SaaS in the Agent Economy: Why AI Just Changed Bootstrapped SaaS Forever

0:00Uh, so, I'm really excited about this talk. And can kind of fair warning, uh, the way I like to think about this talk is, um, like I sent it to Rob and Sonia ahead of time, is I said like this is about as close to like religion and politics as I think we get here. Uh, cuz there are some kind of personal beliefs and biases that I'm going to share, but I think this is the message that you all and I like need to hear. Uh, this is not the message that you might want, uh, but I think this is the thing that we all need. And this is not all good, uh, just

0:36to kind of preface it. Um, but but there is like the silver lining at the end. Uh, and so, in Micro MicroConf tradition, uh, this is my why. Uh, we're all here for a reason. This is my family. Uh, when when I very first started listening to the podcast and following Rob, uh, we were only three. Uh, and now we're four. And my son on the right here is taller than me. Uh, so, 13, 15-year-old kids and my wife Amanda. Um,

1:05this is why I do everything I do, right? Uh, to make as much money as possible and have as much freedom as possible so I can do things with them. Uh, and so, I think as we talk about this really sensitive subject, uh, I want you all to keep your why in mind because, uh, I think it helps us frame what's going on in our lives with

1:28AI and SaaS a little better. So, that's it. I told Rob this is going to be a quick talk. No. Uh, I I do think some of us are in big trouble because of AI. Uh, I think to ignore this, and this is the part that I think you need to hear, uh, is me not doing justice to y'all cuz I'm just like y'all. So, Castos, my SaaS

1:57business, been doing it for 8 years. Um I am worried a lot about what's going on with AI and the longevity of my business. And I wouldn't lie to you and say that I think it's all just like sunshine and rainbows, cuz cuz I think it's not. And if you assume that the next 8 years for me or 10 years for

2:16y'all or Where's Harry? 20 years, right? Maybe? Yeah? Uh are the same, like I I think you're lying to yourself. And so, I'm going to walk through kind of why I think some things are different and will never be the same. Uh what we can all do to protect ourselves and our businesses and our whys against uh [snorts] the things that are coming and then how we can lean into and take advantage of these to be productive and proactive and and take advantage of the biggest technological

2:44shift, I think I've ever seen. Uh since the internet. Um because the fact is SaaS is greatly affected by AI. And it's not just SaaS, right? And it's not just tech that's disrupted by AI. And it's not just our jobs, like I think we all have jobs, right? But it's not our roles in

3:06our companies or jobs in general. I think it's the whole economy. And if you pay attention to the news, like this shouldn't be a surprise. So, this is about 2 months old, the beginning of the SaaS apocalypse. Uh this big drawdown, 285 billion dollars of equity erased from the markets. Anybody remember this? Why did this

3:36happen? This was as a result of a blog post by Anthropic about a bundle of skills in the legal space. So, a blog post erased whatever 300 billion dollars in stock market valuation. Uh you know, whatever. You might think people in the public markets are idiots. I think sometimes, uh but but not like that much. They're not that wrong, I don't think. Uh and so, when things like this happen, I think we should pay attention and generally kind of be aware

4:09of what's going on. Uh this is a pretty cool-looking graph, I think, right? It's like uh what is it? Like uh what do you see when you see this graph? Like this is like one of those dreamscape things. Uh so, this is pretty cool graph, but again by Anthropic. And again, like they're really good marketers, right? Like we're going to talk about some of the stuff they're doing later and why you maybe shouldn't pay attention, uh but this is in the blue, potential disruption of jobs by AI, uh by job type, right? So, up top, like management, business

4:41finance, developers. Uh and then the red is like what's act- been actualized already. And so, if I come over here and see my notes, uh 94% of jobs could be replaced. Um and so far, like less than a third have been. And we've seen layoffs every week, right? Um and so, you know, this is from uh Andrej Karpathy, one of the co-founders of OpenAI. Uh all of the red here are jobs who will be negatively impacted by AI over the course of the next few years, uh with a score of like between one and 10, between seven and higher. Uh and this is $3.7 trillion

5:26of wages that could be lost. Um so, is this like everybody happy? Everybody like, "Yeah, this is [ __ ] great. Like I can't believe we're ending the conference on this." Uh there's just one more. Uh the iShares uh ETF that tracks like, you know, SaaS generally down 30% in the last 6 months. Told you this is like what I think we all need to hear. Um but like hey, we're entrepreneurs, we're founders, we've like seen this stuff before, right? Uh so yeah, you know, web frameworks, Laravel, Ruby on Rails, Next.js, Django

6:06made everything easier for us, right? Yeah, totally. Um cloud computing, right? I don't have to buy the $10,000 server to start my thing. I can just go to Vercel and launch this thing for 10 bucks. Totally true. Uh lowered the barrier, made it easier for us all to do stuff. Payments, right? This is uh Stripe's

6:26seven-line implementation of payments. Like again, uh made things easier for us all to do as entrepreneurs and founders. And I would argue a lot of AI is doing similar things, but I think it's also a little different. And so like all of these have enabled us to do more with less. And I think this time is different. And you might be like, yeah, it's different in ways, and

6:51maybe that's true. But I do think it's different because before all these advancements kind of lowered the barrier, made it easier for us to do stuff. But AI to me is just entirely replacing jobs and roles. And if I take like a half a step back, I think that's the big existential risk to us all is just like there's not going to be people with jobs in the medium term. And I'm going to talk about like the long term cuz I'm actually really bullish on the long term. But I think in the medium term we have a rough couple of years, not because of someone vibe coding our product. Like

7:30that's a medium kind of risk. I think the risk in the next couple of years is just going to be the economy. And there's not much we can do about it. But I want us to be ready to like prepare ourselves and our businesses to weather this time. And so, um Does anybody remember this MicroConf? No, this is a Gemini uh image, but I want to [laughter] But I want to uh I want to see, like, talking about replacing roles. Uh okay, so, i- if you've been in business for a couple of years, raise your hand if you

8:05were like hiring 2 years ago. Come on, raise them. Okay, cool. Uh so, like, half, two two-thirds of the room. Okay. Keep No, keep them up. Keep them up. Uh keep your hand raised if you were hiring a year ago. Cool, like most of us. Uh keep your hand raised if you're hiring today.

8:26Yeah. I don't Okay, that's a lot more than I thought, honestly. So, I'm like, oh, [ __ ] Um I talked to uh Liana mentioned, like, I serve as advisor in residence at TinySeed, where uh you know, companies coming in are between, like, a couple thousand dollars MRR to, like, gosh, I don't know, 50 grand. Um And they just got funding, right? So, they're like, I got to go I got to go spend this money, right? Cuz that's the whole trade-off. It's like, hey, we're going to give up a little equity, get a little bit of money so we can grow

8:57faster. Uh even within TinySeed, I don't talk to a lot of people who are hiring. Uh I certainly don't talk to a lot of TinySeed companies that are past like a year who are hiring. And I don't talk to a lot of bootstrapped SaaS businesses who are kind of established who are hiring right now. You know, the comment um from Oh,

9:19Victor's not here. Victor Keller, I've known Victor a long time. He's like, "Why would I hire anyone anymore because I can do everything I need to do with AI to run my business." Does anybody like resonate with that? Yeah, like almost as many people hands that went up originally. I think that's I think that's kind of the meta theme here is like cuz cuz I feel super empowered and like that's a good and bad thing sometimes, but I think that's like the immediate effect on kind of the

9:48economy and how we run our businesses. Um okay, so I want to talk about three and there's probably more, but there's kind of three that step out to me as like immediate kind of threats to my business and and probably yours. Uh and again, I say all this like I'm not like hopefully preaching, but like I think about this all the time cuz like again, my why, my family, like Castos pays our mortgage and our kids to play volleyball and stuff like that. And so I want to be

10:15really aware of this stuff. Um the first is probably the most obvious, right? Is that like our product just gets commoditized. And yeah, you might like roast me for saying someone can vibe code my product in a weekend. And like yeah, they're not going to vibe code Castos or Salesforce or whatever. But I think a lot of times we know like our customers don't need all of the things that we do, right? And could they do just like this one little thing to get this one reporter or this one way to store a file or this one API or something? Like yeah, like they probably could. And is that good enough

10:49to get on the other hand like the benefit of customization and like owning the data and and kind of having control over thing? Like heck yeah. Um I was talking to one of the Conversion Factory guys and I said like we Castos we charge $19 a month. So like first of all, don't ever [laughter] [gasps] start a SaaS business where you charge

11:09$19 a month. It is just it's rough. But uh ChatGPT's 20 bucks a month. Claude's 20 bucks a month. For us to have the gall, I think in some ways, to say we are equivalent value to the most impressive piece of software ever, I I think is weird, right? Like I don't know like we we we churn hasn't really changed for us, but I think increasingly people are going like, "Gosh, I can do kind of like everything with Claude for 20 bucks a month, and then I'm going to pay Castos 20 bucks a month." Uh I think that like the psychology is is kind of changing there, especially

11:45for like new customers. I think like we're starting to see some early signs of this and talking to with some of my peers, and they're like, "Yeah, like churn really hasn't changed, but like new trials are dropping." That that's that's kind of like the early sign of this, I think. And so like willingness to pay for a dedicated SaaS solution, and you see this on Twitter, right? It's like, "Why would I pay like 500 bucks a month for a HubSpot when I

12:06can just make this myself?" Not true, right? But like not false, either, I think. Um you could use Notion and Claude code [snorts] and Airtable or whatever and get like a good CRM that does most of what you need from HubSpot for relatively free. So I think like the perception of the value of your product changing over time is absolutely a thing we need to pay attention to. And again, I think this is what we need to hear, uh even if you don't like it. Um I think the most interesting one is your team. I'm going to talk about what we do uh in empowering our team to lean into

12:44AI in a minute, but I think the most interesting thing here is us getting kind of like Is the word lobotomized where you just like go brain dead, right? I [laughter and gasps] I think uh I think this is a real risk, especially in kind of creative ventures. Like, "Hey, if all I can do is talk to Claude code or Codex or something like, what real value do I bring as a founder or a marketer or a developer, especially for kind of non-deterministic things

13:09like marketing, right? Creative things." Um And then I think there's this existential threat that we all are feeling right now of like, "What the heck am I doing here? If the robots can do everything, like what unique value do I bring? And what unique value does each person on my team bring when Yeah, not today. Like AI can't do everything, but gosh, like think about 6 months ago. Think about 3 months ago what it was doing. Uh and then if we kind of extrapolate forward 2 years, like where are we going to sit in all

13:41this? And then I think this is the biggest one. This is the biggest one by far. Like I know there are products maybe here today that are basically a skill. Right? You're like a connector for this thing and you do this and you transform this stuff and then it pops out the other end. Like I think the reality is most people have no idea what Claude Code or Codex is and they don't know what skills are and so they haven't caught on yet. But if your business is a skill, the frontier models like just deliver this already. They have enormous distribution. Like literally everyone in

14:14the world uses them. And the functionality for this will be free. Hey, Rob Walling here. If you're watching this and thinking, I wish I could be in the room for talks like this. You can and you should. So much of the magic at our events happens in the hallway track. That's where you connect with two to 300 like-minded founders who are trying to solve many of the same problems you are. I hope you'll join us at our next event. Head to microconf.com/events to grab your ticket. Our next one is in Iceland in September of 2026 and then

14:46we'll be in Austin in April of 2027. So who was it that stole this slide earlier? I loved it. I was like motherf- I can't believe that that you asked this. But I think this is the Formspree. Was it Who's Yeah. This is so cool. I use Formspree. Uh and I I use Formspree exactly how uh this like is described. Uh it was suggested to me by Claude Code. I'd never heard Sorry, I'd never heard of Formspree before like 3 weeks ago when I converted my site over and it was like, "Hey, you need a way to send contact

15:20form and you should just use Formspree. Go here. They have a good free plan getting an API key and and give it to me. Right? Um so, this is a graph of Stack Overflow traffic over the last, whatever, 20 years. That's not great. Um I mean, this is not surprising, right? Like, this was the way that many of you learned to develop is like, I'll just go to Stack Overflow, copy and paste stuff in and see how it works, and we'll kind of figure this out. Uh like, times change, right? This is not the most effective way to get value anymore. Uh and so, as a result, like, the market

15:57spoke, and they just get no traffic anymore. This one's super sad to me. Um Adam Wathan, founder of Tailwind CSS, 3 months ago, maybe, laid off 75% of their team. Very unfortunately, because the agents and the models knew the documentation of Tailwind CSS so well that no one had to go to their site. They had a 80% drop in traffic to their site, and I they sold, you know, premium

16:26products or whatever. But, the models and the agents were so good at implementing their free product that nobody had to upgrade to a paid plan. And like, Adam's a super smart guy. Like, I know him. Like, this not his fault at all. And this is none of our faults. Like, I think that's the one thing I want to say is like, what's happening in this transition of AI is not like my fault or your fault or any

16:49It's not Rob's fault even. [laughter] We didn't ask for this, you know? Like, the good and the bad. Like, the I definitely asked for the good, but we didn't ask for the bad, which is we all are forced to reinvent ourselves. Like, as of today. And I hope that's the one thing that you leave with is like, [ __ ] is going to get bad if you don't think about yourself and your business

17:09in a different way. Uh so, you know, moving to the good way, this is uh you know, the folks at Linear are amazing. Like, this is an image that they put out in their blog post about kind of the new version of Linear. And everybody in here uses Linear. Like we don't use Linear, but I know it's like the cool issue tracker. Uh, we just use GitHub. But like they went from like we're an issue tracker that replaces GitHub issues to like surrounding the entire ecosystem of agents, rules, context, and kind of that customer product feedback loop. Which like to me, as a non-technical founder, is the most important thing in SaaS,

17:47right? Like listen to your customers, make the [ __ ] they want, and deliver, and keep doing that over and over. And they're like pretty smart folks. And they're like, "Hey, why don't we just like say that we are the tool that lets you do that really well." So there's like an evolution, right? Of we're not an issue tracker, we're like the thing that lets you deliver on the

18:06most important function of a SaaS tool. SaaS business. So that's cool. Uh, like these guys are like the epitome of a MicroConf type business, right? Basecamp. Basecamp coming out with a CLI, right? This this is like where the puck is going, right? Is you have to have these kinds of tools to enable your customers to get value from your product the way they want to in the future. And like like Rob said, we are super early in this and very fairly we do not have a CLI or an MCP at

18:42Castos. Uh, partly because like we're not the kind of product that like has to have this today, I don't think. Anthony. Yeah, maybe maybe we're not. We we uh, we we debated it as of like 2 weeks ago. We think we have kind of bigger fish to fry here. Um, but we we considered putting off a very very valuable feature to do this. And it was really close. In the end we decided this is going to wait. Um, but I I this has to be part of the calculus one for discovery. Talking about Formspring, like, hey, how do people find us in increasingly? It's like, well, Claude

19:19Code said to use Formspring or Claude Code said to use Cast Docs or SavvyCal or whatever. Uh that's a that's a as, you know, to Amanda's talk yesterday, impressions maybe going up, but clicks are going down. I think part of that is like, we're just getting the answer from a different way. And if you're not showing up where that answer is generated,

19:39you're missing a big opportunity. Okay, so very early, right? When it comes to MCP and CLI and SDK and ABC, uh we're like really early with all this stuff, but I don't think that you can ignore where this is going. Cool? I promise this is going to get better. Um okay, so is it Britney? Like, what a

20:02girl wants? Is that Is that right? Yeah. What a girl wants. What does an agent want? What does an agent want? Context? Ooh, okay. Uh okay. Humans? Yeah, rent a human, right? Who would have thought we'd have a website that's like rent a human? That like isn't that really like wrong and illegal and uh

20:32okay, what what does an agent want? Anybody else? Energy, water. Yeah. Yep. Uh not a really pretty website, right? Like, uh I hate it for the designers out there, but uh they want an API, right? They want clean, visible documentation. It's just like, pro tip, uh a lot of our documentation is rendered on the page in JavaScript, and agents can't read that. So, if you're not like outputting your documentation as like whatever JSON or markdown or something like that. Like probably take the 5 minutes today to do that so

21:11that agents can see and index that. Uh an MCP CLI or SDK. And then like they want it free. They want it free and at most they want to like pay per outcome or something or pay per token or pay per usage. I think AI is overall incredibly deflationary. Uh we're all expecting more for less

21:36money. Uh we saw this with Ahrefs. Ahrefs historically is like [ __ ] gated everything in pricing tiers and overages and stuff and then they rolled back and they're like, "Hey, yeah, our API is like included in every plan now." They're kind of seeing the pressure of what we all expect in a uh in a like giving more away kind of mode and they're the market leader in their

21:58space. And so question for everybody like if an agent, you know, if my Open Claw tries to go use your product right now, like could it? Yes. Yeah, cool. That's the Formstack guy. That's awesome. Um Yeah, I mean for me for me the answer is no, but we're working on it. Um So this is like [laughter] this is like uh how I imagine myself, right? Like I'm like Dumbledore, right? I'm like a wizard who can run my company from the mountains in France or a boat in the Bahamas or whatever and we do provide

22:40value and it's like 24/7 always on. Anybody else like resonate with this a little bit? Like developers especially. Like I I this is like a developer to me is like, "God, you guys are magic." Like you guys can write code and develop products and all this kind of crazy stuff. Software developers aside, I think the the playing field has been leveled a

23:04lot. I I I I do some coaching and I work with someone here and and they sent me a message about a month ago and they go they they had kind of just learned cloud code and they were like, what the [ __ ] Like, if I can do this, everyone else can do this. And so, what what do I what do I do now? What value do I provide? How can I and they're they're an incredible founder and marketer. How do I continue to add the amount of value that I used to to my product when this

23:38is the new baseline? I don't know the answer. I don't know the answer, but I think that's kind of the question. And so, Alex Lieberman, right? Like, pretty smart guy. Anyone else [laughter] Anyone like, [ __ ] at 30 minutes I have an existential crisis the last 30 minutes than the last 30 years. I think

23:58we're all kind of paralyzed, right? Like, what marketing channel do I go down? What product do I develop? Do I acquire a business or not, right? Like, I think I don't know what to do. I don't have the conviction to follow this course that like I used to. And again, maybe I'm maybe I'm projecting a little bit here, but I talked to enough founders that like they just they don't know. They don't feel really good about this is

24:25definitely the course. And so, if that resonates with you, like, you're not alone. But but I think we'll get into how to lean into something here in a minute. So, talk about motes yesterday a little bit. I think someone talked about motes and I was like, gosh, you guys keep This is the the curse of going last is like everybody talks about what you're talking about. I think there's a there's a couple motes. Maybe there's more than four, but I think there's four for sure that we especially as MicroConf founders and like mostly bootstrapped SaaS companies can really lean into the best is first-party proprietary data. So, if your product generates data

25:04that keeps getting better and more robust and more useful over time, you're good. Like, I don't worry about you, right? So, gosh, what's an example? Anybody have an example? Is this anybody? Yeah. What do you do? I do a ton of design research for my clients, so I gather and synthesize it among all of my other clients and make them all a bit stronger in the

25:28process. Cool. So, if anybody didn't hear, like, design research for a bunch of different clients and you're building this kind of library of, like, what great design looks like. Yeah, totally. So, you have first-party data a bunch of across a bunch of different experiences that you can apply to all your clients. That's cool. Anybody else? Example of

25:43first-party data that you're generating? Yeah, Keith, Segmetrics, right? We have data views and then we normalize data between multiple systems that's really difficult and AI is pretty bad. Yep. Yep. So, Segmetrics is uh attribution marketing attribution tool. Uh I mean, so you have I think you have uh proprietary data on two levels. One is like platform level, all of your customers you see a whole bunch of stuff, like, this is where people, like, you know, Stripe and ChartMogul can tell

26:10us how good or bad our businesses are. Like, that's really valuable. And you in a similar way. And then each of your clients have their first-party data that uh that's valuable to them, right? And that keeps getting better and more valuable over time the more they use your product. Yeah, that's really cool. Um this is where, like, uh we win, right? Castos wins here. We are a system of record. We're a podcast hosting

26:29platform. If you want a podcast, you got to have something like Castos. And if you cancel with us, your podcast goes away. So, if you're, you know, a CRM, uh as much as that's like a bad example, like, if you're the source of truth where, like, this is like this is where, like, our data for our company lives. I I think you're good. Um And because of this, I think these are harder to vibe [laughter] vibe code because like if you vibe code this and get this wrong, like the data that you think you have about your company is wrong and then it's not the source of truth, it's just a source of

27:04garbage. Uh and so that's one. Uh I think especially for this crowd, this is a super interesting one uh is speed. I'm much more bullish about the people in this room with SaaS and AI. I am I don't want to keep using HubSpot as an example, but like a legacy big frankly bloated uh piece of software, right? They just can't move as fast as you can to AF AI-ify yourself and your

27:38team and your product and your company. And so being able to move fast, not having to report your decisions to a whole bunch of people and keeping that product customer feedback loop really really super tight is maybe the universal advantage that everyone in this room has that you should really lean into. And then like yeah, hey, what if my next customer is an agent not a human? Well, like that that's one thing, but the other thing is like if you're selling to the enterprise, which like increasingly I think that's where SaaS should go. Uh again, like don't start a SaaS recharge 20 bucks a month. Um

28:16if you're charging thousands of dollars a month, your ACV's 50 grand or something like that. Uh people always have bought from people and they will always buy from people and maybe even more so this relationship between you and someone else is paramount. You know, agents aren't going to decide to buy a $50,000 a year piece of software. It's going to be you and so like leaning into sales uh I think is a really interesting avenue to continue to succeed

28:48in SAS. Okay, so I want to like another show of hands. Like how many of you feel good about how your team is using AI? Oh. Cool. Okay, a little of this, right? Uh and not like really high. Okay, cool. Um I don't do a lot of things right, but but like we [ __ ] nailed this. Like our team is cracked, super clawed pilled. They're amazing. And so I want to share what we're doing. Um Okay, so uh and I I think like to take a step back,

29:27this has to start with you. It can't be like, "Oh, our developer is going to, you know, have their development team like use all these codes and develop all these best practices." Like even if you're not a technical founder, which like I'm not, you have to lead by example. Like I I always want to be the dumbest person in the room in our company. And a lot of times like I am, but in this respect, I

29:52led the charge. And I think you have to. You have to set the culture of excellence here. And like we're not messing around. Like this is how it's going. And I'm going to talk about some really challenging times we had. But you have to get this right and y'all have to lead this. It's no You can't be like, "Okay, I'm going to appoint like a head of AI in our company." Even if you do that, that person's not going to take you seriously if you yourself aren't

30:17fluent and proficient at these tools. Okay? So, uh in our company, we're seven people full-time. Everyone has Claude code. Support, product, marketing, certainly development. Everyone uses it all the time. Every We have a team meeting every Monday, all hands, all hands all seven hands. Uh every time it's like, "Hey, I was playing around with this new MCP over the weekend and we like built this new

30:47tool." "Hey, I was looking at this API for this other thing we could do. It's really cool. Like, I built this script that we can run to get this report." And it's because that's the expectation. I led by example for a painful [laughter] amount of time being the person that y'all make fun of. Like, "Oh my gosh, Craig's going to talk about AI again and how great it is." And all my god, let's just roll our eyes. I did that for long enough and we had some positive peer pressure in the room to where they were like, "Oh, I probably

31:18should like catch up." Or like, "I'm going to get fired." And that was the conversation a couple times. Um We have an unlimited budget for AI. Um I I can't imagine telling someone that you pay 100 grand a year that they can't upgrade their cloud plan from $100 to $200 a month. I just can't imagine that, right? And we're going to talk about what the kind of bigger companies are

31:42doing around this in a minute. Um we have the expectation, not the allowance, but the expectation that everyone spend at least 2 hours a week learning. And they're like, "Well, where would I learn?" I'm like, "You can watch my YouTube channel." [laughter] Because it's frankly like really good. But if you don't watch mine, like watch somebody's um around like the area where you want to like learn. Uh because you have to keep learning cuz this stuff changes so fast. I was like just before this, like they're going to release uh Opus 4.7 like tomorrow and then OpenAI has GPT 5.4 security, which is

32:19kind of their answer to Mythos, I guess. But like, literally it changes all the time. I was talking to someone about Paperclip and I was like, "Gosh, like I'm in this all day every day and I have a hard time keeping up." But if you as a leader don't uh set the expectation with your teams that they have to keep up with this

32:34stuff, there's no way that they will. And so this has to start at the top and then come down. Uh so, so we have two standing meetings in our company. One is like our Monday all-hands, and then we do this thing called co-working, which is just like hanging out on Thursday. But, hanging out turned into what we did this morning in the lobby, which was tell me what you learned or show me what you did with AI

32:55this week. And this does something really interesting. So, the first time this happens, you got a couple of people who show up and stuff. And then I call on someone. I'm like, "Hey, Jimmy, what did you do this week?" Oh, I don't know. Like, not much. Jimmy, next week that cannot be the answer. Like, you got to show up. You have to show me what you've learned and how you are furthering yourself to be able to do your job better in the future. And and like

33:24this just has to happen. Um and I think this is where um you guys know the the term like right person in the right seat? Yeah. Uh what's the traction guy? Um I was having this conversation with one of you this week that like you're like, "Hey, I'm super on board. Like, I'm Claude pilled, but uh my my development team is not. And I'm having to explain stuff to them." And this was a

33:52non-technical founder. Person was like, "Did Do I have like a people problem here?" I was like, "Fuck yeah, you have a people problem. Like, that is not okay. And and we had several of these conversations with our team. And the the last sentence or the kind of the the question that you always should ask yourself in hiring. I got this from Ruben Gomez. Ruben's like one of the smartest founders I know. Follow him on Twitter. Uh is knowing what I know today, would I hire this person

34:20again? So, I ask myself that of everyone on my team as it relates to AI. Knowing what I know now, would I hire this person again in the kind of AI world that we live in. And the answer is about 50/50. Yes and no. And so, I just told them that. Like, it doesn't have to be any more complicated than that. Hey,

34:43Jimmy, I'll use Jimmy as an example. He's like the most amazing human, uh and he started off like, "This is just not I'm not going to do this." And I was like, "Bro, like you got to get on board. Like, I love you, but like for to work here, you have to participate like this or you can't work here anymore." And since he's the

35:00biggest transformation in our company. And so, if y'all have those folks who are like, "Oh, I'm too good to use cloud code to write code cuz I'm an artist on the keyboard." Like, that person has no place in my business and probably shouldn't in yours, either. And so, right person in the right seat doing the right job had a meaning 6 months ago and I think has a different meaning and context today. And so, y'all as leaders might need to have challenging conversations

35:32with your team. Right on. Thanks, man. I thought y'all were asleep. Um okay, cool. So, 200 companies in Tiny Seed, 200 founders in Tiny Seed. Uh get to talk to a lot of them about what they're doing. And you know, just as a as a side note, I have 9 minutes left. Like, we're going

35:51to go over, I think. I hope that's okay. We'll we'll still get to the bar. Um but, we might go over a bit. Um It's the beauty of going last. Uh so, what other Tiny Seed companies are doing? This is a bit rapid-fire. So, uh hired an AI native engineering manager. I like this a lot cuz this is like, "Hey, we had this job before of like engineering manager, really kind of tough job, I think. Now, it's like, how can I pill my developers as much as possible and get them just token maxing as much as possible, right? This is this person's job. I love it, right? We had

36:24this other job before, now we have the new version of this job. I think there's a lot of that going on. Uh automated like task audits, right? If something happens more than a couple times, we're going to automate this. Because now everyone is empowered to do so much around like I'm going to build a script or I'm going to build a little tool or I'm going to automate this thing or I'm going to get this thing in Claude

36:44code or open claw or whatever, right? This should be happening uh to everyone, not just your developers, right? Because your developers are super good at like, "Oh, if I have to do this thing, I'm going to build a bash script to like automate it in the future." Now everyone can do this, which is a huge unlock cuz then it's multiplying

37:01exponentially. Is that a thing, multiplying? It's happening exponentially. Um token consumption. Brandon, here, right? Uh the expectation is your engineer should be consuming 60% of their token quota in a week or month. Uh yeah, like, "Bro, what are you doing?" Like, I wonder sometimes like, "What are my developers doing if they're not maxing out their $200 a month plan?" Cuz I can max mine out and I spend like a couple hours a day futzing around with

37:30this stuff. Uh what are they doing to not be asking for like a $1,000 a month quota? I don't know. They should be maxing out all the time. Uh this is kind of similar to our our weekly uh kind of show and tell, 2-minute demo of a thing that you did. This can be like a skill or an open-source thing that you did or a way that you're doing like Codex and uh and Claude code like review or it can be whatever, right? Like how this happens in your business can be different, but you should be doing it. This positive peer pressure is hugely uh transformative, I think, in

38:05instilling the culture of kind of like AI excellence. Uh I like this kind of like cross-functional squads. Like we've kind of always done this and now it just has a AI flavor, right? So like pick a thing, have a couple people just go super hard on this for a period of time uh and then share what they did with

38:23everyone else. Gosh, this one's pretty tough. We definitely don't have this nailed. The way that we do this most of the time is we just publish our stuff open source. And so everyone can kind of keep keep track of it. But like how do you share prompts and workflows and skills amongst your company in a way that's like consistent and not like getting in the way. And I know like if you're on the Anthropic team plan, you can share skills and stuff like

38:49that. But not everybody is in that boat. So like have a way to do this however you achieve it, I think is important. Talking about token maxing, evidently at Meta, they have like a whole leaderboard around this, right? Like how can I spend the most tokens possible? Um This is where like again, y'all y'all roast me for being so like into this stuff. This is directionally right, I think. Like yeah, do you want to like use the most tokens possible? Like maybe that's not exactly like proper, but do you want this culture of like I did this thing with AI and it's awesome and I want to show you all about

39:27it. Like that's definitely right to me. Um Jensen Huang, founder of or CEO of Nvidia, we definitely don't have $500,000 a year [laughter] developers. But I think the the gist here is right, right? If your developers aren't maxing out their plans and asking for more, you got a problem. And I would take this step further, not just developers, but like if you have a sales person or a marketing person or an ops person and you haven't given them Claude code yet and they're not banging down your door, I I would fire them, like literally today. Cuz it's like where what are you doing? Like if you're working for a SaaS

40:06company [snorts] and you're kind of plugged into even if it's LinkedIn and you don't know about this stuff and you haven't tried, you know, Claude code worker or whatever and they're not asking you for more access to tools like this, that's a personal problem. And to me, that's like an irrecoverable personal problem. That's a person that just doesn't care.

40:26To to me.

40:32Um I think a talk on AI would be incomplete without talk of Open Claw. Anybody using Open Claw in their business? Not a lot of hands. Yeah, cool. I want to talk to y'all afterwards. Um Open Claw, everybody know what Open Claw is? Yeah, cool. Um so, I I think the thing that Open Claw and now Hermes, which is kind of my uh tool of choice here, did is take us for take us from I'm at my computer doing a thing and I have AI that's really cool, uh Claude Code or Codex, right? To this thing is always on, it's proactive, it can work on a schedule, it does things

41:12with inputs that are not me. So, Peter Steinberger, right, sold Open Claw to Sam Altman and Open AI for a billion dollars probably, right? Like pretty cool, first billion-dollar company, a single-person billion-dollar company probably. Uh not an amazing uh sort of an amazing product, but much more of an amazing mindset shift, I think, for all of us, right? Like Claude Code, still my daily driver and the place I derive most value

41:38when it comes to AI. Uh but, if you fast forward 3 months now, you see Anthropic adding little bits of what Open Claw is to their suite of products. Right? Remote control, dispatch, right? All these kinds of things. Uh and sure enough, like the final form that I think this will take is the Codex app or the Claude app doing things that Open Claw

42:05did. So, [snorts] again, uh Jensen saying every company needs an open claw strategy or an agentic strategy. Does everybody feel like they have a strategy around this? Like for customers or for your team to be empowered with agents? Sweet. Awesome. One. I I don't. Two. Cool. Awesome. Yeah? Awesome. Awesome. I like uh We're getting there. We're getting there. Um I want to share something that that like I do, but this is one place like we've not rolled this out to our team. But if you think about like amplification of someone's skills, like surely figuring this out for your team members, not just your customers, is the next

42:58like huge leverage point. Um so, the place where I'm I'm kind of doing this right now is for myself. So, I call this my chief of staff. I hope the QR code works. If not, let me know. Uh but this is a chief of staff set of skills and set up for Hermes. So, Hermes was like the previous slide, uh is a thing like open claw, but it doesn't suck to me, right? It doesn't break. It's way more secure. It's run by uh a research lab. So, they're like kind of, you know, not just vibe code Well, they're probably vibe coding, but like they're doing it better. Uh the thing I

43:31like about Hermes better than open claw is the memory. It's amazing. It it it learns by itself and has kind of self-improving memory over time. It's awesome. Uh so, my chief of staff here is like what a chief of staff should do, right? It clears my calendar. It checks and kind of categorizes my inbox.

43:51Uh and like then you build on top of it. And so, that's kind of the goal of this repo is like, "Hey, this is a starting point. Get you two steps down the road and then build as you will from here." Is this like resonant Oh man, I got a minute left. Sorry, we're going over. Um Does this like Does this seem like

44:10anyone's company today? This is my company like a lot of ways this is my company. Like you got you got amplitude over here. You got search console over here. You got analytics in this place and that place and none of them talk to each other. And I think that like if you're going to take one thing away from this it should be this concept. So like this is life today because this is kind of like as good as we could do, right? We have all these disparate tools in a bunch of different places

44:36and like they don't talk. And that's challenging. But that's like gosh, we're so fortunate to have all these amazing tools, many of which are free, right? Or cheap. But like this is the vision. This is like where you want your company to go to be AI powered. I think where you want your company and your product to go to be AI powered where you have all of the inputs, right? All of the data about your customer and where they come from and what they're doing in your product and how they're interacting with it and getting value and then you do AI stuff with that information and you deliver a better

45:16value to your customer on the other side. And honestly, that's where we are in our company is we're ripping out kind of everything about how data flows and how we relate that data and value to our customers and putting it in a place where we can AI-ify everything that we do for our customers. Because you think about like here's a really specific example. It's just like your website, right? So your website's in WordPress, right? And you're like, "Gosh, I'm going to write a blog post over here in Cloud Code and I'm going to copy and paste it into WordPress and then I'm going to go to Google Analytics

45:53and see how that did and then I got to look at search console and then I got to look at like Stripe to see like how many people converted and stuff and like if you can do all that and like make sense of it, that's amazing. But what if like all of that was in one Claude Code project and you just be like, "Hey, tell me about my customer acquisition from the last 30 days and how we can improve

46:13it." And it would do a really good job, I think. Like you might need to write it to SQL Lite or something like that. But this is where again like that product customer feedback loop allows for rapid uh iteration and testing and better measurement, I think, for everything that you do in your company. And so something like this probably needs to be where most of us are going to interact with the things that we create for our

46:42customers in a better way. Okay, so um this is where the talk gets better. Uh I'm actually really really really really bullish on the future. Uh there's an asterisk to that, but I'm really bullish on this. Um I think that open claw uh this is uh is it hyperbolic? Changed the world. It it changed how we think about what AI can be from a I sit at my computer and chat GPT and like ping-pong mentally back and forth to like it it this doesn't happen, right? It's like Napster, right? Like Napster was not the final version of what music sharing is,

47:23uh but it changed how we think about it. And I think open claw is the same, right? I think agents doing work and providing value for us to our customers is absolutely happening and will continue to happen and I'm super bullish about it. And this is the thing that tipped the scales and started us going in that direction. Uh I think agent to agent commerce absolutely is happening and will

47:48continue to happen in the future. Uh and you should probably be one side of this conversation, right? Uh is SaaS dead? Kind of. Kind of. I think even Rob kind of alluded to this, right? Like and your Q&A's like gosh, like I'm I'm not like this isn't really SaaS. It's like a an API tool. Like some some SaaS's are kind of just API tools, but like is the value we're going to deliver to our customers just agentic and just programmatic in the future and there's no interface? Like yeah, I think so in some circumstances. And so if you're not thinking like how can I be the guy on

48:30the left here to where other agents like open claw and whatever is going to come in the future uh like I that's where I worry about myself and and like you might should too um because I think that like the way we deliver value through software will be through agents. Or whatever form agents take in the future, but it's probably not an interface where people go to a website

48:53and type things into a form. Um vertical way over horizontal. That's pretty easy. I think the reason here is this is just so much more defensible, right? You have subject matter expertise in a way that makes it harder for other people to replicate the value that you deliver. Uh open weight models, open source models. Anyone running open weight

49:19models, open source models? Cool. Awesome. Um this has gotten a lot better like in the last what? 3 weeks, right? Quinn 3.5 and 3.6, Gemma 4. Uh they're really good. Like I would say gent like y'all again roast me afterwards. Gemma 4 is like Haiku right now. Really good. If you're just like, "Hey, I'm going on vacation to to Italy this

49:47summer, give me a 5-day itinerary." Way plenty good. Uh 90% of the things 90% of people do, you can do with open weight models today. And so, as y'all are thinking about like, I'm going to AI-ify my company, but gosh, like every time I introduce AI to my product, like my costs and my COGS increase. Well, not necessarily, right? Get you some GPU or

50:11run this on a a server somewhere. Zero cost, 100% data privacy. Right? If you have Harris talk, you know, people saying like, "I don't want my data going anywhere." Cool, like local finally is an option here in things like Turbo Quant from Google, which like compresses the memory needed by some of these models by six times, uh really kind of opens the playing field for this to be

50:34uh possible. I'm getting a robot. I don't know about y'all. Like, [laughter] I'm about a year away from doing the dishes and laundry and like mowing the grass. Uh this is what I tell my kids, right? This is what I tell my kids. Uh they both are very kind of like I know I'm over time. I'm sorry. I'm almost done. Uh robotics is 100% where the puck is going, right? AI is going to leave the computer and going into things that we do, and this is where I get

51:03super excited uh about like the future. And so, if we want to take a step back and like get some perspective here, uh I think like it's not as negative as maybe I made it out to be, but I don't hear any wallets like gut checking us. So, I want to really like give us some perspective here. Um I absolutely think the future will be amazing, right? It's going to be like Wakanda, right? Like everything is going to be free, basically. We're going to have food and medicine and resources and infrastructure in ways that like we can't imagine today, right? Cuz imagine like there's been a one-person billion-dollar

51:45company probably already. Like that was unthinkable six months ago. And so with the pace of change that all of this is happening, 100% like we are all going to be better off in a few years. There's no doubt in my mind. And and it's those few years that are the challenge, right? It's these next couple of years where we're going to get through the transition of like the conventional economy to like where things are going and the the panacea, I think that we're all going to be in. And so we have to survive, right? And I got to talk about like project glass wing because like I think this is

52:22where this could all fall apart for us, right? Is like we all run software companies and things like the Mythos model that, you know, supposedly are going to like you know, find all these exploits in our software could could be like very bad for us as software companies. Uh and so I think that like we have to stay aware of like the negative potential when it comes to bad actors using AI to do really bad things uh

52:51with our companies. And so like uh I'll wrap it up here. Uh so MicroConf founders have kind of three things that we should be doing. One, move really fast, right? Shipping it testing all the time. Uh two, like stay in your lane and niche down as much as you can, I think. And then like leverage the human-ness and relationships that you have and can have as founders, right? Getting on calls as the founder with your customers

53:19is something that no one else can do. And so if y'all are like this is cool, I want to like keep having this conversation. I'm doing a three-day, call it a challenge, like where I talk in more depth about what I'm doing at my company to get us ready for the next few years. It's free. I have absolutely nothing to sell you. Uh but you can sign up. We're going to do it on Zoom over your three days next week. It'd be cool to have you all there. I can share a little bit more about what I'm doing and give us kind of frameworks for how you

53:45can set your company and yourselves up to win over the next couple of years. [applause] I've been running MicroConf for more than a decade, and the thing I hear from attendees more than anything else is I wish I'd come sooner. Not because of any single talk or tactic, but because they finally found their people. Founders who get it. Founders who are building the same way they are. At our events, it's not just about the talks. In Portland, we organized waterfall hikes, food tours, happy hours, real time with other founders outside the sessions. And those relationships don't end when the event does. They turn into the people you text

54:23when something breaks or when something finally works. If you've been thinking about coming, I hope you'll join us in Iceland and Austin. Get your tickets at microconf.com/events. If you got something out of that talk, make sure you hit subscribe. We're putting out more of these every week. And that wasn't the only great talk from

54:40Portland. Go watch this one next.