot there's a whole school of thought that says that AGI won't be achieved through language models alone that world models and understanding the physical world around us has to happen first I I I've theorized that having a mind and being able to like speculate about the world around you and project thoughts onto it is probably step one uh I don't know your thoughts there well I yeah that's interesting and I still really kind of go back to free will right if free will is an illusion which it is um then what does any of this mean okay folks now we lost everybody uh for all the folks that are like I haven't had my coffee yet this morning let's break let's bring it back down a little bit fun things to think about if we just put all that aside right and then just talk about it in regular human terms because because it feels like we have free will so let's just say we do it feels like we have consciousness so let's just say we do and then let's kind of compare it I'll tell you my little thing and then we'll go on okay consciousness I would say like if you break it down you basically have neurons in

your brain there's a fixed number of them it's high there's a method that they work on they use you know the synapses and it's electrical and it's chemical um the time it takes for a thought and the number of neurons is a physical thing it's actually really slow um so they've got all this measurements right in a computer you have the same things going on they're they're using different it's all electrical it's not chemical um but you can you know if you wrote a program to simulate neurons and to do the exact same thing that a neuron would do in terms of like a swarm of stuff going on and uh you know memory in one place and division you know all that then why wouldn't it be conscious yeah I mean I think you know and we talked about this um in our and actually I would I if we remember to put into the show notes the link to that fun little Star Trek one that we did I think we should uh the thing that I had said in the context of that conversation was really I don't disagree with anything you said I don't my my I think I think that um uh that uh language is too small of a pipe for us to explore this uh so there's only so many words we could throw at it that's a that's a Kurzweil uh quote that I think

fits really well here but what I will say is as I dove in to understand how brains achieved a human brain to achieve consciousness it was through evolution like it took an entire evolution of the brain we didn't just have Consciousness all of a sudden it was it was created to solve problems and in awareness and and survival and stuff like that so and even there like is a at what point does a embryo fetus newborn two-year-old five-year-old um if someone is in a coma sleeping yeah you know but but you know what so to maybe put a bow on this what you're doing right there is describing Theory of Mind too like because you're being like there's a gradation too right it's not a yeah like so like like an insect some people will be happy stepping on an ant but they would never hurt a bird or something like that right size of the being right whatever our cuteness I think people have like a cuteness scale and so like you're projecting on to these other things and so then like we all we have to play some real sophisticated mental gymnastics to uh to say someone in a coma is still a conscious human being as if they were up walking around but it it takes us you know we have to re-re-reassert there it's a continuum which I which I guess means that what the Dalai Lama would be you know a tan

perhaps now how this applies to our models I I do think that the debate will go on forever because there's no definition that's just like you said but then part two is that the more that the models can understand and describe empathy and understanding to things around them the the more useful they'll be to us as you know in our agency moving forward so that's that's kind of like how we tie this whole thing back now now we've officially lost everybody yeah maybe but I mean if you can if you can simulate Consciousness to the point where it's indistinguishable from Consciousness then yeah but that's the key the indistinguishable means that we have to have our own definition and be anchored in something to be able to do an AB test you know like you can't dip unless you have a ground a ground truth so um I the whole thing's fascinating I mean literally I could talk to you about this yeah and then there's also life right so like the you know I remember way back in biology like the definition of life is you know it has to work to keep itself alive it has to um you know be able to reproduce um I can't remember there's a few pretty basic things that basically apply to all life um well I get I mean the question is is AI alive but I think we're clear that that that's not the case but uh but dude it's getting closer than it ever has honestly um we haven't even talked since

like fit well yeah we probably talked let's get some useful things yeah we probably talked when the mythos scare happened I think that was maybe the last time we had a discussion I'm mixing it up between these like little cabals that we have with the folks from the community and in our actual recorded podcasts but um while mythos isn't quite here we have had Fable for some time we had a brief respite where we did not have Fable available to us now it's fully available it's part of uh Claude prop is it part of the 20 a month I don't even know I don't know um I will tell you this I have the 200 a month plan and I used 100 of my Fable yesterday yeah yeah you sent me a slack on that um so my first thoughts uh there's a couple things so Fable for me I mean the for me my my personal world and I think and I know I feel comfortable the AI world changed dramatically um in you know January we talked about the 4.7 just sort of like complete end-to-end coding uh um updates that came out in two different models at the same time it changed everything I think the Fable model the Fable 5 and I'm not gonna uh discount um the 5.6 soul right it's also extremely impressive I mean

I'm arguably more impressive in some areas but but just my my observation is this Fable 5 for the first and I'm trying to try not to be like coy or too self-deprecating here but for the first time ever when you get like you do a bunch of coding and you get back a summary right and it's like telling you what it did and it like it's the first time where it was like dude I don't know what you're talking about like can you dumb this down a little bit more for me right like it was just like like I don't know if that actually is more of a statement about how it was it was tuned a little bit too much to not consider the recipients yeah or if it actually is so intelligent um it actually to me because you can really just set that you know I say very very frequently when I work I said set verbosity you can really just set that you know I say very very frequently when I work I said set verbosity to 40 you know yeah it wasn't even verbosity it was the it was the the technical level I I have the most ultimate pro tip actually before I forget let me just share this with you and folks so um so pro tip uh use ASD STE 100 that stands for simplified technical English uh instead of like a you know like if you're doing personalization or you're putting this into your prompt instead of saying like stop doing or talk to me like a 10th grader or whatever all I found is if you go

down here this is where it is um write all technical Docs read me this helps for your PRS too in the comments in your in your PRS uh but I found that simply saying write all technical Docs read me's and code comments in ASD STE 100 which is simplified technical English active voice sentences under 20 words one idea per sentence so on and so on I found that that actually massively like it's not dumbing it down by any means but but this whole STE 100 it was meant to take very like high level uh uh science and uh terms and actually speak to and quote unquote plain English I found that's been a very helpful prompt in my uh Fable 5 coding mostly when I you know when it's done and I and I like I do this thing maybe we've talked about this where I go at the end of this run whether it's a loop or a goal was achieved or whatever or just like just you know one uh series of commits I say give me a human readable brochure style HTML report uh so I can see everything that it did and this is where I'm like you know what percent complete are we did we solve for these things show me all the back tests and all that and this is where I applied that I applied in my report that's just for Chris right and it's been a game changer and I've shared this with a bunch of people and they've given me the same feedback that they've

experienced that as well too but now you I mean you had you had some real oh my gods um in our slack exchange on Fable 5. talk a little bit about that uh oh yeah yeah yeah so one of my projects this is a good story um is the decipherment of linear a I'm not sure how much we've talked about that it's an ancient language first discovered here in Crete um you might say it's the oldest written language in all of Europe um so languages that came before it were basically considered hieroglyphics and they you didn't really they weren't really sentences and grammar and stuff like that okay linear a was basically considered that right because there's the structure of it that they can see but there's not that many signs uh and there's not that many documents document in this case it's just a clay tablet basically um so of course it's not likely to really be deciphered because of the amount of that but what I built was um a repository that had every published paper that I could find every document all the other websites that had all the details about it a cool company or a cool place in Italy actually did a bunch of super high resolution 3D scans they published the data and just all of that everything and then when fable 5 came out I said hey there's a bunch of papers I didn't get they're

on the list go find them uh it doubled my count of papers to over 500. so it found a way it didn't get through paywalls or anything but it found a way to get to whatever kind of sites in whatever kind of academic places that had the document that was available that had some other wall or whatever to get to it and found it and then my little process to do what I do is it downloads it as a PDF uh translates it chunks it embeds it and cross references it like other stuff so basically you can ask a have a conversation with 500 papers you know oh dude I was blown away by the fact by by what it did and then when it described exactly what it how it did it I'm like that's insane it really really sounded a little uh yeah something's going on there uh the the the way that the on the the day zero when the model came out I used to have like little parts of the story by the way oh yeah go ahead please yeah yeah yeah so lead up to the part you slacked me too actually um well I can't remember the part I slacked you but there's something new that just happened so a guy oh um claimed to have deciphered Linear A using AI using

the same tools I'm using so I read the story contacted him uh he's gonna be hasn't published as a detailed results yet but uh I sent him you know he seemed really nice I saw someone he responded and I said here's the questions I have I wasn't able to resolve how did you resolve them and I had my AI agent write up a super big document to ask his AI agent so that his agent can then respond so basically the two of them can converse and communication in the future I I don't know yet but I think I've actually done much more research than him on it uh and I think I've just proved his his theory but I'm gonna hold out actually says yeah his conjecture yeah I don't think uh based on just a little information that he provided I'm not gonna get into the details but like a particular uh anyway um well I could talk about it but yeah I'm really curious well okay so this is your personal first person version story of some of the stuff that we're seeing happen all the time not not the least of which is um you know some of the breaches that are happening but yeah um how about like I mean you're kind of talking about some of the conjectures that have been solved

like uh and also one just came out a couple days ago opening I's been testing Astra this is going to be their Fable that's going to come out pretty soon and they stopped they solved like all basically uh 10 long open math problems that have been unsolved since 1999 similar to what you're talking about here and they did so in like 24 hours and for less than two thousand dollars retail that wasn't their internal cost and I mean this is the same kind of stuff we keep hearing about we keep hearing about you know that type of stuff or the you know the the protein fold or the you know all these math conjectures yeah all this stuff it's just falling math is just falling and and and and I heard I heard some smarter people than me mention you know math like a foundational math is probably dead and evolved and there's probably some mathematicians that are very sad about that right now but applied math and like even applied materials and applied physics like the the it's just going to grow exponentially from here on out because now we have all these baseline of understanding to apply to these things so it's crazy and how how do we bring that down to our own how do we bring that level of intelligence down to our own workspace um for me honestly the way the the first thing I do when a

model comes out is I have some day zero things queued up I'll go back to old coding sessions and I'll reinitiate them with the new you know I'd like to I'd like to have them stopped at nice clean places yeah if in fact I ever stopped and on the fable one I had a bunch of uh um code bases I I would say the the most sophisticated uh code bases that I've worked on they were equal parts machine learning and equal parts just you know code and web technology and I would I said on the day the fable came out um hey go give me your assessment of these look for vulnerabilities uh stability you know items whatever and I have done that I I did that same prompt this is how I can tell the difference personally I did the same prompts on Opus 4 7 and 4 8 right prior to that that that I mean I've done back testings in like the thousands of back tests on these code bases and now all of a sudden the first time I use the new model it comes up with a ton of stuff that I've never seen before right yeah and and then I started pointing in that like true problems that I was having and I was just trying to muscle through these and it was just like whipping these things out of the way and solving for stuff and it was like I mean we've talked about the moment you know in in January when all these things came out and it became

clear we can go end to end into production with stuff this was a different like Epiphany like sort of psychological reaction to it that I had personally was like oh my God like this is they're like oh my God moments with the AI are shocking and frequent in both our lives it's crazy let's talk about I'm gonna talk about the conference you just went to yeah this is good I think this actually fits with what we're talking about because um inspiring people to be hopeful about AI is I think an important responsibility that folks who've had the epiphanies like you and I are talking about yeah then we're going to dovetail to you and file maker training after that yeah yeah yeah first of all a conference summary yeah so um uh you know we don't have engage or devcon in our Claris Community space this year in the U.S yeah um in the U.S we engage you here which is huge and getting bigger yeah well that's I think the interesting thing is that the community-based uh conferences are are really emerging and um and I I'm speculating only I don't have any other information here but I speculate that that's probably what we see over here in the states um and for for I know that we had there was one in Canada uh just over the border earlier this year and then the I think the only one am I and forgive me for

if I'm missing some but I guess that's kind of on you if you have a conference that no one's heard of but uh the one we're talking about is one that happened in California right near uh Claris uh it was uh John Howell's uh full access this was the second year he's done it um uh John Howell did a great job with this he really kind of creates a style you know what it was it was very much like um I haven't really been to the pause on errors um that's probably the one I'm thinking that's conspicuously absent uh but uh pause on air is more of like uh it's like a retreat yeah yeah yeah you go out somewhere and you stay somewhere that's what this was so happens um if that's your jam you know this was right up your alley um and it was like a four day long thing I was only there for a day a day and a half I had some uh speaking commitments that I collaborated on with John but um but it was right in the backyard of Claris so the Claris folks were there um John was able to get um uh Ryan McCann CEO of Claris uh uh uh um who else was on there oh Juliano Icabelli the VP of product um uh let's say Clay was there as well too Evan who's the VP who just became officially

the VP of engineering who I'd love uh they were all there so he had a panel of like the the the greatest minds and the the most influential people in our community uh there as a panel and and did like an hour-long Q a with all those folks on on day one so that was super interesting um and then um there was an AI panel um the second day that was I was I was just a replacement um I know I was not part of this panel originally but someone dropped out so you know they went into the JV squad and pulled me up there but it was uh myself it was done by um Michael Wallace and uh and his bot uh Nova uh was actually one of the panelists which was funny uh then myself uh Ronnie Rios now with full city by the way um Rick Hellman uh and Todd Geist of course too so and then Todd did some other um uh like uh proof kit uh type presentations later in the afternoon so a couple of days I was um Claris you know insight and um some AI discussions so um uh and I think there's recordings or something I'm not entirely sure how that works but um but it was good I think I think this is what we'll see Community like more you know smaller regional uh conferences

in the United States much like out in in Europe and whatnot so by the way you need to come to mama for the engage you conference and beyond panel I'm moderating so I would love to I mean honestly I would absolutely love to every year I kind of put it on the schedule and I think about going and then like I don't know I just never get on the plane I know I should just do that one year you know and and Malmo would just be incredible um just absolutely beautiful so yeah the conference was good I don't know if you caught wind of any of of it or anything but um um you know it's nice to get together with the folks and it's hard to summarize right it's like you know here's a five minute summary of each of the 40 sessions or whatever you can really yeah and I and admittedly I only went to the you know I was only there for the first day and a half so I missed you know a lot of the meat on the bones but it wasn't like an AI conference I mean we're talking about AI and the things I just mentioned were AI but it was you know mostly you know FileMaker um presentations throughout so you know OG FileMaker content for those that are interested and maybe some you know leaning forward uh stuff um for those that are interested in the AI stuff it was it seemed well attended I don't think it was as big as the the Europe ones but maybe after a couple of seasons of some yeah John's built to be successful

with that he's just going to keep going for it yeah he's just so committed and into it um yeah he's all in for uh the next one too so as are Yotis and Johan who who run uh engage you uh it takes a special breed uh no question about it uh yeah and a team well I I will say that um during the panel uh the AI panel someone had asked a question about um oh I think one of the questions I'm sorry that that the moderator Michael uh presented was um something about how to learn FileMaker right like where to go to to learn FileMaker and um I'll get my answer in a second but but first I want to give you an opportunity to plug what I think is one of our interesting um I mean would you call it a class or I mean I know it's open enrollments the the the uh the time that you have uh was it every two weeks go ahead and explain what it is because I think more people should be involved I try to do it every week so it's a free AI FileMaker integration class uh I've done 20 of them now I think um I tried to do it every Thursday um but I you know when I was moving I took some time off hey we live in our new house now that's right

congratulations by the way uh most incredible view uh probably looks similar to what you've got in your background there I'm guessing that's the sea we have in the mountain view um wow so I I spent a ridiculous amount of time prepping for each class um and it's so much fun to do it it's so much fun we had 52 people in the last class it just keeps growing wow um so yeah and there's a class tomorrow so just kind of different Technologies tomorrow I'm going to be showing you this thing that I called Clio CLIO it's FileMaker logs on the AI age basically I'll give you a short description please so uh it runs there's a website I'm in love with called fly.io and you talked about that quite a bit on the yeah in the classes yeah I'm sure I do that did uh it's a because it's just such a perfect kind of combination that connects with FileMaker so well um the virtual instance of Linux or whatever and runs a Docker um no uh friction at all to set one up and I did a bunch of work to connect it to FileMaker MCP and other ways basically oh data um so to get that connection running well really well then when you deploy it it runs on their private cloud or you can run a Docker on your own computer so you

get different levels of security so the Clio app is simple it runs uh basically as a web viewer and it syncs um it gets its log data by different mechanisms but the normal way is on window transaction so every time that fires it does a call to then send that log record with all the detail that you want into the Clio instance which is a which runs in a sequel database and that's the boring part but then what you get is you're basically doing that as a mirror like you're you're consuming data parking it in that SQL or Postgres I mean basically I wrote FM log forever ago right and it's been it's been integrated in dozens of systems same basic idea right everything a user does that you care about you log you store and because at some point in the future you need to know something and by the way did you start doing that because of like HIPAA regulation like because like because like HHS before then but I mean there was a huge one was part of HIPAA so that one not only tracks like record edits and things like that but it also tracks record views so you can put there's different levels of stuff right someone looks at a patient record detail you need to know that um so once the data is in then you can easily view slice dice that's instantaneous but you also get an AI chat bot tool

so basically you can have a conversation with your logs you know what's Matt been up to this last month uh chart the number of new invoices created in April um you know compare um you know compare these two tables it can do all kinds of of you know really really an advanced type stuff just like you would guess including create creating relationships on the fly of multiple systems of data because you can connect it to multiple databases um and then it has a rules-based system so if a certain rule gets tripped like people logging in on a weekend uh records being deleted um here's an important one the log system going offline like if a system hasn't been heard from in a day yeah um you know so uh welfare check yeah exactly exactly um and then it can alert you by various different methods um and what else is key in there uh oh yeah and then the communication but that's actually kind of again this is as cool as that is and you can get it right out of the gate um the really amazing part I think is that you can have the source code so uh you just basically make a git repo so if you want the source code you would need it yeah I would recommend a GitHub account

you definitely clone all the files to that you point your cloud code at it and then anything that you want it to do that's different you just prompt it into existence so so what it might be collecting for example you're saying yeah is what might be different yeah I want to log to this system as well you know if this particular event happens I want to connect it to our weird and terminal communication system that's in whatever place and alert so and so um I want it to summarize the email for these two systems for these tables and create a PDF and email it to someone you know things like that you would never just expect they're going to be in any kind of a tool you can just prompt into existence and they'll go and then it'll be able to you know because it's on GitHub as I add new features it'll be able to merge them with yours and not break your new things that's the thing that's amazing I think that's what's amazing um and it's ridiculous essentially versioning yeah and then it's very very few changes to FileMaker no plug-in no custom function no add-in no uh any of that so basically if you want to yoink it there's just a couple things to change and then you delete

the instance and you're done right I love I love easy destruction if something's not going to work or non-destructive integrations is really what you're talking about because you can just unplug it and it's you don't exactly take a bunch of Deadwood out of FileMaker try it yeah yeah it takes all the Deadwood out of FileMaker because it's taking that anyway I'm super excited about so well so can you talk about how people can get to that if they want to engage with it and then also your class how somebody gets on your list too because I really think there should be I mean there's a ton of people that come but there could be 200 people that show up and they would all get benefit I'm gonna have a problem if it's 200 I'll have to up my zoom account but I can do that um so r.ai my website and not only can you download clear right now but you can actually try it so I've actually made an instance that you can click and play with the logs it's got three systems connected to it and you can try all the things I've just been talking about and it's actually using real AI calls on my real AI key API key um so it limits you to 10. so you can try 10 prompts and then it like closes it anyway um and that's to manage tokens so it's yeah exactly right yeah so people don't go crazy and so basically so like a bot doesn't try to figure out how to co-opt it for whatever purpose oh right yeah that makes sense yeah I have a lot

of rate limiters on all my projects for the same reason so I don't know that's all small stuff I think I think you have even though it's my total world right but it's what I love to do but you've got much much bigger changes two specific things I want to hear about one uh talk about your career what's what's your day job yeah um well first uh just to you know put a bow on what you're talking about I am kind of fascinated with the idea of just AI against logs so well done uh you know it's just the ever expanding universe of information there so well done on that um yeah so I you know one of the nice things about getting to the conference uh I had one of these exchanges that I used to have at all dev cons uh where I'm literally just sitting in line at lunch and to I think I was talking to Ronnie Rios and somebody taps me on the shoulder behind me and says you're that guy from lynda.com right like like the voice is uh more known my voice unfortunately is more known uh than even myself uh but I don't you know that that comes from all the training that I've done and the online stuff that I've done and I used to do classrooms like crazy which mostly got killed by covid but right around 2021 or whenever uh uh GP custom GPTs came out I actually

spent a week or so compiling everything that I've ever done the one the one nice thing that makes what I'm describing here possible is I'm always like every video I've ever done every time you've ever seen me on stage at devcon which is like 50 something plus times I've scripted every beat you never hear me read from it but I but it's important for me to collect my thoughts your your flavor of ADHD or whatever that actually keep and organize that is I don't even understand it because I've also done a ton of presentations and classes and blah blah and I have no idea where those files are and and that's just it it was the ADHD that that hopefully benefits the you know that created this knowledge base so this goes back to so every devcon presentation I ever did I have the script every LinkedIn learning uh course I ever did I have the script um every user group I ever did I mean literally every demo and then now user group yeah gigantic amount of data hundreds and hundreds of words actually I can quantify it actually uh uh as a matter of fact because I've been doing this uh the thing that I'll that I'll tee up here uh but um and then of course once we get into the um the Zoom transcript era you know every important conversation or presentation or whatever and all of our podcast scripts and all that type of stuff um I kept it all I created one corpus of data of this and when the GPTs first came out

I created ask Chris a very limited in what I was attempting to do it was just a hey look here's everything I've ever said about FileMaker in the way I say it and what I mean by that is you know you can go to the FileMaker help system but people have taken my who've taken my classes or watch the videos they know I make some crazy stuff up like you know this concept of query relationships isn't real but anyone who's ever taken a class for me knows that I talk about it as a way to help frame up and really understand the massively compute confusing um you know table occurrence concept and context and FileMaker like what is all that and I've different ways that I talk about FileMaker and analogies I use and anecdotes it's really the same material every time um and I wanted to have like a voice in the training so people weren't just getting stuff that I recorded but like me talking about it and if they like my style or you know found it useful that they could really get me uh training them so I did I I collected a bunch of information what I found out back then Matt was that I had 2.5 million tokens of non do of unique uh FileMaker utterances essentially and I really couldn't do much with it so I did as much as I could and created a GPT you can still go to ask Chris it's out there somewhere but what I did recently is uh my machine over here we we talk

about our claw machines but you can actually see one in the background there that's TARS TARS has been working all week on uh taking that 2.5 million tokens of Chris talking about FileMaker de-duping it doing a ton of research on you know what stuff has been deprecated so that it's current and it's not confusing it with old versions of things um and creating CLIs and endpoints to be able to communicate with it and what I sort of flippantly probably talking about too and yeah probably the same idea oh yeah it fully chunked um you know the just making it really accessible to use essentially open sourcing my my training and what I sort of just I need this is a great song but we gotta get to the drop Chris yeah so the drop is that um I mentioned in the panel that I'm officially retired from FileMaker training uh and I'm going to open source all this material yeah uh and it's provided to anybody that wants to learn about FileMaker subscription free gate free uh they can like be annoyed by me talking about FileMaker till the cows come home and or if you want a human I still do FileMaker so yeah I'm not I'm not poo-pooing training in general just sort of saying you know yeah this is amazing I mean making that available is amazing I think also just in general using using AI to learn FileMaker is amazing you know my kind of job I teach AI a lot right

that's the main thing I've been doing um because it's really hard to get from zero I'm interested to I'm now making programs with an integrating FileMaker there's a there's a lot of mystery stuff in there um well and you know I've had the benefit also of I got hired to do a couple of like presentations and speak at some professional development conferences on the topic of AI education and AI and so I really rabbit holed a lot on research around AI and one of the things that I found that was the most compelling was uh there's this paper in the late 80s from this guy named Bloom I think it was I don't know what his first name was but it was Bloom and he called it um the uh the two Sigma problem um and what he was talking about was back then uh if you look at students in a classroom there's this con we've talked about this before the guide on the side and the sage on the stage so the the sage on the stage and this is something that Linda Weinman from lynda.com was like very prophetic about um and she saw that the video training was the the guide on the side now AI is I think way more profoundly that but what the theory is this this uh two Sigma problem was that this guy Bloom in his studies found that a student uh who achieves a certain level of uh you know test scores or whatever the measure was um in a classroom if they had their own one-on-one tutor

they go up two whole orders of like two Sigma increase meaning that you could effectively go from the bottom 50th percentile to the top five percentile in in comprehension and understanding pretty consistently by just simply having your own tutor now the reason that he called it a problem is saying well this is 1984. we don't have a teacher for every student and we're never going to have a teacher for every student well it's 2026 now man and we do have a teacher for every student they just need information and there's no there's no shortage of that information and so my my idea is to solve the two Sigma problem with FileMaker in our community by providing a corpus of data to allow someone to have a one-on-one yeah um tutor 24 hours don't have to be about raising your hand um I I go further I mean not only do we have one teacher for every student we have one teacher with X number of assistants right kind of describe AI is like um you know AI is basically like having um a PhD a thousand PhD candidates who have who speak every language and in every single field and their PhD candidates so you have to tell them what to do right by getting smarter I like the

math part that's it that's an important you know asterisk and they're eager right like they're motivated to to help they don't they don't sleep they don't get bored uh and for people to really understand I I think so the the spawning multiple agents thing I I I had an example that I think is uh pretty grounded that that I can share there I was doing some like redesign of a thing right and I was trying and I wanted to know like what would users think of this just to be to be totally baseline on this so I went into Fable 3 and I or Fable 5 sorry and Fable 3 sorry uh vibes thinking of and then I said um hey spin up 20 I want you to create a uh a focus group of 25 users they they have the following things in common and I'm paying you a million dollars to run this this uh group for me and these are the things I'm curious about that I want to see people's opinions on and so it's spun up uh the 25 different personas and then you can like if you you know if you if you want to see it if you're inside I'm still a desktop harness like yeah I like the experience actually I I'm made for that not full CLI so in the actual harness and I'm sure in some other harnesses too but you can exposed and see what all the agents are doing and they have all these cool

little status bars and you can see what they're actually you know up to is so cool and I and then I got back and then I did the same thing where I go give me an HTML brochure you know stylized with all the reports of what everybody said it was so cool to like literally I think I upped it to 50 on another one too but just that concept of spawning a bunch of agents what really hits me about that is the that would not be at all that what you even just your prompt would not be at all obvious to to someone but we're gaining all this expertise as we use it to know how it can do things and what it can do like I mean even like my example I was working with a client who had all these papers but they wouldn't know to say chunk these PDFs and embed them using such and such model cross-reference them put them in a repo put them in super baseball blah blah blah you know I I know that because I've done it a bunch of times and because I've done a bunch of research on how to do it but the user might not even know to ask hey how can I make this useful what can I do um so I'm I'm loving building my expertise so but I want to get on now I want to get on to your day job because you you uh uh yeah so so I retired as a filemaker trainer right so that's one less thing that I'm doing um but I guess

uh so I get before I before I share that um uh so it's not about me it's about like this moment in time that we're in right now you know one of the things that I was reminded of getting together with the group you know that was at the conference this past week is something I have to be reminded of like the the AI but I I know you and I we both live in AI bubbles and I'm not talking about some like fictitious like stock market crash thing that people have been predicting for six years that hasn't come yet but um the AI bubble that we live in the echo chamber of AI and I and I'm not saying that like our ideologies of AI are just repeated back to us I'm just saying like yeah advanced AI understanding and knowing what's capable if you just pop your head out of that bubble like I know super smart people in other regions of the of the country even in the U.S that aren't on like you know in California that aren't involved with Silicon Valley somehow that are like wait what's going on now dude like what do you even mean by an agent and they're super smart people and then of course it's like a very you know unfortunately this technology topic has become very divisive even politically so there's some people that are just trying to tow the political line that are keeping their head in the sand on technology advanced which is crazy to me think about if the iPhone was politically divisive at one point everybody carried an iPhone one way and

other people were like boo iPhones like it's it's just as silly places I mean you know like the iPhone's not popular in a lot of countries uh well yeah but I mean the depulsive nature of this technology is no I get that I totally it is it is really weird but I don't I don't I don't really perceive it's along political lines I guess I think it's other I don't know what I tend to simply talking about the capabilities like what is actually what is truly possible right what you really need to worry about what you don't need to worry about yeah which by the way before we're done one of the things I think you don't need to worry about is solutionate the hallucinations and we can we can certainly uh end with that I want to point some people to something that I just put on LinkedIn this week but before that um I also think that the negativity about like the concerns about oh my God I'm gonna you know I'm gonna lose my job you know these are all real I'm not I'm not taking these away from anybody people can own their own feelings and concerns I'm saying I started there okay like you know six years ago when I first stumbled across this I was freaked out dude like I thought it was all over I thought what am I going to do for a living how am I going to get to retirement what am I going to tell the people at iSolutions that I work I didn't even say this discovery of language models that I made I didn't utter anything to anybody at iSolutions for like six months dude you know when I first

stumbled across it I was freaked so I was as negative as you could possibly get the only thing that has made me an I guess an optimist I'm just a practical on you know uh practitioner I guess of AI now but the thing that changed me on the continuum was simply learning more about it actually working with Technologies understanding what's possible and even more importantly working with customers and seeing the technology actually move the needle create a productivity create efficiency and see the look in the face of customers who are now having the same Awakening we talked about this and maybe our last episode together something big is happening I I said there and I know it got picked up and a lot you know and I'm proud of it but like I've never felt more alive and productive and like connected with my own brain um than I am right now in technology and the biggest expression of that to sort of dovetail that same thing we were telling people back in February was I now have because I I let myself learn about AI and I let myself like try some things I since we talked last I start I have a new startup and it is uh on the precipice of being the most successful endeavor I've ever done business-wise but even

more importantly I'm able to be in the most like the rooms that I'm so excited to be in like the the domains that I care the most about that I'm the most fascinated with I dare say my hobbies have now become revenue generating streams for me right and really I've been able to take my life's work this is the thing it's not just the AI is doing this it's that I've been working on machine learning and I've been working in the domain here I'm talking about sports I've been doing this for like 20 years and the problem is and with all due respect to the people I might be referencing they're not listening to this anyways um I've had to rely on other humans to take my thoughts and ideas and concepts and turn them into something real and then I've been constrained by budgets like I didn't have tens of thousands of dollars to pay people to do these things now I have no constraints I don't even have to telephone game explain my ideas to another human I'm explaining it to somebody something that is considerably more wise than I am technically and now I'm taking my whole life's work and I've put it together into this startup and dude oh my God man like people don't care it's sports nobody gives a crap about sports in a community but I promise you I do and so I was able to just say make it so and I want this idea to happen and then it just kicked doors open for me and now I'm in like like literally my world

like the celebrity effect and the oh my God I can't believe I'm sitting here in fact I've checked all those boxes already in the sports world so essentially um I've created what is referred to as a decision layer uh for sports operations and on the field operations that is that is effectively it lives above your data above the ontology layer of above your data and then it's this decision layer that users can interact with and it's you know it's got some chat components to it but my my angle on it is that um I do a I do uh machine learning so instead of just uh consuming all the data that that it's proprietary and organization I consume all their proprietary data and then I run it through proprietary machine learning models that I not only just run one model at a time but like I have this uh mixture of experts approach where I take multiple models that then create this like super rich and proprietary context that people can chat against I take my philosophies of uh personalization to it and a lot of the other technical stuff that you and I have talked about here into this whole context engineering uh mechanism that or but huge professional sports organizations my favorite ones in the world are finding very useful and I

can't believe people are like oh my God you're on to something this is one of the and people like really smart people who have moved the needle in Industries are saying this is one of the coolest applications of this that I've seen dude I can't believe it and so I've been building 24 7. thank you I mean I mean I mean like I don't want to jinx anything you know the bottom could all drop out but but enough about me what I'm really saying here is people can change their lives for the good I did I mean I I mean I could have just woes me my way into you know how am I gonna let FileMaker you know get me to retirement or I could be doing the most exciting thing professionally I've ever done in my life I chose that that Avenue now granted people got to open the door kind of always been hardwired to do sort of stuff like that you know pee on the weeding edge but anyway but I mean yeah and you have to have that entrepreneurial vibe vibe but you know that's what I would say it's like lean into your entrepreneurial vibe like get it you know get outside your comfort zone um and then a lot of people are like yeah easier said than done well no it's actually super easy to do too uh but you just have to be like in a mindset to pull it off I don't I don't think I agree with you it's not it's not super easy to do because there's that there's there's a huge amount of stuff that you forgot that you had to go through to get where you are that's fair um and um it's a slow burn a lot of mistakes a lot

of mistakes and a lot of other things funds you know you've been building all of this other stuff um which is why you're you know you're so far ahead of me as I'm coming along you're staying you're keeping you know actually it's probably more like this well but I mean the the I guess I guess the lesson then is your person take inventory of your personal backlog you know we uh that group that were in the cabal of uh AI people we we kid around about this concept of taste we all went to a conference together earlier this year and someone said you got to have taste what they really mean by taste I think is your personal backlog like my personal backlog was I love sports I'm fascinated with it you know it's the intersection of data and unpredictability that has always fascinated me and then like I had all these concepts you know machine learning was important to me I saw some areas where in sport I thought that things were broken and I I hypothesized on how they might be fixed but these are all things that just bounced around in my backlog for years and years and years and so when this tool came out I said hey backlog let's start pulling some of this stuff out and see what what takes and and whether it's a small little project or the fascinating one you're doing with language you know ciphering yeah I mean that's personal backlog

and so that curiosity and I know but I mean there's like you know I I completed a lot of that stuff and then that opened up all these other doors I just now looked into the folder I've got a folder of all my cloud code projects and each one is you know an independent thing there's 48 folders in there now dude like some of them are some of them are related like in you know like Clio is a folder and then there's Clio customer installs and within there there's a bunch of customers that are installed um so now um anyway I totally get what you're saying right and once you once you start getting going there's all this other stuff now some of them are you know they're going to be not worked on yeah some of them are just dead ends and they're fun and like hey look what I built but there are tiny websites I built a I built a website for a friend of mine who owns an Airbnb you know yeah I would say start there online make a link buy a domain start with websites yeah uh move from websites into SAS software which are effectively websites uh uh anyways and then you know I just continued that progression with machine learning and then brought it all together into a platform you know and layer application into platform yeah yeah man I mean like but I guess I guess the thing is

like there's two pads in front of all of us right now at this time and one of them you can go by but in the long run uh in uh yeah well there's still time to choose the the road you're on I mean I'm I'm trying not to get into Stairway to Heaven or anything either here but uh but uh um I mean I I guess the choice is there the choice I guess I'm you know I've been to the future and I've come back and I've said hey if you choose the choice of just like lean in follow your curiosities that little entrepreneurial spark that made you a business owner in the first place follow that and just see what happens and look I got lucky there's no question like there's just as much equal parts luck and and all the story that I'm sharing with you but it is a potential outcome someone you know did this like right like it's possible and I'm not the only one I talk to people who've I mean I have um customers that I've been following on their own journey that have completely changed their career trajectory like become you know gotten to where they wanted to go professionally by just doing the same type of thing so whatever your goal is even if you're not an entrepreneur you're a ladder climber everything's out in front of you right now and and I think what to bring it back down into the file maker folks like if you take inventory of what you're really good at I think you're going to find out that maybe it's

not you're not just good at file maker like developing you might actually find out you're good at a bunch of other stuff and then lean in on those and see what comes out of it but you know this is really it's rarefied I'm as you're talking I'm thinking about the village I live in um a few hundred people later I I would say that there's probably not a single person who lives here that this would make that would work for their life maybe a handful of people you know because these are totally good yeah yeah yeah I mean their lives are great they don't make a ton of money but they're you know they're fifth sixth eighth generation hundredth generation this town was was uh founded in 1280 I learned recently uh around a spring a natural spring with the best chasing water I've ever tasted uh when you're sitting at the restaurant eating souvlaki and you and you they water or the server wants to give you water he walks over to the fountain that's right there fills up the thing and comes back dude that was yeah that's an amazing water's fine thank you yeah yeah yeah but but I think you're but like yeah okay so I would qualify that by saying you're absolutely right not everybody needs a new journey I mean of course not like in the right place at the

right time the right you know back no doubt I'm I'm I'm saying that that really what I'm talking about is probably maybe closer to some weird timeline switch in the simulation than anything I did don't get me wrong like the luck in the you know I I my religion is um uh opportunity is it or I'm sorry luck is where opportunity meets preparation like I believe very strongly in that and so all I did was just have myself prepared and then the luck of the world and then timeline switches and all sorts of like crazy voodoo happened and all of a sudden cool things and cool things emerged but I guess this advice maybe isn't for the people in your village this is for the people who are experiencing some consternation about their profession that have concerns about what it means you know if they're if bad things are gonna happen or if good things are gonna happen God bless the people who don't think about it at all because you're fine and you can just keep doing what you're doing and maybe you know as a consumer AI will impact your world in interesting ways but um but frankly uh the people who are worried are the ones that I'm really trying to talk to and I'm trying to say look there is an alternative where doing the most amazing things you've ever done before is a reality in your future and it can be done through AI and by the way I am I'm this adventure that I just talked about team of one me

like this isn't a whole new Venture where Chris has a whole new team now now granted there's you know sales people and project managers but like that I think is my favorite podcast right the moonshots podcast yeah yeah I love it I'm a huge fan they talk about like you know there's going to be right now there's a company that exists that's going to be a one-person company like yours that's going to have not only a valuation of a billion dollars but revenue of a billion dollars a year yeah and some at some of those guys opine that it might act I mean they're saying it might actually exist already they just haven't reached the billion yeah it's really easy to think of a million right like a you know a crazy high rate attorney or doctor whatever can can happen in uh income of a million a year or even 10 million a year um you know like real estate developer or whatever but even if you get to that point now you got employees 100 million for one person can't can't imagine it right now yeah I mean I say one person but like you know I still need a project manager because like I need people people customers need to hear back while I'm gallivanting around it yeah yeah but anyway I'm kind of you know sales people like just reality stuff but like actual like you know the investment that I had to make the the biggest part is I no longer have to rely on other the telephone game of explaining my ideas

to other humans and me being constrained by budgets it's so weird because these are core concepts but being constrained by money to be able to make my ideas a reality it's almost like the 3D printing uh which this is the first thing I ever 3D printed by the way right here but it's like the 3D printing of ideas instead of actually having to have it manufactured and uh you know like rely on someone else to go build it and ship it to me and get it into my home so I can experience this product now I just go oh I'd like a project and it goes it comes right to me I think like AI is kind of the 3D printing of your ideas and concepts and potentially business so the first thing I printed 3D was a I printed a 3D printer that prints 3D printers they have those but uh they don't have chip they don't have 3D printers that print chips yeah dude we started with philosophical concepts and we now we're bringing it back we're bookending with philosophy again so um uh no I guess well yeah um so I guess I would say you know people picked up on this in the last time we recorded together but I I'm just here I just want to be a voice of just there's a positive outcome I just want people to just think about the fact that there's a potential positive outcome to all this and I'm and I'm living it and it's glorious it's affected every aspect of my life in a positive way and and

it can for you too yeah I agree with that message you know it's it's improved my life in a lot of ways it's improved the life of many people very close to me and I think that's what's going to change and I think that's what's going to change for me and I think that's what's going to change for me and I think that's what's going to change for me and I think that's what's going to change for me and I think that's what's going to change uh many people very close to me um in a very positive way I think about the guys I know here that own 400 olive trees and a few pieces of property here and there and they're go back Generations and their life is good you know the they have a good don't fix it from that yeah exactly and they're scared about AI because they don't know what it is and it's not going to impact them really um I think they're probably pretty insulated those guys are um yeah but you know I mean people in the world I think it is yeah and it's fine um I just think it's dangerous if you're one of the individuals who will be impacted and you're you've chosen to just turn your back to to it and and just you know like actively I know me too I I really do and and I get it and and I'm here to also say I get it I was just as afraid as you are those people are the company owners who uh and we're talking to all the database developers

and so these are the customers and the people who own those companies and some other individuals who are scared because yeah no it's true a genuine fear I get it I mean I I it's it's I've been there I lived it I I was afraid I had the sleepless nights I freaked out I I looked at I changed my opinion of retirement like all of it I get it I'm not saying I breezed past I'm not saying to you but I'm saying to those people I I was there I lived in it for gosh I was there for probably two years I never did get freaking out yeah never did freaking out yeah but the reason you will I am not freaking out right now is one big difference between me now and me then actual hands-on application with the technology what it what it does what it I know what it doesn't do I know it's possible and then I started applying it to in my own life and in my own business and I've seen the amazing things and then I then in that kind of Theory of Mind concept before I projected this out to other people my customers and even people who I've given this advice to and I see them projecting it back with joy in their face and oh my God my you know I've saved my company or I solved this problem or this was fun all of those emotions all positive and and it's come full circle for me so like it's just undeniable and I wish that for everyone I do too

and that's why we do this you know who else totally does Alexi Folger who I saw in person she said yep I do listen at the end so this is us formally saying to Alexi um thanks for listening thanks for everybody who's listening she comes to all of my classes and she asks the best questions of course yeah yeah yeah they're like thought-provoking they're absolutely right on and they're you know they're usually about yeah but how does this work in the real world what are the security implementations you know yeah yeah it's great well before we leave we should probably tell people again how they can join your class because I think it's so interesting that the Thursday ones here in the States in the morning um 6 p.m for me uh navar.ai right there and what do you call it when they get there what are they looking for um right off the top of the site join free classes okay and it's Zoom and you've got like 50 people there there should be 200 people there I pop in whenever I don't have a conflicting meeting or even in between it's a lot of fun uh so once you join the list I basically send out a meeting invite so rather than sending an email with a link I just copy all the ones into

a Google well actually now it's going to be a Google app strip um and it it sends you an invite a calendar invite so all you have to do is say yes I'm going to go no I'm not going to go uh or you can do or just yeah you just say decline or whatever right yeah so and then when the date comes just like oh okay well there it is on my schedule might as well go right it's so much less friction that makes that makes sense you didn't need it it's just self-selecting um I like it I enjoy it whenever I'm able to make it I really enjoy it it's always fun to talk to those same people we've we you and I have done some rabbit holes together too you're always calling me out I can't just sit in the back of the room and the judge uh you're always pulling me forward um the only other thing yeah so yeah that's an hour of presentation and then half an hour of just general FileMaker AI chat or whatever yeah and I think do you record the first part and then not I record the first part but I only make it available for people who are in the class and ask me there we go there you go I do not want those to be public because you know everybody's faces are on there because it's it's a class it's an interactive class yeah as much as I can it's like hey we've got this product go to my website click this download it open it we're all going to kind of do it together um and then you can see I think I hope everybody does

that uh I'll leave with just the one plug that I have that I'm pretty excited about is um I just posted something on LinkedIn um it's called uh you'll see it pinned on my my LinkedIn profile it's called uh uh my my article title is hallucination isn't a probability problem and um it looks a little mathy but it's actually cool uh approach to and that's actually the core tenant behind this uh this startup thing that I was telling you about this is how I I solve for this and allow people to control truth in a very aggressive manner there's certainly some proprietary stuff I fly by in here but for those that are interested in tapping deep into the core of my AI thinking this is exactly where I'm at right now and and everyone I think everyone should think about this problem as well too so that's my plug and then of course you know I hope that we can get together much sooner I got a couple AI conferences coming up in September by the way but I'll be um well I got invited back to the open AIs dev day uh where I was got that trillion token award last year if you recall that craziness I will not be getting one of those this year I've definitely decreased the well you know I solved that well I added the toe I added the exit condition on that Loop right so uh that was fixed so no more trillion tokens but uh I'm going to

that in September and then the very next day is the uh AI it's one of the other AI conferences I'll be I'll be joining with Ronnie Rios so uh do back to back I'll have some fun stuff to report from both of those conferences next month but hopefully we can have a conversation in the meantime too and keep keep uh you know keep things fresh as all the new models like Astra get released and whatever comes after Fable and all that and in the meantime I'm inviting tons of other people to have conversations uh I got I think I have more time than you with all the stuff you've got going on so anybody who wants to talk about stuff on the podcast hit me up to actually schedule a podcast interview with me click a button boom love that and all that will be in the show notes I hope for this particular oh who knows get AI based show notes uh yeah one last plug right so this is all this is all I do right I do the free classes um I invite anyone who wants to set up a free 30-minute appointment with me to talk about AI or whatever uh and then um I do consulting basically like hour and a half sessions uh to actually implement it in your system using my huge toolbox of things everybody needs that yeah and it's really fun so and that's really that's a class that's one-on-one teaching to get you from

wherever you are to wherever you need to be in so far as I understand it right as I'm keeping growing the learning part has been ridiculously fun maybe more than anything else oh yeah oh yeah for sure or just the the vehicle for learning is probably the biggest game changer here I think a lot of people think that children are not learning right they're actually using it as kind of a cop-out or at least some people think they are uh whereas like we know enough to use it to actually learn as opposed to use it to just answer questions anyway yeah my wife's a elementary school teacher as we've talked about before and they're in a very big no screens movement right now which is opposite of what we're talking about um but with that in mind the whole two Sigma problem is interesting if anybody has a really interesting uh way to utilize this open source knowledge base that I'm giving out um I'm happy to please engage me reach out um you can reach me on LinkedIn if you don't already have all my contact info um and uh I'd be happy to uh donate this uh corpus of of data to whatever anybody's crazy idea might be as well too so in the injury in the interest of as you have your classes in the interest of moving people's knowledge forward

as always fabulous conversation thank you sir we did it buddy thanks Alexi for listening Matt miss you brother glad we had a chance to chat looking for it forward to our next episode