Subscribe
Share
Share
Embed
Built This Week is a weekly podcast where real builders share what they're shipping, the AI tools they're trying, and the tech news that actually matters. Hosted by Sam and Jordan from Ryz Labs, the show offers a raw, inside look at building products in the AI era—no fluff, no performative hype, just honest takes and practical insights from the front lines.
Obviously, a super innovative name, daily check-in. I vibe coded and replicated before. I now, instead of having to read every single one, I get a summarized form and, you know, hopefully, keeps me more up to date faster.
Jordan Metzner:What Clay does is it takes your leads and it allows you to write customized messages to them based on, like, specific data and calculations. OpenAI released, what they call OSS 120b and OSS 20b . And, what these means is these are models that have essentially zero cost to run. Built this week, breaking it down. Built this week, we show you how.
Jordan Metzner:A fresh idea,
Sam Nadler:Hey, everyone, and welcome to another episode of built this week, the podcast where we share what we're building, how we're building it, and what it means for AI and startups. I'm Sam Nadler, cofounder of Ryz Labs. And this week and every week, I'm joined by my friend, business partner, and cohost, Jordan Metzner. How are doing, Jordan?
Jordan Metzner:Hey, Sam. How's it going? Super excited to be back.
Sam Nadler:Yeah. Another we say this every week, but another crazy week in AI. We should probably stop saying that. But just to give you the docket really quick, we're gonna have, quickly present a tool I created this week, a little pivot than our than our normal routine. Then we're going to go into one of the tools we use frequently, Clay.
Sam Nadler:We've just started using it, but, it it just recently also raised some money in the last few days at a a pretty large valuation. And then we're gonna go into some of the latest AI news from OpenAI. Any thoughts before we dive into the tool of the week?
Jordan Metzner:Yeah. Like you said at the beginning, kind of a crazy week. You know, we've seen already releases from Anthropic, Open, OpenAI, and Google. OpenAI has a huge press conference, coming up today, day of recording this episode. Expected, you know, GPT five.
Jordan Metzner:We'll talk about that soon. But, yeah, it's been another hot week. You know, a lot of demand for LLMs and and new LLMs, and, you know, it just puts pressure on on some of the other LLM providers who are not up to speed. So, yeah, lots to talk about this week. Super excited and super excited to what to see, know, kind of what some of these opportunities bring as well.
Jordan Metzner:Just wanna say really quickly, don't forget to like and subscribe, and follow us on YouTube, Spotify, Apple Podcasts. And, yeah, let's get going, Sam. Let's let's check out what you built.
Sam Nadler:Yeah. Let's do it. You know, one of the things I love about sharing every week, you know, hopefully, we can inspire, you know, operators to use AI to build tools that have a, you know, really quick impact on their organization. And some of the tools we present require a bit more effort, and some tools are I really think that we present, you know, anyone can do with just a little bit of time. Obviously, the more familiar you get with AI tools, obviously, the faster you can produce these tools.
Sam Nadler:But what I'm gonna present is something, literally, I built in the last week, launched, on Monday, already made some improvements from feedback from our team, you know, iterated and have a, you know, v two out. So let me give a little bit of context. You know, after operating Ryz Labs for four years now, both you and I have a lot of different teams that we need to engage with and communicate with. One practice we developed quite some time ago was that they fill out kind of a very quick five minute form. It says essentially a digital stand up where, you know, there's really three questions.
Sam Nadler:One is, what did they complete yesterday? What are their priorities for today? And are there any blockers, immediate blockers? But the there's really two goals with this form. One is to have our individual teams really think about, even if it's a quick five minutes, what is the biggest priority for today, and that we can see that.
Sam Nadler:And then number two is they can communicate any blockers immediately. They don't have to wait for a time where, we're in a meeting to bring it up. You know, if there's something that's blocking their progress for that day, I get it first thing in the morning. Over time, our teams grew, and it it, you know, somewhat became challenging for us to read eight, nine, 10 of these forms on a daily basis. So I said, why can't we just have AI summarize this and post to a Slack channel?
Sam Nadler:And that's essentially what I built, was a little tool. So instead of using a Google Form or Typeform, I vibe coded and replicated the form. And then I, basically have a job that summarizes all the forms, and post to a Slack channel for you and I. So, it kind of takes the most important things, and, you know, part of my prompt structure for the summary was to highlight the blockers first because in many ways, that's the most important thing. And now, you know, I don't have to read every single form.
Sam Nadler:It still gets post to their original Slack channel so that if I do wanna dive into a specific form, I can easily. But, I now instead of having to read every single one, I get a summarized form, and, you know, hopefully, it keeps me more up to date faster. Obviously, a super innovative name, daily check-in. As you can see, it's a really simple form. I had one question, which is energy team level optional if they don't wanna report that.
Sam Nadler:As you can see, they fill it out, and every I I can jump into you know, I don't it's not my plan to come to this web app to see their individual responses, but you can see when they filled it out. You know, it it has everything that's going on for that team, recently completed, current priorities. But the real feature here is I get this AI summary, and this AI summary automatically posts to a Slack channel for you and I at a little after 10AM central, 8AM Pacific. And I can always go back to I have, like, the historical check ins. You know, we launched this week, so, you know, it doesn't go back too far.
Sam Nadler:But if I select the date, I can always have them here. And I get some metrics. You know, there, I would say, is a light correlation for the really diligent teams who who are on top of this versus the ones who sometimes forget on on who's completing, what they need to do during the day. It's not perfectly correlated, but I I do have some metrics here on, which teams, how how often they're doing it on time, their completion rate. And then team management here, if a new team comes on board, I can add the team.
Sam Nadler:But, yeah, really overall, really simple build. Built it in Bolt Bolt, and, yeah, we use Supabase for the database and for the edge functions, and I used GitHub for the cron job. That's about it. I mean, I literally built this kind of in between meetings over a few days last week, got it live on Monday, and the teams are already using it.
Jordan Metzner:So, I mean, this is a full stack web app that you built from scratch. So, I mean, I guess the first question is, like, maybe, like, you should give a little background on how much software development you were able to do prior to AI and, you know, prior to Cursor and some of these other tools.
Sam Nadler:Yeah. Prior to AI, I would say I knew a little bit of SQL. You know, I could probably edit a few SQL functions. I'm by no means a developer. Now with AI and, you know, I jumped on some of the tools pretty early, and I've built a few of these kind of small little tools for for our own teams, made it so much easier for me to build something.
Sam Nadler:You know, I wouldn't say this is, like, all that, obviously, all that revolutionary. But, you know, previously, I wouldn't there was no way I would have been able to do this, build a custom little web tool that saved you and I time without AI. And, you know, now I'm learning how to build edge functions and cron jobs, and it's all pretty easy with AI. Cumulatively, probably took me two hours total.
Jordan Metzner:Well, crazy.
Sam Nadler:Crazy. But, yeah, overall, I built a kind of cute little web app that saves us time and in, let's say, an hour and a half, really.
Jordan Metzner:Okay. Cool. Can you just walk me through kind of what the teammate has to do to, like, fill out the form and how it works?
Sam Nadler:Yeah. I mean, it's super simple, but let's let's say you're the accounting team. Bada bing. You choose your team. You fill out, you know, your priorities, like, today's top priorities, recently completed would be maybe closed.
Sam Nadler:July books. Any blockers or concerns? Maybe, you know, client x y z overdue invoice, and team is on fire. And then they just submit the check-in.
Jordan Metzner:Any ideas for kind of, like, additional features or, like, where this is going?
Sam Nadler:Yeah. The I definitely need to add some permissions. You know, some teams have sensitive information, and right now, I don't have those permissions, but I will layer those on. And I've just asked the teams to kinda withhold any super sensitive information for the for the time being. And we have some other tools where this may be a better fit integrated into those tools, so that would be kind of another phase to this project.
Jordan Metzner:Cool. Yeah. Super cool. Great job. Yeah.
Jordan Metzner:And then another idea I thought of just listening to you out loud could be leveraging AI when it does the summaries to do a look back against the last, like, seven days of the work that that's been completed, or the forms that have been completed, just to see whether or not, you know, kind of what they said yesterday has correlation to today, and etcetera, and if there's any kind of something that fell out of the out of the mix or something like that.
Sam Nadler:Yes. Yeah. That's a good point. Like, are we are we completing what we're prioritizing and have AI review that? But one, like, really quick thing that I do wanna bring up just as I as we're talking, This did work really well.
Sam Nadler:Yesterday, one of our teams reported a blocker. So specifically right here, our sales team, and this popped up in our AI summary, ran out of credits for a tool we use. And I was immediately able to get in, top off the credits, to bridge the app for when our kinda plan renews, and that's exactly what I wanted it for. You know, previously, not saying this is, a huge deal, but maybe I would have read that form a little later in the day, but I got the summary. It was top of mind in the summary, and I was immediately able to jump in and fix the problem for her.
Sam Nadler:Is it game changer? No. But I think it's always helpful when your team knows that you're that much faster to react to whatever's blocking them from doing their best work.
Jordan Metzner:Well, great job. Honestly, I think it's really good. It's really cool. Alright. Let's jump in the tool of the week.
Jordan Metzner:I think this one goes back to you as well. But, yeah, let's talk about Klay. They just announced I don't know what series it is, but like a $3,000,000,000 plus evaluation. So obviously, very popular in the sales community. Tell us a little bit about how we're using Clay and what you like about it.
Sam Nadler:Yeah. So Clay is a tool we recently started using. And just to give, like, a high level description, it's an AI powered prospecting and workflow automation tool. So what does that really mean? You know, you can pull leads in and then really enrich those leads with job titles, etcetera, and then build custom workflows.
Sam Nadler:Then you can, you know, connect your, you know, kinda it looks like a spreadsheet. I forgot. They call it a clay table. Then you can use OpenAI to help, like, personalize each of those leads based on different, essentially, functions you write that could be summarizing LinkedIn profiles or, you know, different insights. So the way we use Clay and I would, you know, also caveat that we're we're somewhat at the beginning of using Clay.
Sam Nadler:I think we can go significantly deeper. But, you know, for our outbound sales initiative, you know, previously, we were targeting our ICP and basically asking them, hey. Do you wanna use our services? This is what we offer. And, you know, to some degree, that's a spray and pray approach and, you know, honestly, not that effective.
Sam Nadler:With Clay, we're able to take our ICP, understand exactly what roles they're looking for. So we're able to know and target that they're have two, let's say, front end engineering roles. And then we're able to layer on kind of a function that says, how much they'd save if they'd hire those front end engineering roles for us. So instead of just saying, hey, Jordan. Would you like to use Ryz Lab services?
Sam Nadler:It says something like obviously, the copy's a little bit better, but it says something like, hey, Jordan. I see you have two front end engineering roles available. Did you know if you hired with Ryz Labs, you'd save approximately $20,000 a month on those two roles? And it and, you know, we can be way more granular, way more personalized, directly describe the value proposition, and and and still have you know, still be contacting hundreds of people a day, with this kind of personalized outreach. So, there's probably dozens and dozens of ways, we could get more sophisticated with this, but that's how we're using it right now.
Sam Nadler:And, it's been early, but there's a little bit more traction than our previous method. It did take us a while to, like, get set up and figure out what our angle was.
Jordan Metzner:But so just to summarize, what Clay does is it takes your leads, and it allows you to write customized messages to them based on not just like who they are, but like specific data and calculations and all different types of customization so that each email feels as custom as possible? Exactly. Okay. And then once it writes those emails, does it send them for you or what do you do from there?
Sam Nadler:No. You have to you have to push them to your either your CRM or your email tool we use instantly. So it doesn't send them for you. It it enriches. It uses AI to create that custom personalized message, and then you push that to your email tool.
Sam Nadler:For us, that's instantly that pushes it to, like, whatever sequence you're using. You know? Two email sequence, five email sequence, whatever.
Jordan Metzner:Any other feedback or comments on Clay? Overall, pretty cool tool.
Sam Nadler:I'm sure we could go way deeper and make it even more personalized, you know, potentially call out different things on people's LinkedIn profile. So I think there's additional ways we can make it look like we're doing a lot of research on on our ICPs, which ideally, you know, is is the best way to do it. It's just, you know, hard to do that at scale. So, you know but, yeah, we're we're excited to use it and see where it goes. Yeah.
Sam Nadler:Let me queue you up. I know I think it was yesterday, OpenAI released two open source models. I know that was exciting for you. What does that mean?
Jordan Metzner:Today, as we're recording this, it seems like OpenAI is gonna announce even more new models. We'll talk about that in just a sec. But OpenAI released what they call OSS one twenty b and OSS 20 b. And what this means is these are models that have essentially zero cost to run. So you're only gonna have to pay for the servers to run them versus the previous models from OpenAI were available only through OpenAI's APIs.
Jordan Metzner:And so with this open source opportunity, you can actually host the models on your own computers. And for the 20 b model, you can run that on most computers. And what that means is that you can, you know, use things like have it program all night long while you're sleeping and just use the power of your computer. Or for the one twenty b model, it's a pretty powerful model. It's equivalent to almost 4.1 or even close to o three style model, but you can get it running on a much cheaper infrastructure, something like Grok, or I saw, you know, AWS announced it.
Jordan Metzner:And so what you're able to do is do things that have almost, you know, negligible cost except running the server to to run those performances. And so you basically get, you know, the ability to do almost anything you want with Chatuchiki at a cost of zero. And if that means, the cost is zero, it means you can scale up out, you know, you how often you could run, you know, these these servers, and you can run them frequently and concurrently, and, you know, you can go horizontal and vertical, but, you know, you could just have a server read every single line of code nonstop, like, in your entire code base all night long. And, you know, once it's done, you could have it, you know, start again type of thing and just look for bugs. Right?
Jordan Metzner:And you could have another one that's just working on documentation, another one that's just working on something else, and, you know, literally imagine that the cost of running these LMs is going to zero. And and I just wanted to give a quick demo of it. I don't know if we'll have time to to chop it up or put it in, but this is Grok Cloud. This is a open source cloud solution provider where, you know, they have hosted models that are open source, including Facebook models, and here we have the OSS one twenty b model. And what's cool about BroadCloud is it's incredibly fast.
Jordan Metzner:So I wrote here you're a marketing expert as my system prompt, but you talked a little bit about clay and customization. But, you know, let's just say I wanted to do something like, you know, write me, let's just say, 10 emails to tell people about our new advertising for our podcast built this week, the AI podcast with Sam and Jordan. Cool? So, you know, it may not know about built this week since it's like a pretty new podcast, and it might not be in its repertoire already. But if you just watch the speed, and we're gonna go one, two, three, let's go.
Jordan Metzner:And boom. I mean, email's done seven, eight, nine, ten. And there you go. There's 10 emails. And so you could just see that, you know, maybe they're not perfect.
Jordan Metzner:Maybe we need to go back, you know, maybe I I even wanna say, like, you know, no dashes. And then, you know, let it run again. The last one took five point three seconds. So now it's writing it again with no dashes, and that one took four point two seconds. So you can see just see the power and speed of leveraging these models, and the cost is, you know, going to zero.
Jordan Metzner:And if the cost goes to zero or, you know, close to zero, it means that you can do a lot more things. So yeah. Any questions about that one specifically?
Sam Nadler:Yeah. From its strategic perspective for OpenAI, what do you think this means?
Jordan Metzner:Well, I think what it means is that, you know, if you want to have a commercial model of a LLM, you need to have it better than OpenAI's open source model. Right? And so it creates a moat around their product saying, look, our premium models are gonna be better than our open source models, but, you know, if you wanna compete with us, your models better be better than our open source models also. Otherwise, why would anyone talk to you? So it it it raises the tide, and it creates a moat around OpenAI.
Jordan Metzner:And it makes it look like from their perspective, hey. These are just our open source models. Like, these ones don't even mean that much to us. You know, we'll we'll give these ones away for free. And, you know, eventually, these open source models, you know, even now, are are good enough for many tasks.
Jordan Metzner:And, you know, when the cost is so cheap and so fast to run, then you start to open up your ideas to your mind as to, like, you know, what what can this do or what can we do to, you know, be more efficient within our organization.
Sam Nadler:Yeah. Exactly. And, you know, that's a great segue. Like, hey. These aren't even our best models.
Sam Nadler:We're giving them away for free. What is being launched today?
Jordan Metzner:Yeah. So hard to say. We're recording right now, and I think the launch happens in just about fifteen, twenty minutes. But, you know, rumor has it that GPT five is, you know, supposed to come out today. You know, obviously, that means we'll see something even better than these open source models.
Jordan Metzner:Excuse me. Alright. Yeah. So GPT should be launching today. There were some leaks on the GitHub website showing that there's, you know, multiple GPT five models coming out, including a mini and nano, a chat based model, and then including a standard model for complex coding tasks.
Jordan Metzner:I think it's gonna be impossible to believe that they're gonna launch a model that's not better than what we're seeing from Claude Opus right now on the market. I mean, otherwise, it just, you know, wouldn't grab much attention. So, yeah, I mean, I think it's gonna be like I said, it's gonna be another step forward in LLMs and what people can do with them. I think we might see an update to OpenAI's Codex, which is their competitor, Claude Code. So it's hard to say what's coming, but it's obvious what's coming is another leap forward improvements.
Sam Nadler:Super exciting, just the progress. When did they when was the last major launch from OpenAI?
Jordan Metzner:Well, yeah. I mean, except the open source models came out week. We saw yeah. You know, we saw Gemini launch, you know, some new tools called Genesys. We launched we saw Anthropic launch Opus three point three point one, I believe it's called.
Jordan Metzner:So, yeah, we're continuing to see improvements from, you know, all the major laboratories. But but, you know, OpenAI has been kinda leading the way except in the coding space, and I think this might be their first opportunity to to really take it to the next level.
Sam Nadler:So so, anyway, Jordan, great episode. Don't forget to like and subscribe. Built This Week anywhere where you find podcasts. Any final thoughts before we wrap up this episode?
Jordan Metzner:Yeah. It's been a super exciting, fun episode. Don't forget to like, subscribe, turn on the notification bell so you can get our new episodes that drop every Friday. And, yeah, Sam, great episode. Great job on what you built and super excited to see what's coming up next week.
Sam Nadler:Awesome. Thanks, Jordan. Bye, everyone. Thanks, Sam. Bye.