Hank: This is pretty good. I’m curious which design, which design.

Gonto: My favorite one is the middle one.

Hank: Mine too. So that’s why that’s the one I’m working on now. I’m iterating on this.

Gonto: What’s going to be the URL?

Hank: It’s going to be codetomarket.ai.

Gonto: Okay. You need to buy all of the codetomarket domains that exist. Basically, each domain is going to be used for a different thing.

Hank: I mean, why not? We’ve got a little universe going on here. I think we still have a couple in the bank. I’ve still got .tools and .dev in the bank.

Gonto: That’s awesome.

Hank: Today we’re doing another fun screen share thing. You’re going to want to watch the video. We have the video on Spotify, we have the video on YouTube.

Gonto: Click pause, go to your computer if you’re on mobile, or just listen to another podcast, and when you arrive at work, nobody works the eight hours, just watch the video as soon as you get there.

Hank: No, Gonto’s wrong. If you can’t watch the video, listen to this all and then watch it again later. We need the double count.

Gonto: Yeah, we need more.

Hank: We need that engagement. Okay, I’m going to share my screen though. So first up, we’ve got, boom, manifolds.run. I just thought this was a cool little site. Look at this. It’s an ASCII torus bouncing around. That’s all. The simplicity of the site is fun. What do you think?

Gonto: The website is actually white for me. It’s not dark. I don’t know if it’s because of the time of day or what, but the background is all white.

Hank: Oh, really? Are you on light mode? Because it probably matches system, and I’m on dark mode.

Gonto: I have light mode during the day and dark mode during the night.

Hank: Oh, look at you.

Gonto: So I 100% like the website a lot. What I like about it is, go back to the home for a second, they’re actually pretty fucking clear. It’s an enrichment service. How much do you have to explain about enrichment? It’s like, look, we use proprietary data, we use crawling and stuff like that. The only thing I would add to this is show some stats or data on why and how they are better than others. We’re using them with Clerk and they were so much better than Apollo and Clearbit and all of the others at matching rates, at least for developer tools. So if you’re a developer tool, try Manifolds. They are pretty fucking good at matching people. They’re pretty fucking expensive though. And just put that data on the website. And pricing, also very simple, but very, very expensive. This is more than Clay charges. They are really good, but very expensive. But what I do like about them is they only charge you when they have a good match. If they don’t have a good match, they don’t charge you. They charge for success, which I also thought was good. Not very well explained on the pricing page either.

Hank: Yeah. See, all of the goodness about this company, you’re explaining verbally now. Which, first off, we’re not sponsored or anything. Maybe we should be, but just, disclaimer there.

Gonto: Yeah. So they could probably tighten up their copy.

Hank: But yeah, I just thought it was fun. Okay, here’s an interesting one. Cognition.ai. They relaunched their site. It’s totally different than it was before.

Gonto: There was an episode when we chatted about the PlanetScale websites, and we were talking about how when everybody is zigging, they were zagging, and their zag was: we make it full text, we do the explanation. I think we’re starting to see more and more pages that are all text. So I think the zag is slowly starting to become the zig, and then the question is going to be, what’s the next zag? Having said that, there’s one main thing that I hate about this landing page, and it’s the fact that this landing page is about Cognition. They explain a bit in the text. They don’t explain as much. They don’t talk a lot about differentiation. They have the logos, but they want people to go to either Devin’s website or Windsurf’s website. If you look at the Devin page, it’s basically the same as any product. It has the hero, it has the image, it has the enterprise, the pricing, the header, the nav, the login, the contact sales. And if you go to Windsurf, it’s literally the same. So they have one main home that is just text and different, but they link in sort of a hidden way to two products that are completely separate and completely different, because they were lazy to integrate them, because Devin and Windsurf work really well together. So what they should have done is actually update the full Cognition page to talk in that page about Windsurf and Devin, and talk about what each of them brings, what they do well together. And they could have just picked one style, because this website is completely different to Devin, completely different to Windsurf. So everything is a mess. I don’t like it at all.

Hank: I agree with a lot of it. One, yes, great point. Everybody’s starting to zag, so it’s time to zig again. And come on, this is so much text. This was made by a brand agency or something, or a designer who got excited about manifesto sites. I do like how they do the logos. When you hover the logos, they only highlight the ones that they have a case study for.

Gonto: Oh, that’s pretty cool.

Hank: I don’t like that some of these case studies are ancient. Like, this one’s from 2023.

Gonto: You can’t have everything in life.

Hank: You can’t have everything in life. Anyways, that’s kind of cool. And then, yeah, I would love a Devin and Windsurf explainer. At least give me a little toggle here to be like, oh, Devin is this and Windsurf is this. And if you’re coming from the site, instead of going straight to the homepage, maybe it gives you a different context or something. There’s stuff to be done here. Well, I work with them, so I have some insider info. I think, look, by the time we publish this podcast, and with the number of people who see it, these brands are going to get merged, so there will be less of a problem there. But there’ll be some interesting stuff on the Cognition and Devin side. Good feedback.

Gonto: Yes, great feedback, and heard ahead of time. They’re not stupid. But I mean, it was an acquisition. It’s kind of bound to happen that you merge the brands, and then it will be more similar to Anysphere and Cursor, or Anthropic and Claude, etc. Cool. All right. So those were two interesting sites. Now let’s do a quick review of our site. Remember this guy? How are we feeling about our site these days? My take, at least: I really like the colors and the icons on the left and right. What I don’t like is the CTA of notes and transcript, I think it sucks. And then how it goes to the second page, I think it sucks. So I think we did a good job with the home. We didn’t do a good job with the linking and how people navigate. It’s not like the whole card is linking. It’s so fucking weird.

Hank: We hired a designer to help us. Shout out to Yousef, Hassan’s little brother from Together AI. He’s great. We put all of our budget onto this page.

Gonto: Yes, no budget on the other pages.

Hank: And what’s great though is this has been on Framer. It’s been easy for us to use. But as you’ll see, I think we can do better on our own. So I think we’re going to do some cool stuff. And yeah, everything opens a new tab. If you click blog, it opens a new tab. If you click any of the episode links, it’ll open a new tab. It’s painful, but hey, it gets the job done. And I think you’re right, we’ve got to get some improvements. And I think, still, after almost two years of doing this, it’s still kind of a refreshingly different site than most sites you see. So I’m going to give us a pat on the back for that. But yeah, there’s a lot to be desired here. You can tell that we didn’t spend much. But hey, that’s why we need you guys to watch twice. Okay, this is codetomarket.fm. Let’s go to codetomarket.app, which is where I put the Code to Market Summit.

Gonto: Ooh, you got effects.

Hank: Oh yeah. You want to see the effects? All right. Now look, I’m just one guy, but you ready?

Gonto: Yeah.

Hank: Zoom.

Gonto: Wow, that guy skis as well as me. That’s a pretty steep gradient there. I also like this box. That’s nice.

Hank: I’m working on this. Yeah, this is just me and Claude Code. What’s fun about this is, and all of this copy is going to change in the next 15 hours for us, for anybody who comes to codetomarket.app, which is also the place where ctm.fm/summit goes. If you go to one of those addresses, you’ll land here. When I tried to do this for the first Code to Market Summit, when did I make that site? November. I made a lot of this design in Figma and I couldn’t for the life of me get it to work in Framer. I was going to do this just as a new route on our Code to Market site, and I was trying Lovable, I was trying to work with Claude. I hadn’t really gotten Claude Code correctly, I think I had tried a little bit with it. Now, building this was so much easier. It’s so much easier than trying to get an engineer. I’m amazed. Even the dumb stuff, like hey, if you scroll, let’s get that ski guy to zip. Simple stuff.

Gonto: Is it Claude with Next.js?

Hank: No. It’s going to blow your mind. This is React, Tailwind, Inertia, and Laravel. This is a Laravel app, baby.

Gonto: Ah, come on. This is great. It does look good.

Hank: Well, hey, I still get credits at Laravel, so I’m using my account. There are even little things, like I wanted this little CTA box to accelerate its scroll so that when you get to the bottom it tucks nicely here. So that was fun.

Gonto: Okay, it looks awesome.

Hank: Are you ready for the Easter egg on the site?

Gonto: Yeah. What is it? Do you have a code?

Hank: When you submit the form. Oh, and I’m going to add it as a Konami code, too. Are you ready?

Gonto: Yeah.

Hank: Here we go. Do you remember this game? Did you play SkiFree?

Gonto: I did play this game. How do you get it to show up?

Hank: Well, you can just go to codetomarket.app/skifree, or once you submit the form, it’ll forward you to this page as the confirmation page.

Gonto: It’s a pretty good confirmation page.

Hank: I know, right? Also, this is another thing I tried to do a little while ago. Oh, there’s the Yeti. Oh no. You can see I hacked the end screen. I’ve got a CTA: the Yeti got you, ski a real mountain, no Yeti included. I’ve also got a post on X, which has, this is a dynamic OG card, because of course.

Gonto: That’s awesome. I’m feeling some Vercel conference vibes.

Hank: You’ve always got to be looking for the little hacks. So we’ll see. Okay, I’m going to continue shilling my own projects here. Which, by the way, you’re going to come to the Code to Market Summit this next year, right?

Gonto: Yes.

Hank: Awesome. Okay, next thing. Hold on, maybe I can find something really cool. Here we go. Yes.

Gonto: The shetty was awesome, by the way.

Hank: The Yeti, yeah. I love how you say Yeti.

Gonto: Shetty.

Hank: Okay, you ready for this?

Gonto: Yeah.

Hank: I told Claude Code, one, this is an announcement also. Me and my buddy Mitchell, and you’re going to help out a little bit, we’re doing a live workshop, Claude Code for Marketers, in June and July. And check this out. Just another site now, yeah, the site’s cool, whatever. But I’m curious if you would have picked this, because this is what I had Claude Code do first. I’ll share this tab now. I told Claude Code to get into my Figma files and to look at the Summit site and our FM podcast site and to come up with different styles. I was like, I need a new style, but still in the Code to Market cinematic universe, and give me some options. And on its own, it was like, oh well, by options, I’ll give Hank this cool side-by-side thing, because I told it to just spin something up in HTML. So it gave me this thing where I could look at these different options, and I thought that was so cool. It wrote this copy, which is very Hank vibes, this copy for marketers who’d rather ship. And the reason it had some of the ideas on some of this copy is because I’ve been spending the last week pulling my Granolas and the last 10,000 emails I’ve sent, and doing all sorts of different stylings and stuff. I’m trying to be like you with your Gauntlet Bot and OpenClaw. I’m trying to train it to speak more like me. So I was like, wow, for a one-shot, this is pretty good. I’m curious which design…

Gonto: My favorite one is the middle one.

Hank: Mine too. So that’s why that’s the one I’m working on now. I’m iterating on this.

Gonto: What’s going to be the URL?

Hank: It’s going to be codetomarket.ai.

Gonto: Okay. You need to buy all of the codetomarket domains that exist. Basically, each domain is going to be used for a different thing.

Hank: I mean, why not? We’ve got a little universe going on here. I think we still have a couple in the bank. I’ve still got .tools and .dev in the bank.

Gonto: That’s awesome.

Hank: So I could see us using .tools for, like, oh, when we have a bunch of…

Gonto: Yeah, we create tools for people to use, for marketers.

Hank: Yeah. Okay. So you like this one? Why do you like this one?

Gonto: I’m done with black.

Hank: Same. Yeah. I mean, we were done with black when we made the podcast site.

Gonto: Yes. I’m done with black.

Hank: Also, this isn’t going to be just for devtool marketers. This is going to be a little more generic. So yeah, even though the copy, you can see it, it knows our brand and it knows me. Oh, you market to developers. So I need to do a lot of copy fixing, but the components of the site, these are fantastic.

Gonto: This looks pretty good. You should show Mitchell, too. It’s just you on the website.

Hank: Oh yeah, here’s a more updated version. I got Mitchell. I need to put pictures in. So there you go, everybody should come. Oh, this thing. You can do so many things, guys. Like this form. If you put your email address in and you select this, it goes to a different Stripe link, because I’m not doing a full Stripe checkout, I just send you a link. But it will at least prefill your email on the Stripe form and stuff like that. There are so many things where you can just smooth the rough edges. I probably should just do a Stripe checkout.

Gonto: You should.

Hank: I don’t know if I can do that by tomorrow. We’ll see. Maybe by the time people listen to this. Okay, I have one more thing. This site is really crappy. It’s a project I’m really excited about, and I hope to get other people excited about it too, but I also hope to show that a non-coder like me can do some interesting stuff now. And if you’re not doing interesting stuff, well, one, you should come to Claude Code for Marketers, and two, you should be trying this stuff. So let me get this. This next project is called Tokens for Good. So, Tokens for Good. Here’s the idea. You have spare AI tokens, and I have a charity that I’m working with called Fierce Philanthropy. What they do is research and grade charities, and this guy was doing it one prompt at a time in Perplexity. He would do a chain of six prompts to kickstart his research, and then he has a further review process. And I was like, Todd, we can automate that. So I automated it, and then he was like, well, Hank, if I wanted to do a million charities, how much would that cost? If I wanted to provide the research so that donors can really understand what they’re doing. I was like, well, that would cost like half a million in tokens. Or we could crowdsource it. It’s similar to SETI@home, or AlphaFold’s Folding@home, where instead of giving your GPU cycles to folding proteins, you can use your unused tokens on Claude Code routines, and soon other types of cron jobs and schedulers, and it can effectively cost you nothing, or pennies, and you can give this research. And you can see here, if I go to contributors, I’ve been testing this out already on 212 orgs. I’ve got my little brother, and I’ve got his Qwen. This is his locally hosted Qwen. And it’s very interesting. I could go at length on how it works, but let me ask your marketing advice. This is the page that really needs improvement, because this is the onboarding page. What do you think I’ve got to do here to make it good?

Gonto: I don’t think that people understand how they can donate their spare tokens and what that means. Like, is it part of their subscription? Is it part of their API usage? I would explain that a bit more.

Hank: Yeah. So I need to explain really clearly that if you have a Claude Code subscription, this will cost you nothing, because it uses unused tokens from your session.

Gonto: Exactly. That’s the main thing I would focus on, and that it’s not about the API, it’s about using the subscription. And then, because of that, I like your how-it-works, that it’s step by step. I would just explain a bit more that it costs nothing, and I would give them some strategies on when to do it. Like, for example, the idea to me is, if I have tokens left, like I only have four hours, that’s the time I would give my tokens. Or one day, or whatever it is. That’s how I would think about it.

Hank: And the way it actually works is when you install the MCP, and I should put a GIF here, or a recording of me onboarding, when you install the MCP, it first asks you, okay, is there a cause that you care about, that you want to specifically research? Could be poverty, health, pets, whatever. And then it asks you how often you want to do this. Weekly, like on Sunday nights? I’ve got plenty of tokens. Even though I just showed you a bunch of stuff I’m working on and I’m building this MCP, I’m on a Max plan and I still can’t use all my tokens. I don’t know how these guys are token-maxing. But I do it every hour, which you can only do 15 times a day on a Claude scheduler. And then next time you log into Claude, it just sets up the scheduler for you, and you don’t actually have to think about it. But the interesting thing here, I think, for our audience is, hey, if Hank can build an MCP that automates using other people’s tokens to do charitable research, just imagine what you could do. Imagine what you could do in your company to activate your squad, to share skills and workflows. I don’t know. I think we’re really just at the tip of the iceberg of learning how AI can affect our work.

Gonto: I agree. Coding is fun again. Like, I do apps for everything now. That’s cool. I like that you’re doing some philanthropy besides all of your work projects. I feel you’re working more and more every day.

Hank: I am. I mean, I remember telling you, I’m not going to work that hard this year, and then there’s just too much cool stuff going on right now. I can’t help it.

Gonto: I agree. We have one last topic for today, and this time it’s a tweet from Chris Sev, who used to be the head of DevRel for Laravel with Hank, but now he leads DevRel at Auth0, actually. And his tweet, we’re showing it now, has a GIF where he shows that he changed his title from Director of DevRel to Director of AgentRel, because I think he focuses on agents. What’s your take? Do you agree? Do you think the future of DevRel is AgentRel or not?

Hank: I mean, I don’t know that I like the term. DevRel is a cooler term. I think he’s right, though. At the Code to Market Summit, this guy Dave Kiss from Mux came, and he had a really good little bit that we all talked about, how he’s going through an identity crisis, because he writes a lot. He really finds a lot of joy in writing. And he’s like, but I’m not writing for humans so much anymore. A lot of what I have to think about now is, what does the agent read? We talked about this on the last podcast, with the llms.txt and the agent tools you give on your website. So there is really something to this, that DevRel is more and more about agent experience. AX is another term coming up, and maybe it is about AgentRel. But you don’t like the term, or you don’t like the concept?

Gonto: I just disagree completely with the concept. When I think of developer marketing, I think developer marketing is two things. It’s DevRel and DX, or developer experience. The DX side, which used to be writing better docs, having better API guidance, doing the SDKs, I think all of that part, that DX, is dead. All of that becomes AX, where maybe you don’t need SDKs anymore, maybe you just need skills and specific prompts and the MCP and how to connect. So on all of that, I agree that the DX is gone and it’s now AX. But DevRel, which was all of the other side, relating to developers, I actually think it’s even more important now. And I see this all the time in SF, where before, humans would talk to humans or read stuff from humans all the time. Now the web and everything is so full of AI things that I think humans want more human contact and human touch. So I think it’s even more important that DevRel is human than before. I think they should focus more on in-person events with other developers, doing a drinkup or a meetup or something like that, or they should do more side events at specific conferences. I think they should focus on building human videos for YouTube or for Reels or for TikTok, where a human actually sees a human. Where they are tweeting themselves as a human, because people can notice when a tweet is probably a human or not. So I actually think that DevRel should be more and more human now, because that’s what humans want as all of the web and all of the digital experiences are becoming agent-focused.

Hank: I think I agree with pretty much everything you said, and especially on the events. Events are getting so critical. We probably won’t show the tweet, but I saw swyx make a comment somewhere about how, man, a year ago he wasn’t sure if he could keep doing the AI event circuit, and now it’s just blowing up. It wasn’t in dire straits a year ago, he just wasn’t sure about it, and now it’s everywhere, and people want to go, and they’re getting record attendance, and everybody’s putting on conferences. We might need to do an episode just on: so your CEO asked you to put on a conference. Because that is a common theme among my clients right now.

Gonto: Mine too. And it’s either that, or side events, or owned events. I think everybody’s moving from events where they go and have a booth to owned events. And those could be conferences, they could be something else. We can talk about it in another episode. But I think on this, just to close it out, Chris is completely wrong, and if he thinks it’s going to be AgentRel, he’ll be fired.

Hank: Or did they just give him a promo? Did they just ask him to retitle to that? Who knows. Curious.

Gonto: I know his boss. His boss is Ari Shapiro. He’s Argentinian. So I can ask his boss if it was him.

Hank: There you go. Well, I’m sure Chris will answer in the comments. And if he doesn’t, tag him. Everybody get in the comments and tag Chris.

Gonto: Okay, great. Thank you for coming to today’s episode. The first half was Hank sponsoring his stuff, but we’ll come back with more meaty stuff soon. Thank you.