Drowning in DevRel Demand

Shout out to Framework for having the Twitter handle Framework Puter, which is just a funnier way of saying computer. That cracked me up when I saw that.

Oh, when I read Puter, I thought of golf, the putter instead of computer. But yeah, exactly. So probably something very different. You're right. Welcome again to Code to Market. I know we've been very inconsistent lately, and we're sorry we've been drowning work and having a baby.

But what I can promise you is that we're busy, okay? You know, we got lives. No apologies. Sometimes you got to disappear for a while to be wanted again. So listen up, people.

We are not sorry, but we're so back. We're 996-ing this and locking in on the podcast to do this again. So you'll hear us from us again probably weekly, I would say. So we're not sorry, but we're coming back. With that, we have two or three topics today, depending on how we do with time. And we'll start with the first one. I don't know if you were on Twitter this weekend, but there's Ayla Croft, who is just a developer. She has 4,000 followers, who literally tweeted something that we're showing now if you're in YouTube, that says, Hey, Framework, if this post gets 10k likes, can you guys send me a semi-beast mode DIY setup? Framework is a company that sells computers that are like Mac, but for Linux and Windows. And they look awesome. And she basically said that she's broke. She has no money, but wants one. And Framework literally answered like, Sure, if you get 10k likes, we'll do it. What's the likelihood of somebody with 4k followers getting 10k likes? Very, very low. But she nailed it. If we look at the stats, she got 12k likes, 1.2 million views, 557 retweets and quotes, and 278 replies. I loved this tactic from Framework. But I'd love to hear from you firsthand. Like, what do you think about it? What will you do differently, similarly?

It's such a great lesson in just engage. When people tag you on something, just answer them. And, you know, answer in the most generous way if you can. Like, the cost of this for them, one computer. If she hits this crazy goal. The impressions they got, the press they got, and the types of people who then are suddenly like, Yes, let's make this company give her a computer. We're going to retweet. We're going to quote tweet. I think he told me ThePrimeagen, like, quote tweeted it several times. DHH quote tweeted it. You know, people start pushing your product for you just because you replied in the generous way. Such a simple lesson.

100%. And to me, what's fascinating about it is DHH has already been pushing from Framework since they have a partnership with Omakub, like his Linux version. So he was already talking about it. And I think this is a great way also for him to highlight Omakub. So what I love about it is, as you said, people engaged and then people just started retweeting. The ThePrimeagen did five retweets, which to me is insane. Toby retweeted DHH, but they had so many people. And in the end, they hit the objective. Framework also shared that they moved past 100,000 followers thanks to this post. That they increased so many followers this last week. Just think of this post. And people were trying to come and help and be there. And what I just said on the thread that I really liked is it worked for GameStop. The community always comes through. That's what she tweeted before she knew that this was going to work out. And it did. And then what I think is fantastic is because she got help from DHH, the ThePrimeagen, and basically everybody who is speaking at the Omakub conference from DHH, she also tweeted a couple of days ago that she first installed Omakub on the Framework computer that she will get. So even better, because DHH and ThePrimeagen just did it for help and they got extra publicity. The last thing to share on this, and I love this, Framework is not only giving her the computer, they're actually flying her to their company to show how they assemble the computer and show them the office and stuff like that so that she tweets about it. Because they know that they tweet from, with a video or something like that, or tuning the offices, showing the team and stuff like that, will probably get a lot of attention as well. Yeah, it's content now. So it's a second chance for them to do it. Exactly.

Yeah, they can just ride this wave until it crests and gently crashes into the sand. Beautiful stuff. Also, shout out to Framework for having the Twitter handle Frameworkputer, which is just a funnier way of saying computer. That cracked me up when I saw that.

Oh, when I read Puter, I thought of golf, the putter instead of computer. But yeah, exactly. So probably something very different. You're right.

Yeah, that's great. Next topic.

Yeah, let's move to the next topic. Talk about it. This one is a bit longer.

All right. So a year ago, Swix wrote how DevRel is dying or something. He might have written that two years ago. Whatever. A while ago, Swix wrote about how DevRel is dead. And now he's writing about how it's still back. And I actually think it never went anywhere. I do think, yeah, there's like trends in the interest. But I think DevRel for the last year has been critically important. But we can talk more about that. I don't know. Do you think DevRel did die out for a little bit? Is it coming back in the same way? Is it coming back in a new way? I want your takes on that first. And let's just talk about DevRel.

Yeah, I don't think it's ever disappeared. And if you look at, like if you read Swix, if you read John's post, like what Swix talks about is that people stopped talking about bottoms up and it's something that they stopped caring about. And I don't think that's true. Like I know most companies were still caring a lot about bottoms up. What I do think is that DevRel changed. Like in the past, you had infinite money. So you could hire an infinite amount of DevRel people. And back then, DevRel would just write content and do things without linking it to revenue, without linking it to specific KPIs. And that, I do agree, like stopped. At the same time, I think DevRel has evolved. When Swix wrote that blog post, DevRel was basically doing blog posts and creating content that you own so that people came. That I think was the first wave. Second wave, during the time that Swix said that DevRel was dead, I think was sponsoring YouTubers, was about sponsoring people talking on Twitter and stuff like that. So the thing with that is that it seems that's not done by DevRel and that's more done by growth. And it seems that instead of having DevRel, people only care about these influencers and these community people on YouTube and stuff like that. And the reality is that DevRel was actually working out with them, but behind the hood. So you didn't see them as much. And I think the third wave started with Lee in Vercel and then eventually now moving to Cursor. What's your take?

I think there are always these ebbs and flows with this type of role when it comes to what type of content is the most important to have and whether it's easier for you to, I mean, it always starts out easier to buy content. So I think we're in this transition right now of everybody's sponsoring YouTubers and people with big Twitter followings. And as those line items start to add up and finance starts to ask questions, and as you get tired of just continuously giving these external people money, you always have to ask yourself, well, should we create our own channels like this? Like this happens with everything. Like often when the company starts a blog, they might hire some external writers to write on that blog or something. And then eventually you go, man, it'd make a lot more sense if I just hired someone full-time to do this. I think DevRel often follows that pattern. You do some experiments, you see what channels and types of content works. Maybe it's templates and demos, maybe it's videos, maybe it's written stuff. But eventually you do want someone in your company to be creating the content. And we have this new wave of importance on YouTube and even short form content, TikTok and Twitter and Instagram. And I think there are now the types of people you can hire to make that content and to have your own media channels.

And I think besides that, in the beginning, like I started doing YouTube sponsorships maybe three years ago. When we started doing YouTube sponsorships, for example, with Clerk, there was nobody doing YouTube sponsorships. Same with Vercel. They were one of the first. So then the deals were cheap and most of the influencers were not sponsored. So everything seemed organic. Like I remember we did a full video on Clerk didn't work. So we asked them to just build an app and use Clerk without mentioning it. And that still works a bit. But now I would say most developers know that absolutely anything that an influencer is using in YouTube is paid. And to me, what's fascinating about it is even when they are not being paid, like Theo fanaticism for Vercel, people now think he's being paid even though he's not. Because we move to the other side. We move to the side where everybody thinks, oh, influencers don't have real opinions anymore and they're just getting money. And that's, I think, the point where Sean was talking and thinking about DevRel is dead. It's this idea of influencers are now fake because it's all paid and it's all like that. Well, because of that, I think we switched.

And when he originally wrote, you know, you got to remember with the end of Zerp and the end of free money, there's a lot more scrutiny on these type of roles that don't have direct returns. And this is a thing when I had DevRel under me at Vercel, I would have to defend it. You know, people would ask for what are their metrics of success and say, well, those need to remain a little nebulous. It needs to be more, you know, more focused on content and that sort of thing. And it's also not the type of content that's going to directly drive even Next.js downloads, but especially not Vercel sales in directly attributable ways. Now, of course, some of that is directly attributable. Like if you build the templates marketplace and you have like templates with one click deploys to your product, you can get some attribution there. But, you know, for example, right now at Laravel, if we make a video just about some cool framework feature, that's important to have. It'll get 10,000 views, which is kind of a vanity metric, but I don't know how those views necessarily translate into adoption and such. It's just nice to have in the community. And so there is this ebb and flow of how cheap is money and how tolerable is it for a company to have someone who doesn't directly tie to revenue outcomes.

At the same time, I think I've tried to measure those things. Like we do like two or three standard deviation signups to see if something is happening. We try to see how many people click on the links on YouTube and think like, okay, if they got this number of views, compare it with others. So you can do some magic, I would say, on top of it, but still not the accurate number. And to most people, that is voodoo magic.

You and I have talked about attribution before, and I still think one day we need to have a deep dive on attribution, like an episode on that, because to most people, attribution is so hard. It's hard enough, I think for us, and we like obsess over this stuff, but for the people who haven't really obsessed over this, it is magic. Like you don't know how to translate that stuff. And so it's hard to justify.

I agree. And what I was going to say is I think that now that everybody thinks that all of the influencers are bought and all of their opinions are bought, then why not hear influencers who work for a company? Because at least for those folks, I already know they are valuable for their platform. Yeah, exactly. But maybe they use other tools that is not their platform and they would love them. So what I mean by that is, if they already have a salary from somebody and they're promoting that thing, I assume that the other tools and platforms they use are not promoted. It's because they really value them. And that's not what you believe for influencers. That's why I think DevRel is back. DevRel is back because now people care about the other opinions from the people who are working. I think a great example of this is Josh. Like Josh, for example, who was a YouTuber, works as head of DevRel full-time at Upstash and has a lot of videos at Upstash. And everybody knows he's an Upstash guy. But regardless, then he has other opinions about other things. And given that everybody thinks that everything is paid, there's only one thing that matters now and that's attention. Attention is all you need. So if attention is all you need, that means that it matters more if you're showing up on Twitter, if you're replying, if your video is becoming trendy, if your YouTube short is working more than anything else. And we also see more and more companies thinking about UGC, like user-generated content on TikTok and similar, because more and more developers are 20 to 25, 20 to 28, and the 20 to 28 folks use TikTok. They do not use a lot of the other platforms.

Indeed. Yeah. Good takes. I'd love to hear, you know, if you're listening, Sean, we'd love to take, you know, what do we get wrong or right about what you've said? What are we missing here? What's, what are the elements that you understand that maybe we and our audience don't?

And we have a third topic for today. Do you want to present it? Yeah. Also, targeted to you.

Also SWIX, AKA Sean related. So Sean sent me a message, you know, asking what I think of the strategy of launch everything twice. He said, so, and Google does this in particular, they launch a preview of something and then the preview is the real launch. And then the GA comes and it's, the GA launch is kind of a flop. And that is a phenomenon, I think with AI, because AI companies, you want to be first to move, right? You want to get as soon as, like, as soon as we saw Claude Sonnet 4.5 come out, it was two days later, cursor and cognition. And like, everybody's announcing, all right, we've adapted to 4.5. Here it is. Like, and whether it's GA or not, they're just pushing it out because they don't want to look like they're being left behind. And then the problem is, well, when it's actually stable and ready in GA, the wind's already been sucked out of that launch because the people, the early adopters already used it and the people not on Twitter, like didn't care in the first place, maybe. So there's this question of, can you do a preview or a public beta launch well and a GA launch of the same thing well later? And my initial thought is, it's best if you hold something back. It's best if there's something in the GA that's like a sweet surprise that you held back from the masses in your first thing. But maybe that's not possible if you're trying to be first on everything. Like, then the alternative is, well, just don't try hard on the GA. Just acknowledge that your preview version is the launch and the GA is just kind of a changelog update. But I don't know. What are your takes on this?

I agree. Like, what I would say is, GA launch for AI platforms don't matter anymore. Like, the main thing that matters is, I'm the first model. Like, I love this meme on like, Anthropic just announced the best model ever. Then it's Google. Then it's OpenAI. Like, it's literally going through that. And I think everybody pays attention on the preview. And for the preview, they do the work of talking to the influencers, giving them access before. So then when they launch the preview, everybody already is connected to them and you can already try it out. And that's when it's new. So, it's either you wait for GA or you just ship preview. And my take is, you just ship preview. And typically, I imagine you don't ship GA because when you ship preview, you start seeing what happens in real life. Based on what happens in real life, you build evals and then you improve your product. And then once it's at the better place, you ship GA. So I think for these things, you cannot ship GA mostly because you're improving it until it becomes GA. And that level of improvement requires a mass audience. And if you require a mass audience first to improve it, then there's nobody gives a shit about GA and I would just not announce it. Like, I would just ship it, do a tweet, and that's it. I would forget about it.

I think for products, so here's where I 100% agree with you. If it's a product where you're reliant on someone else, so like if your product relies on, say, Anthropics' latest model, like, and you just got to ship to get out. Yeah, get as much out as quickly as possible. Or if there's some other, like, external force or very competitive nature of like, yes, all of our beta stuff needs to get out as quickly as possible. If you don't have that, then I would never do a public preview. I would only do private previews. And like, I have the luxury of doing this at Laravel. You know, it's all, you know, we cannot just do private previews. We can let hundreds of people into those and they feel special because it's private. And then when we GA it, everybody else is very happy. The GA and we've already worked through most of the bugs and stuff.

But I think you can do that because your GA doesn't require a ton of traffic, a ton of people using it and stuff like that. I don't think that's true for the AI models.

I agree.

I don't think they can do that.

Yeah, exactly. I think in which case, I would say the GA, unless you have something that you've held back or something fancy to add, don't even worry about it. It's a change log more than anything else. The preview is the launch. The GA is just an update.

I agree. wow to show. Maybe you show it with videos and stuff like that, but you don't let people use it. And then once people start using it, it will become trendy again because it becomes trendy through crazy usages similar to what Sora did with the app and the cameos. Yeah. So that's the only other way. But I don't know that. I think just ignore GA.

Yeah. Interesting. Well, there you go. Those were our three topics today.

Thank you, all of y'all, for listening. We're 20-something minutes, so we're right on track. As we said, we're back on schedule and you're going to start hearing from us a bit more often. So sorry that we had a lot of episodes that we missed. Thank you.

Bye.