Gonto: We wanted to talk about our tech stack a bit. We don't have that many topics, mostly because my only Twitter feed this week has been the World Cup. I don't know if you saw, but Argentina lost. I'm sad, but happy that I didn't buy the ticket and the plane, which I was planning to.

Hank: Oh, wow. That would have cost a lot. We're just messing around. We're just talking tech, though I have a completely not relevant to our podcast, but I saw some people having this discussion. I wanted to have it with you to see if you even know who this person is. Do you know who Joey Chestnut is?

Gonto: No.

Hank: He eats hot dogs for a living.

Gonto: Oh, such an American thing. Yeah, we don't have eaters until they throw up here in Argentina.

Hank: Oh, he doesn't throw up. You got to keep it down. If you throw up, you lose.

Gonto: Oh, wow. Impressive.

Hank: Okay, so he ate 76 hot dogs in 10 minutes.

Gonto: Holy shit.

Hank: Also, he's not like a fat guy. He's like a pretty fit guy.

Gonto: That's basically 7.6 in a minute, which divided by 60, it's like it's one. It's just fucking insane. Every five seconds.

Hank: Okay, so here's the question. You know who Usain Bolt is, right?

Gonto: Yes.

Hank: Fastest man alive runs 100 meters in 10 seconds. The question is if these two went one on one and they have to eat a hot dog and then run 100 meters, who wins?

Gonto: I think Usain Bolt wins. Who do you think will win?

Hank: How fast do you think Usain Bolt can eat a hot dog?

Gonto: No, he's a slow hot dog eater probably, but I think he can run a lot faster. He just fucking flies.

Hank: Yeah, but think about this. Joey Chestnut, he eats a hot dog five seconds. Let's say he takes 20 seconds to run 100 meters.

Gonto: Yeah.

Hank: He's done in 25 seconds. Usain Bolt, let's say he's having his best day ever. He's going to run nine seconds.

Gonto: Yes.

Hank: So if he takes more than 20 seconds to eat a hot dog—

Gonto: If it's just one, he can just choke on the hot dog and keep it while he runs. He doesn't even need to fully swallow it. He can keep it in there and then run.

Hank: I think that's a crazy take, but I think, I mean, yeah. Okay. Interesting.

Gonto: We should ask the AI what it thinks.

Hank: I'm asking Joey Chestnut versus Usain Bolt who wins. Thank you everyone for humoring us. Yeah, by the way— GPT's on my side.

Gonto: Oh, it is.

Hank: Same analysis. It's like, it's interesting. It gives a range for Usain Bolt as 18 to 30 seconds because it's like if he eats it in eight seconds, yeah, he wins. But if it's 20 seconds, he loses.

Gonto: Yeah, it's hard. Of course.

Hank: Joey Chestnut, eat one hot dog, three to five seconds. Run 100 meters, 17 to 22 seconds. There you go.

Gonto: Okay. But for this week, we wanted to talk about our tech stack a bit. We don't have that many topics, mostly because my only Twitter feed this week has been the World Cup. I don't know if you saw, but Argentina lost. I'm sad, but happy that I didn't buy the ticket and the plane, which I was planning to.

Hank: Oh, wow. That would have cost a lot. It was like 20k a weekend. Ghana, New Jersey.

Gonto: Yes. So sad, but that was the only thing. So we wanted to chat a bit about the tech stack, and one that we were chatting about before was actually our email stack for marketing in general. And you were saying that you've been using Resend. Share a bit of why and how you're using it.

Hank: Well, I'm using Resend across a few projects. So I have personal projects. That's a no-brainer. So things like Tokens for Good, tokensforgood.ai. Use your spare tokens for charity. Thank you. And I just have it send out an email every week saying, oh, here's the charities you researched, how many you did. If I have some news item, I'll include that. But then I have this other company I'm working with. They have this new product, and I didn't want to get bogged down by their current HubSpot and Apollo junk. And their SDR was slow. So I was like, just give me that list. And they had a Resend account for other things. And so I was like, just give me an API key with this new domain. And boom, now I'm just sending out emails every day. And so I'll get an export of, okay, who's filled out the form and whatever. And then it just sends stuff out. And it's so much easier, and it's so much faster to iterate. And I don't know, it's gotten really smart.

Hank: I've also, with that, built a little local CRM so I can track. With this company, I'm helping with a private beta. So I have a mess of Google Sheets from the different leads we've gathered. And then I'm putting them in a private Slack. And I'm seeing who's actually activated their beta key and how much have they actually used. And so I'm just tracking all that. And then I'm sending different emails based on where they're at in the funnel. But Resend's been really easy for that. I haven't tried. There's another breed of new email platforms. You were saying you're using Loops.

Gonto: I am. We were using Resend and switched to Loops. To be fair, not everywhere. All of the product emails are transactional. We still keep them in Resend because I think the API for them is really good. But for all of the marketing ones with agents now, before Loops, I think, was really good because it's built for developers. But now that we have agents, they have really good APIs that I think help do some of these specific loops. They do have the same thing as Resend to start, like the Contacts API, like Events API and stuff like that. But then things that they have that I like are, for example, one is Triggers, where you can actually pick specific triggers based on a field that gets user-created or something changed or they're activated or something like that. And it's basically built in on the API. And they have also an API that's called Branching, where it starts on a given flow and then you branch to separate ones based on a field. For example, have they activated or not? Have they used this feature or not? And then in each branch, you continue to do things and you can create actually some pretty cool workflows and playbooks using that. That has been like really, really powerful. They also have a really good…

Hank: Yeah. So it's similar to like an n8n or a Tray, but it's more purpose-built for email specifically.

Gonto: Correct. Exactly. You can create workflows on all emails and they have both a drag and drop interface. So if a marketer wants to use it and doesn't like agents, they can drag and drop the conditions and everything. But then if you actually want to use an agent, you can use their MCP and just get shit done using their MCP by creating these branches. I personally use it more with the agents. They also have really good segmenting and filtering and stuff like that. But my favorite thing, to be honest, is how we can connect already to existing services like Clerk for Auth or specific databases or stuff like that. And then how you can actually build these branching and triggers and playbooks. And it was just because it was built in. I think for marketing emails, it was just easier to use without anything on top rather than Resend, but it's not as bloated as a HubSpot. I think it's something in the middle that I like.

Hank: Yeah. Which is funny because HubSpot is the thing in the middle between like Loops and Marketo.

Gonto: Yes.

Hank: I think Marketo is done. I've got some tech that I think is done. Oh, and then I have a question for you.

Gonto: I agree.

Hank: I agree that Marketing Cloud is done in B2B. It's still going to be around for a while in B2C.

Gonto: I think HubSpot is now the enterprise choice. Like, for example, in one of the companies that I help, we're considering moving from Loops to HubSpot, mostly because the Loops integration to Salesforce sucks. So I think when you grow up and you start needing Salesforce and routing and SDRs and stuff like that, you used to use Marketo for that, which was shit. Now I think you use HubSpot and HubSpot is the grown-up version, which the API is worse, but eventually you need Salesforce. It's the sad truth of every B2B SaaS.

Hank: So I disagree on the eventual need for Salesforce. I think HubSpot, CRM, this is how I explain HubSpot now. HubSpot is the team of, if you're using multiple hubs, they call them, if you're using Sales Hub and Marketing Hub and even Customer Service Hub, although I recommend against Customer Service Hub for developers because the way you do emails doesn't render markdown and code and stuff like other things do. But anyways, let's say you're just using Marketing Hub and Sales Hub together. That's like having, okay, it's a bunch of B players, but together they're an all-star team. They know how to pass the ball and it's mainly because they share a database and it's an actual shared database. Whereas like Salesforce with their marketing products, Salesforce's Pardot or Marketing Cloud, they're actually completely separate databases and they have like different data structures. So you actually still have the same integration problem that you would with anything else.

Gonto: Which is crazy. Yeah, and with HubSpot and Salesforce, you have the sync, the API and stuff like that. I've honestly never used, with any of the companies that I've heard, HubSpot CRM. So I don't know. I've used the marketing automation side always with Salesforce. I've never actually even tried it. So I don't even know if it's good.

Hank: That's crazy. I think it's good. It also has the same sequencing tools like Apollo or Outreach. Outreach is dying, by the way. Apollo has done surprisingly well because they're really good at the enrichment part, which Outreach never added. It's interesting. But yeah, I think everyone, I think the default should be HubSpot CRM and HubSpot marketing automation. And then you have to justify moving away from that. Salesforce, I think is, I don't know, it's going to be interesting over the next few years. I mean, I know companies that are like really trying to think like, oh, do we, do we need Salesforce? Like we're spending 400K a year on this. Like we could build our own. I tell them, please don't use Attio. Attio is—

Gonto: I actually used it for Hypergrowth.

Hank: Oh yeah. I mean, it's, I think they're adding features fast, but they're still a little behind.

Gonto: I think Attio is good for VCs. Like it's a lot more based on relationships and contacts and it basically imports your emails and can recommend who can nudge something and stuff like that. So I think in that sense is better for VCs and it's a bit more lightweight, but I agree. I don't know if I would use it for just a typical product. The other thing I've tried recently for emails have been AgentMail, which is basically a product that is built for giving your agent an email. And it's just built so well for that, like the hooks, the API, how it works, it literally designed for your agent to use it. And it's the other thing that I've been exploring lately, for example, for my scheduling assistant, my scheduling assistant now is using AgentMail.

Hank: Yeah. You were telling me about this. So that's better. Yeah. If you have an agent that you want assisting you with stuff, you can just give it an email address. So it kind of has its own entity and that way it has more content. Like basically you can CC it and then it knows this is when I'm supposed to be involved rather than just guessing by reading every email. And also you can give it accounts with different access and different roles.

Gonto: Exactly. And I think the pricing is pretty good. Like I honestly use the free plan from AgentMail, which gives me 3000 emails per month. And for my personal executive assistant, three emails per month is more than enough.

Hank: Yeah. And this could be very good for like if you're experimenting with, you know, AI SDRs or schedulers or stuff that could be interesting. Imagine if you gave every, every AE on your team an assistant. That could be interesting. Huh. Yeah. I want to check that. Hey, here's a question. Do you have a lot of clients using, this is a two part question, one using Clay and two using it, actually not just using it for the enrichment?

Gonto: Most customers that were using Clay are actually moving away. This is my hypothesis. I used to have a lot of customers that were using Clay, mostly because I think Clay was the best way to do outbound. Something that I saw a lot was that companies were doing using Clay to try and see what enrichment and outbound they could do. And then they would scale it in other platforms and they would use basically Clay for experimentation. But lately they can just do experimentation with any platform, like with Resend as you're doing, or with Loops or with any that has an API. So I think Clay is in a tough spot now where they will need to add something to the product or they'll be killed. But what I see at least so far with my customers is people are moving away from Clay because a lot of the things they could do with them, they could do with somebody else. Only thing some of them were keeping was the enrichment. But to be honest, now most of them are actually leaving enrichment too, because it's very easy to just have multiple, like Apollo and Manifolds and a lot of them, you rate them based on how good they are. And then depending on the type of lead we just pick instead of using Clay, like it's not hard to implement that. So I don't know, it's going to be tough for Clay. What's your take?

Hank: Yeah, this is a thing. So like a year, year and a half ago when it was really taking off, me and my buddy Mitchell had a conversation where we both thought, you know what, I think AI is going to be good enough that Clay is just temporary. I think Clay is a fad. And also all the people who were most excited about Clay and talking to me about it, they were just using it for enrichment, which I was like, that's the most expensive enrichment tool I've ever seen, because they were charging, you know, they were charging for tokens. And so I've been holding on to this hypothesis, because I instead of learning it, usually I'll learn a tool like that when it's automation focused, and whatever. But I was like, I know n8n really well. But I think even that, I think n8n still has a lot more legs. But you know, I was like, I'd rather invest in learning how to work with agents. And now it's so easy with Claude Code and Codex that I think all the people who are real power users of Clay are going to move away from Clay and just building stuff directly in code.

Gonto: I agree. I think it's either going to be in code, or even it could be something like n8n. Like I'm working with n8n. And one of the things we've done that we shipped recently for everybody is the chat experience. But the idea that's different to others is that it's actually twofold, where you can talk to get changes through AI, but you could also do it manually. And then you can create your own boxes that you can reuse for other things. And I think what will work is sales and marketing still doesn't believe in code, because they can't see it. So but they do believe in workflows. So if there's a way for developers to build workflows very easily with n8n by actually chatting with it and doing something manually, but then having something that is viewable for the others, I think it can work. And one thing we're planning to test sooner is something similar to like, hey, do you know what you want to build? Tell us. If you don't know, like record your screen on what you do on your typical day and we'll help you automate all of that. So it's like a different type of onboarding to help you do the things that you typically do by just creating skills and learning some prompts from what you're doing. So I have a lot of faith on that, to be honest.

Hank: Yeah. Interesting. I mean, I like that. I've always liked with n8n, it's very good at exposing the code. And before you had this chat tool, I haven't tried the chat tool. But before you had that, I could go to ChatGPT and be like, hey, I want an n8n workflow that does this, this and that. And it would just give me the JSON that I could just upload into it.

Gonto: And now it's even better because now you have two options. You have like chat embed into n8n, but we also built an MCP, which is an official plugin in OpenAI and Claude. So you can actually just chat with Claude if you prefer, and it will create it for you directly using the API and you will just see it in your account.

Hank: There you go. Yeah. Wow. Yeah. It's interesting because I feel like the nuance between like a Clay and an n8n, like it is nuanced, but I think it is a real difference, not just to help you shill your stock here and your clients, but I do like them. Interesting.

Gonto: The other thing that I think a lot about is the GEO platforms, because something that to me was fascinating is I've been helping and working with Scrunch. They were recently bought by SidePoint. So they will continue to do what they do, but they will also start helping SidePoint. Profound, which I think is the big winner in the category, now switched to an agent for marketing. I think if you move to be a full agent for marketing, you're sort of forgetting a bit on the GEO space. I've recently used Gauge, which is now one of my favorites together with Scrunch because they solely focus on this and not on other things.

Hank: How do you spell?

Gonto: It's G-A-U-G-E.

Hank: All right. How do you say that?

Gonto: Gauge.

Hank: Gauge, then.

Gonto: Now I understand why you don't get it. Yeah. Gauge.

Hank: There you go. Perfect.

Gonto: So I know to me, it's going to be an interesting space because to me, Profound was the winner, but moved to a different category. And I think we will need one that is specifically focused on GEO because it's going to be so important that you will need one that is fully dedicated to that. No one that is like, oh, by the way, it's just part of our product, which is what Profound is doing. I have customers use them all. I have some with Scrunch, some with Gauge, some with Profound. So I know it's going to be interesting who's going to be the winner and where they focus on. The Scrunch bet in the future was different, which I like more Scrunch vision in this case, which I don't know if they'll continue now, but Profound's vision was, we'll be your agent for marketing, we'll be content, we'll do everything to help you with GEO. Scrunch vision was different, which is like in the future, agents don't want to browse your human website. They want to browse an agent website, which is markdown, specific connections, different things. For Scrunch was, we're going to help you build your agent website, which is constantly mirroring your human website and we'll just keep it all the time updated. I'm a big believer in that vision. And I think if you focus on that, that is a huge product in the future.

Hank: I think so too. All right. You've given me a couple of branching thoughts. One, Flint, triflint.com, they're doing some similar stuff, like tied into your Google Analytics and like building out. They were focused first on landing pages and just being an easy, like, hey, we will absorb your design system and through your website or Figma or whatever. And then we can spit out landing pages, which are really useful for SEO and ads and comparison pages or whatever. That was a cool product. But now they're building this automation loop of like, we look at your Google Analytics and we turn that into pages that you need.

Gonto: So I like, they also do ads the way they help you build ads and landing pages for your ads as well. That's one of their biggest new things, which I wasn't, I wasn't positive on that.

Hank: So I didn't want to say it, but that is very interesting. So like automating the basically keywords, whether you're paying for the keywords or whatever, having the software that helps you create those loops. Now in theory, you could go build this stuff yourself. Like in either of these cases, you have to have the know-how and like to connect it all up. But that'll be an interesting build versus buy discussion that I think everyone needs to have. Like if you're not going to build it, you better buy it from one of these companies. So that's interesting.

Hank: A related interesting thing is one of my clients, we just were in the migration this month of getting rid of Framer and we were going to go with a CMS. And then I was like, yeah, CMS or hell. We've had our conversations on this. I was like, guys, like let's really live in the future. Like Vercel, GitHub and whatever coding agent the team is comfortable with. Now they're co-founders out of Replit. So they mostly use Replit. I was like, okay, then let's use Replit. And we got this one coding agency that's helped set it up. And I just watched the demo, the latest demo. And the team is like they're frothing at the mouth to like actually activate this thing because it has like Replit at least has the shared skills and MCPs. So I'm like, okay, that's good. I would have preferred just Claude Code or Codex, but it's good enough. And you could still use Claude Code with it. And it's got the shared system. They've been bringing their SEO agency to these calls because they're just going to unleash the SEO agency onto this. And like, it's looking pretty sick. And it's the future.

Gonto: I think I've seen other companies do it. I'm 100% sure it's the future.

Hank: It has to be. And yeah, like it's so streamlined. I'm like, I mean, this is just a Vercel, GitHub. And these marketers don't even need the GitHub seats. The agent has the GitHub seat. Like they don't even have to know what's going on there, which is really interesting. So once this works, I'm going to be like, I got to get some of my other clients doing it. We have to do it with our website, by the way. We're still on Framer.

Gonto: We need to remove Framer and just move to a Claude Code version. I agree. Framer, I think is dying slowly.

Hank: I think it is. Yeah. Because they, you know, I'm curious what's going to happen with Webflow and Framer and a lot of these.

Gonto: Some of them will continue to exist. Like I was seeing the stat today. Only 2% of the US population is paying for AI. 2%. So we live in a fucking bubble. So maybe Framer and Webflow is dead in our bubble. But for the rest of the world, I think it will still continue to exist for some time. So I think it still will eventually stop making sense. But they have years until that happens.

Hank: Framer is in our bubble. But I mean, Webflow, the thing is, like if Webflow and Squarespace and Shopify, they're not dumb. Like I think they've seen enough in this.

Gonto: I agree. But at the same time, most people are not using AI. Like it's only our bubble.

Hank: There's no way it's only 2%.

Gonto: 2%? I saw a stat today, it blew my fucking mind. 2% of the US is paying for AI.

Hank: Wow. Guys, you're not cooked yet. There's still, you still have such a head lead. This was, I had someone that we know come to, someone you and I both know really well, came to me and Mitchell's workshop on Claude Code for marketers. I was like, just come, because, you know, they were feeling like they're behind and an imposter syndrome. I was like, just come and you'll see the people who really don't know how to use this stuff. And at the end they were like, okay, yeah, I'm actually ahead. I'm like, everyone who came to that workshop is now in the top 0.1% of marketers using AI. Like, people are so scared of this stuff to actually use it beyond just, hey, please write this article.

Gonto: I use Hermes as my assistant for everything. It writes my emails, it does everything, and I still feel I'm behind. But the feeling that I'm behind is what makes me be on the top. Like if I weren't feeling I'm behind, I wouldn't be investing my time as much maybe into these things. Well, I don't know. Maybe I would just because it's so much fucking fun. But yeah, I don't know.

Hank: Yeah.

Gonto: Interesting. I'm with you.

Hank: All right. Well, that was fun talking tech and stuff. One, one last note for people who are still with us. We might be doing like a Code to Market meetup, possibly around Browserbase's conference on September 10th in San Francisco or around Resend's conference, October 21st. And then maybe around Supabase's conference, which is October 2. So we might do one of those three, one of those three.

Gonto: Yeah. We need to find a time for the two of us to be together in SF, but I need to go, he needs to go. So we'll find the time and we'll announce it soon.

Hank: Yep. All right. That's the pod today. Thanks everybody.

Gonto: Thank you.