Leveraging AI

242 | Inside the Friday AI Hangouts: Real-World Wins, Woes & WTF Moments with AI

• Isar Meitis • Episode 242

Share episode

Is AI supposed to make things easier… so why does it feel like chaos?

You’re not alone. In this special episode of Leveraging AI, we invite you into the room where the real conversations happen — our live, unfiltered Friday AI Hangouts.

Hear real business people share what’s working, what’s breaking, and what’s making them want to throw their laptops across the room.

💡 Whether you’re a solopreneur, in a scaling team, or inside a major enterprise — these discussions will sound very familiar.

One takeaway? The learning curve is real. But the gold is in the room.

🎯 In this session, you'll discover:

  • Why AI Projects are replacing Custom GPTs — and when you shouldn’t switch
  • How business pros are using Claude skills, Comet browser, and Make.com to automate faster
  • The difference between frustration and progress when implementing agents
  • What goes wrong when AI makes up your part numbers (yes, really)
  • How to choose between N8N, Make, and Zapier — without wasting weeks
  • The one piece of advice for managing large AI-driven data systems
  • Real talk: Why AI tools still fail — and how to debug smarter, not harder

This episode features you — the voices of our global AI business community. Want to be in the next one?

 Join our next Friday AI Hangout, open to all.

👉 Register here (https://services.multiplai.ai/events0) — it’s free, it’s fun, and it’s the realest AI talk you'll find anywhere.

About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

Speaker 43:

Hello and welcome to the Leveraging AI Podcast, the podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business, and advance your career. This Isar Metis, your host, and we have a special episode for you today. For over a year now, a community of amazing business people from all around the world who are all AI enthusiasts, who wants to learn together, have been meeting every Friday for what we call the Friday AI Hangouts. There's usually 15 to 25 people every week. As I mentioned, different places around the world, different sizes of companies, from large corporations to solopreneurs and everything in between, and everybody brings ideas, questions, solutions, things they've learned, systems that they failed, and so on, and it's just a fascinating conversation. Every single week, it's completely open to anybody to join. And there are endless number of golden nuggets from these meetings, and I thought it would be interesting for you to listen to segments out of these AI Friday Hangouts to see if that's something you might be interested in joining as well. So this episode is going to be snippets out of different community meetings. Many of them are extremely tactical and valuable and you're gonna learn a lot from them just like you learn from our regular Tuesday episodes. But it will also give you an idea of what it feels like, so you can decide if you want to join us or not. If you do wanna join us, there is a link in the show notes click on the link register. It will appear on your calendar, and then obviously you don't have to join, but whenever you're available on Friday at 1:00 PM Eastern, you can do that. I will obviously like to see many of you and to get to know you better and to welcome you to our amazing community, and now to the episode.

Speaker 10:

I wonder if anybody has any experience, um, sort of taking what they did in a, like in a custom GPT into the new GPT cus uh, chat GPT projects. And the context is this, I built sort of like a prototype for a client. Um, they had new apprentice starting, they've never had an apprentice before and they had no clue how they were gonna do with it, deal with it. And so I ended up finding like what the apprentice's like three year plan was and put that into the GPT and then I found out what systems they used in house and created this thing that they can talk to it, figure out how to teach the person, give it to the kid to learn his things and stuff, and. Now, now they love it and, um, are gonna wanna use it and build on it. And I realized that it would be so much better to have it in this new, like, chat GBT projects because then everything stays in the same place and I can control it, but have more different users. So I'm just wondering, do the instructions still work as well? Has anybody tried that migrating?

Speaker 3:

I will let other people answer before I answer Anyone? No

Speaker 6:

one. Okay. So a few, how many of you need an overview of projects in general before we dive into what are the differences? Okay, so the, the biggest, uh, uh, two seconds on the biggest difference between projects and custom gpt. Uh, custom gpt follow instructions and do them. And that's it. And that's the end of the thing. And a project is a think about a project as a mini private version of Chachi pt, and in that mini private version, you can give it its own universal context. You can give it special set of instructions, but very different than a Chachi pt. You can have conversation basic, basically a regular chat, just like a regular chat with GPT only inside your little universe. And it has also,

Speaker 4:

it has its own memory. So it has its own memory.

Speaker 6:

Here's where I'm going. Uh, it has many benefits right now over custom gpt. Uh, the first one is you can have free form conversations with it versus in chat GPT, it just follows the instructions, right? You can't just have a regular conversation in there. Uh, so that's a big benefit. The other big benefit is that you can upload more files than you can upload into custom GPTs inside projects, which is obviously a big deal. Uh, the third thing, which is huge is what Jonathan just said. It has its own memory in its own little universe, which means in addition to the information that you gave it, when you trained it to do the thing, and you gave it access to files, it learns over time because you, you give it more information as you're having conversations with it. So as of right now. There's really no reason to create new custom gpt. It makes a lot more sense to build them inside projects because of all the things that I just said. Uh, those who already exist in custom gpt, the answer is it depends. And the reason it depends. If it's working, it's working. Like there's no, there's no reason to move it over. Like if it's doing the thing that it needs to do, let it do the thing that it needs to do. But if it, if it doesn't, or if you wanted to have a, the ability to have free form conversations with that in mind, or you want it to learn over time or all these kind of things, then literally all you have to do is upload the same files on the project side, uh, copy the instructions to the instruction sections, and, and that's it. And you're, you're off to the races. So you, it, it, it can work. And if you think about it, projects in Claude. That's what they've done for a while. Like in Claude, you don't have these two different animals. There's one thing which is projects which you can use, kind like projects in chat PT or like custom GPS in chat pt, depending how you want to use it. And so I think when Chachi PT came up with project was to partially mimic, but they should have just converted cha GP like custom gpt into projects. And I think they just wanna confuse people go like, whoa, where about Cha g pt? Like custom gpt go. So, uh, right now it's kinda like duplicate, but the only

Speaker 10:

thing is that you can't, you can't do connectors or something right at the moment with these new projects. So there's

Speaker 6:

a few things. There's a few things you cannot do. Uh, like the, the only real benefit of custom gpt right now is the whole section of the bottom where you can write code and do all these kind of things. Uh, that does not exist inside of projects yet. Uh, but again, I, I, I think what makes sense is that they would merge the two things into one thing that has all this functionality. Uh, but I think that's the only thing right now that doesn't exist.

Speaker 10:

Okay. And have you tried sharing it with people and using the different permissions and stuff? Oh, yeah.

Speaker 6:

Hold on. So, so that's the biggest thing you cannot share. Oh, no. Now you can share everything. They changed it like a couple weeks ago. Yeah, exactly. Yeah, it's brand new. It just changed. I share like two or three weeks ago you couldn't share projects. You could only share custom gpt and now you can share projects as well. You can even share them. I'm confusing, I'm using so much of Claude recently in Claude. You can share projects with anyone, like you can share projects with the world and they don't see the instructions and the data behind it, but they can use the project in chat. GPTI don't remember how they did it, but I know you can now share projects in chat GPT that you couldn't a few weeks ago. So yeah, you, you can share projects.

Speaker 10:

Okay, well I'll, I'll, I'll keep working on this and I, I will report back how it goes. Awesome.'cause I also wanna like have different permission levels. I think it would be good if the Apprentice can use it, but only in study mode or something. Right. Whereas the people that have to work. Yeah. So I

Speaker 6:

think in that case, you probably want to create another version of it. Maybe that would be for the Apprentice. And I think that would make it a lot easier to control where you give it slightly different instructions.

Speaker 7:

Yeah. We'll see. Okay. Thank you.

Speaker 10:

And just before, before while I still have the floor, I just wanted to say thank you to Ken for sending me a podcast after the last, um, session, which was really helpful to my last, uh, awesome la last, last week's issue. Thank you, Ken.

Speaker 6:

This is not about me. This is about us all helping each other and there's nothing that makes me happier than, you know, uh, that if I die tomorrow, this team can continue being helpful to another,

Speaker 38:

Actually I've been listening to a lot of your, um, podcasts, uh, every week and. Trying to use the tools and, uh, the frustration level is unbelievable. Um, it took Gemini to write me a script to merge a bunch of doc Google docs into one doc 38 script attempts, and everyone said, I got it right now I got it right. It here to just cut and paste this into the script. It's the JSON, no problem. Oh, it didn't work. Oh, send me the error stuff. Okay, here's the error. Oh, I know what went wrong. This'll fix it. 38 attempts. So the frustration of these things just don't seem to be ready for prime time every, I did another one where I asked it to create a prompt where I could prepare for a meeting for someone. You, one of the use cases that you, right. So I said, gave me this prompt and, uh, so I gave me the prompt, I gave it the person's URL, uh, on LinkedIn, their name, all this background, and it went and gave me a wonderful brief. Snapshot brief of the, you know, questions. I could ask this person to start the conversation except it was the wrong person. And then when I questioned it and said, oh, I can't look at LinkedIn profiles, that's all private. Well what if I gave my login? Oh no, I can't do that either. So the frustration level I'm having actually accomplishing any task that I'm trying to automate is unbelievable and trying to learn any to end. And your last podcast you had last week with the guest for the French lady, like this is, uh, when she said the learning curve is steep, it's unbelievable. And I'm a relative. I've been in, you know it a long time, relatively savvy guy. This is, but I'm not a coder. It is unbelievably steep learning curve. And I don't know what the group's experience has been, but the consensus seems to be that don't go with the easy ones like Zapier make go with the harder one. But it is a way lot more complex.

Speaker 39:

Oh, um, okay. So I mean, ma make.com is significant, is still e even the other ones have a bit of a learning curve. So if, if, uh, NAN is considered hard at that point, then, uh, yeah, I'd definitely look at, uh, you know, make.com is actually really nicely, uh, um, laid out and, uh, it's fairly and it's very powerful.

Speaker 30:

Yeah, so I agree. I, I always tell people, get started, start with make, uh, from a value to investment. It is by far the best one right now. Uh, it is not as flexible and capable as NA 10. It is way easier to use and definitely way easier to get started. Uh, even just connecting accounts, like connecting your Gmail account or your Microsoft Outlook account make. Is, you go in, you give it your password, and you're done, you're connected doing it in, in NA 10, you go to Google Cloud or to Microsoft Azure, and you have to open a project and you have to assign permissions, and you have like, it's freaking, it's like 20 steps that you need to be an manager to do. So, a hundred percent. Like if, if you're not exactly, if you're not now in the process of I'm gonna automate my entire business and it's gonna save me a lot of money, uh, doing it through NA 10, just go with make, uh, there, there, I'm, I'm

Speaker 39:

actually in the process of automating a lot of our business. Uh, but I'm designing it all in Make first. Yeah. Yeah. And once I have all of that in make, I am going after the ones that use the most operations and I'm gonna move those over, uh, first and use, you know, use it as an optimization problem. But designing it and make is so easy and so fast that it's just to get stuff out there as a, as a initial tool. It's great.

Speaker 38:

It seems now I understand why there's a whole business out there for people to build your N eight N agent for you.'cause it's so complicated. So there's the buy, buy, the build it, there's the build it yourself. Um, and then there's an easier way to build it yourself apparently. I'm wondering, is there anything where it's pre-built, you just saw, you know, we all want to go, I wanna check my emails, I wanna monitor my emails, and if this email comes in, I wanna do X with that email. I wanna reply and I wanna customize reply. So I need a little bit of a brain. So I might have agent kit, but like what's been pre-written, these seem like pretty standard bus business processes. Business is simple. Get the business, do the business, get paid a lot of similar processes. Is there a library out there of pre-written and in agents that do specific things?

Speaker 30:

A hundred percent. So both on Make and on Zapier and on NA 10 and all of those, there are thousands of existing templates. Most of them are free. Uh, and some of them will cost you, okay, pay me 30 bucks. I'm like, okay, I'll pay you 30 bucks to save myself. The three hours of agony. And what I do in many, many cases is I mix and match segments of these templates. So there are long templates. It does all these things. I'm like, okay, I just need this. But now I have this building block that I can reuse. And then there's this other thing that does the other part that I need, and all I have to figure out is how to connect them and make them talk to one another. And the same is true in Make and in NA 10. Uh, yeah. Literally thousands of people are, are sharing them. Oh, a lot more than thousands. Probably tens of thousands of people are sharing them on the internet. And you can, if you just search, uh, for template. This. Now, the, the thing that I'm still puzzled by, and I'm sure it's coming and there's, there's third party attempts, but not inside of NA 10 and make, there's zero reason right now. Why you shouldn't be able to say in English what you want the system to do, and then it will just create the automation for you. A to Z soup to nuts, the whole thing. Test it out. Verify, check that it's working. Ask me for what you need from me. Like I need your API key for this. Here's the instructions on how to go and get it. I need your login to this. Just gimme your login. I need, and you will give it all this stuff and it will run, uh, the close, the two closest things that I have right now that I'm using. And I shared that with you in the past, but since you're new, I'm gonna share it again. Uh, one is I have a Claude project. That has an NA 10 MCP server with a whole set of instructions on how to use the MCP server that I never got to work a hundred percent, but it gets me 85% there, which is way better than I can do on my own in five minutes. Uh, it takes a little bit of time, but it's not a big deal. Like to get the MCP set up and connecting it to Claude and, and then running it. Uh, so this is one good shortcut. Again, it's not perfect, but it's gonna save you X percent. Um, the other one is I now do 100% of my entertainment work in the Comet browser that is now free and available to everybody, so you don't have to pay or be a member of anything. Uh. And comment has been really good at helping me solve NA 10 issues that I'm running into. Uh, sometimes it'll be similar to what you experienced, like, oh yeah, I solved it for you. Like, okay, it still doesn't work. Ah, okay. I know I done wrong and kinda like that. Uh, but usually within three or four attempts with the right prompting and the right reference materials, which is a very big deal. I'm gonna get to your first point in a minute. Uh, it actually helps me a lot. So I do all of my NA 10 work in Comet and it helps me solve a lot of problems I don't know how to solve on my own. Uh, and it's kind of funny because despite the fact I'm asking it to do the thing every time, it's like, oh, here's how you do this. I'm like, no, just do it. I'm like, ah, okay. And then it go, it goes and does it. And you have to do it every single time. But it, but it works. Uh, I am, I, now going back to what I said in the beginning, I'm shocked, literally shocked that anything doesn't have a built-in tool like this already. Uh, and it could be a paid separate tool, like, pay me 20 bucks a month to build. Yeah, sure, no problem. I'll pay it. Uh, just to have that tool run and build automations and test them and verify and everything. So basically run your own MCP server, know how to test it, know how to install it, and so on. Uh, I'm, I'm shocked it doesn't exist yet, but I will be really, really surprised if it doesn't happen in the near future. But I've been saying that for at least six months

Speaker 6:

So how many of you used, uh, Claude skills? Oh. Okay. So about about a month ago, when, when, when the new Claude came out. I don't know when that was, like x number of weeks ago. And then we'll go to you, Richard. But, uh, they came out with skills. So what are skills? Skills are like little pieces of instructions. That could be anything because it's basically like a fancy prompt with knowledge. So think about, okay, the best way to describe it is like a library of prepackaged, custom gpt or broad projects in this case that it can pull at will whenever it needs to get access to it. So think about, I think I showed you this many times before. Do you know, I, I showed you that in a regular chat. In chat GPT, you can use the at symbol to call a custom GPT. Did you know that? Yes. No. No. Okay. So I'll show very, very quickly and then I'll show you, uh, I'll show you the Claude thing. Cool.

Speaker 11:

Do you see Chachi PT or am I sharing something else?

Speaker 6:

Chachi PT Chachi.

Speaker 3:

Okay.

Speaker 6:

Yeah, so if you, if you hit the, a symbol in a regular chat in Chachi pt, usually maybe it needs a little

Speaker 3:

refresh to wake up. Obviously, when you want to demo something, it doesn't work.

Speaker 12:

Yeah. Uh, in theory. I had the same thing. Um, if I was trying to do it in five or five, one, I could only, I'd have to change to 4.0 if I created the gpt. No, no. I've

Speaker 6:

did this a gazillion times in the past two weeks. So not in five one, but I definitely did this on five. So let's go to, uh, GPT five. Uh,

Speaker 3:

thinking at

Speaker 6:

Interesting. Well try it out. It, it conceptually works. Like it gives you a, like a dropdown menu of your, that never happened to me that it didn't work, so I'm wondering why it's not working. Maybe the update took the friend away again. Yeah, maybe. Uh, so usually what happens when you hit the add sign, it opens like a dropdown menu that shows you all your didn't work for me either.

Speaker 7:

I just tried it. Doesn't work. It never de know it existed, but it doesn't do anything for me.

Speaker 6:

Weird. So nevermind. But what it used to do is to allow you to maybe

Speaker 7:

still in an old model,

Speaker 6:

it allows you to use several different, what you're saying in four. Oh, it works. Let's try just, just to show you that it's working. Uh, four. Oh, it definitely

Speaker 12:

worked in, in four. Oh, but they didn't

Speaker 6:

in five. So it is working. I used it in five for sure. Uh, do you, but, but you can kinda see here, right? So you can, you can add whatever this, and then you can give it whatever you want to give it. And then the, the idea behind it was that you can use several of them in the single conversation. So if you had like a three step process, you can use the step one and then the step two and then the step three, uh, just by adding it. And then it knows what happened before and it uses that as the input for the next step. Um, the, so Claude. Uh, scales is kind of the same concept, only it knows on its own to use the scales. So if you look at cloud desktop, which is the way to get them or create them, uh, if you go to your settings and you have to install cloud desktop, but if you go to your settings and you go to, uh, capabilities and you scroll down, then you have scales and then you need to turn them on. And then you can have all these things that it already comes up with out of the box and you need to turn each and every one of them on. Um, canvas, is this why it's doing

Speaker 10:

so good on your Excel and Word documents? Is that like a new skill that it has?

Speaker 6:

So, so, yes. But I think some of them are, are, are built in and you don't even have to turn them on.

Speaker 10:

Okay.

Speaker 6:

So, but you can see two very interesting ones. One is called MCP Builder and the other is called Skill Creator. So if you have Skill Creator on and you tell it that you want to create new skills, you literally have a regular conversation with it and it will give you a package that you can download. Once you download that, you can upload it here and create a new skill. So I have a new skill of my brand guidelines, the tone, I like to use examples of that, the actual documents with my color coding and so on. So if I, if I now go here and I start a new chat and I say, uh, please create a document that will promote the Friday AI Hangouts, which is a great community of people talking about ai. It needs to be no more than five paragraphs long. Describe the benefits and why people should join. Create it as a well-designed Microsoft document. Oh, sorry. Stop. Wait, you know what? I, I actually wonder if what I've done it anyway. Uh, use my brand guidelines there. You'll see what it's going to do. Thinking, thinking about referencing brand guidelines, da, da, da, da. Identify task requirements, prepared to consult brand, and now it's reading the file'cause it knows reading, multiply i brand guidelines directory to understand the brand identity. So it's actually, so these, these, uh, scales are, are packages. They have instructions and knowledge base, just like all the other stuff. So in the knowledge base, it has my brand guidelines, the color coding, the fonts that I use, all these kind of things. It has my logo, so it can use my logo because it's gonna write HTML code and it can take the logo and place it inside the html. It's like stuff like that. Uh, but you can create them for literally anything like you can have, and it knows how to pull them in real time. Now, the biggest benefit of putting them in real time is that until it needs them, they don't use tokens. They don't take part of your context window. So it's freaking absolutely brilliant. So that's plot skills and it's, it's, again, relatively new, but you can create them with the creation scale inside of scale, which is kind of meta. This is, this is how it works. And to call that scale, you literally want to tell it. You want to create a new skill and then you'll say, oh, I have a skill to create skills. Then you will, you'll load that and then it knows how to create the package. But then you can download and re-upload, which it by itself is a real stupid, like if I told you I wanna create a skill, I don't need to download it and upload it, just save it wherever it needs to be saved. But that's a whole, a whole other thing. Uh, so here we go. Let's see what it did. And I will see that's gonna have the colors of my shirt and stuff like that in my logo and pretty cool. Right.

Speaker 8:

And how did it get the context of that data? Did it, you uploaded something before or it's going to your website? Yeah. So

Speaker 6:

you, by the way, you can see it here. So Perfect. I've created the f say, here's your document. The document includes brand align, so it knows how to pull the colors. So these things come from that scale that I created. And the way I created the scale is having a conversation, asking it to create the scale. And then I gave it the brand guidelines by tone of voice, stuff like that. It creates a package. You re-upload the package, you're done. And you can do this for literally anything specific formatting of invoices, like it should literally, anything it needs to know, you can create a scale and we'll call the scale when it needs it, which is freaking amazing.

GMT20251107-174847_Recording_gvo_1280x720 (1):

Um, all right. I need help from all of you geniuses that work, uh, developing and doing all the cool coding stuff and that happen to know the evils of Microsoft. So, I know I probably lost half of you guys, but the other rest that are interested in, um, we develop, we had to go back and develop our own chat bot, and it's a technical chat bot for all of our part numbers and all of our stuff. And we're doing it in co-pilot studio, which is it, it's not as bad as it sounds. It's, it's pretty good. What we're finding is that I see you laughing Isa. I'm, I'm not, I'm not, I, I had good and bad experiences with, with Microsoft copilot, so, so what I'm finding is everything went fantastic. We uploaded the whole catalog in. This is, uh, thrilling stuff. I know all of you guys are losing your minds over how exciting this looks like. I know. I'll send you guys one. Um, it has all of our part numbers and each part number has different variations of, we uploaded everything and then halfway of reading the chart, it starts to hallucinate and create part numbers by adding a one. So for example, part number 1 0 1 6 4, instead of oh four, it says oh five. The weirdest thing. So we got in with the Microsoft folks, uh, with the co-pilot studio team. They looked at it, they gave us a couple suggestions, it didn't work. And then they sent us to deal with OCR. We've been playing with OCR, but that thing is monstrosity. Uh, we're way over our heads with this. Any suggestions? Has anybody had experience with OCR? So I, I do, but my question is, don't you have a digital version of this thing? Why do you need OCR? Well, according to the, the co-pilot studio folks from co-pilot studio, they say that we need to add or add this toggle button in co-pilot studio to allow OCR as a connector so that it can read that table better. But Oh, okay. I'm, I'm, I'm going back. Like I, you're talking about the master document, right? Yeah. Like if the document, if you have the raw data of the document versus the final version, kinda like print ready copy of the document. Yeah. Just use the raw data then post, yeah. If we uploaded that, yeah. We have the digital catalog. For some reason it cannot read. It stops reading like halfway the table, and that's when it starts hallucinating. Uh, how much data is on there? Again, I, I don't, I don't, I know these are, uh, 450 something pages. Yeah, but it shouldn't, you know, I, I would, I would assume it has nothing to do with OCR and again, I, I don't know enough to answer, but OCR relates to, I have a page that I wanna scan, or I have a PDF file in, in its final flattened format, that the data may not be as easy. Yeah. And I wanna read what's on the file versus. I have the source data in a digital format, like a CSV or something like that, or directly for the team. Yeah. I assume you guys have the raw data and not just the final formatting for the book. Yeah, and if that is the case, then it has nothing to do with OCR because it doesn't need to read the page. It actually consumes the data. Mm-hmm. And so my guess would be. That so, so the way all these tools work mm-hmm. The very first step is called chunking. Mm-hmm. And I don't want get too technical, but it's basically taking your data and cutting it into smaller pieces, pieces that then the AI knows how to retrieve based on the vectors that it's generating. Yeah. Knowing how to do chunking for large data is art. Yeah. Okay. And if you don't know how to do that, it is not going to work. I literally just read a very interesting paper that does not necessarily relate to what you're doing, but it relates to what you're talking about on a bigger picture of somebody that was trying to get the AI to also know what's in actually charts and graphs outside of a full document versus just the text and what he did in step one. Is that he created a metadata for each and every one of the charts, and it still didn't work. And what he did then he changed the chunking to include the two paragraphs before and the two paragraphs after each chart. So now it has the context of what the chart is, of what it's, you think about a book. It's like, okay, you're talking about something, they're gonna show us something. Then sometimes you'll analyze the thing. It gives it more context on what's in the chart, so it could actually now understand the chart better. And so knowing how to do chunking, especially if you're talking about 400 pages of raw data, is the trick. I don't know how to do that, but there's people that, that's what they do. And if I had to guess, this is your problem and not, mm-hmm. The OCR aspect, because if you have the raw data, then OCR is not even a part of the equation. Mm-hmm. Okay. So this reminds me of the problem that you saw that you and I, uh, connected on when I was trying to process that 2,500 page word, uh, PPDF document. And even with the raw data, it was almost impossible'cause we couldn't find a platform that was able to handle the, the amount of data. So can, you can definitely chunk data that is way, way bigger than that. Like people are creating databases of like the entire fricking universe, you know, think about open ai, right? They're like, what the hell is cha pt? It's like they, they took the data off the internet and they chunked it. So it's, it's doable. You just need to know how to do it. And, and again, there. People who, that's what they know how to do. Uh, and the, the basic out of the box tools are great if you wanna upload 20 pages, 30 pages, 50 pages, a hundred pages. Once you start getting to more than that, or you start getting into very, very specific information. Like, again, think about the difference between just creating a chat bot that talks about, uh, HR knowledge. How do I, you know, sign up for a PTO. Versus I need the exact part number. Mm-hmm. It's a very different kind of retrieval, which means it requires a very different kind of chunking and data set up. And, and again, sadly, I don't, I I, I'm open to anybody else if you know somebody or a tool that does it, but I don't

Head Shot

Isar Meitis

Host