Squandering a Billionaire's Attention

There's so many things I don't like about their like Twitter profile and stuff, but like they got the important thing, right? The product is good. They broke a pattern. They got engagement. Nice. Good job. Taalas, Taalas, Taalas, whatever you are.

Yeah. I don't know the name, but I agree.

Today on Code to Market, we've got a crappy launch that did well, and we've got a whole bunch of reactions to downtime and how different companies approach it. And then we also have, you know, a nice guy and a mean guy fighting on Twitter. But yeah, Gonto, you've got the first one.

Yeah. Let's start with the launch. Like I actually tweeted about it. I was blown away. Like there's a company that actually even the name of the company sucks because it's Taalas, but with double A. I don't even know how to pronounce it. It's like Taalas or I don't know. But they basically spent 30 million in developing something that allows you to put an LLM weight in a physical hardware. And what that means is that this physical hardware is so much faster at replying and responding to chats or whatever you send. So they actually had a demo that's called Chat Jimmy that if you tried, it's just mind-blowing. Like you send a message and it responds almost immediately, whereas most of the other ChatGPT club take forever. Having said that, as we're talking, like if you're in YouTube or in Spotify with video, you can see this tweet. And it's just 24 dedicated people, 30 million spend. This is what we ship. And if you look at it, like there's no image on the post. There's no Twitter open graph. There's nothing. There's a link that says details that you can't see anything else. There's a thing that says demo ChatBot and that's it. But that was the entire tweet. And I felt it was a really, really bad launch. But they still were so popular. What's your take, Hank?

I mean, yeah, their launch, it's one of those things where it's like, yeah, they did multiple things I would have advised against. But I guess what do I know? Because they got more impressions and more engagement on this than I think anything I've, you know, like had my hands on directly. And they've got like no following. So, you know, it's hard to know what's going on here. Like there's something about people just have to engage with it because it does check a lot of boxes in the bundling. We've talked about bundling before and they've got the whole like, okay, this is kind of a, it's not quite funding, but they've got the like funding style, like 30 million and people and like a launch thing bundled with it. And then what's weird is they have three links, which you're not supposed to put three links on a tweet because it does what this did. Like it's got a broken preview image. And the main link that Twitter's pushing you toward is the API request form, which looks weird. And then the Chat Jimmy thing. I don't know. So I guess before we get to that, like that's all weird, but then people just wanted to engage with it. You wanted to retweet it. I commented on your tweet after trying it. They've got some like heavy hitters in the replies, almost 500 replies, almost 4 million views, almost 6k likes. Pretty crazy.

I 100% agree. And to me, it's like they had so many opportunities. Like the chatbot was so fast. They could just have done like a GIF where you show the message and the reply. Or they could have shown a picture. Like the blog post had a really good benchmark on how much faster it was and how much better it was. Like they could just put that picture instead of something else. So I think the opportunity was so big on things that they could do.

I mean, the traditional thing is you at least put in like what's interesting or valuable about this. I guess they put extreme specialization speed and power efficiency, but it's really easy to like banner blindness style skip over that line. Because the formatting of the tweet is even weird. Maybe that's the key here. It's a pattern breaker and it kind of catches your eye. Like what's this hot mess?

I remember why I saw it. I saw it because I saw somebody quote tweeted saying like, check the demo. It's incredible or something like that. So then I clicked on the link. And then at first I didn't understand what the demo was. But once I saw the demo, it was like, holy shit, like this demo is really good. But I think in this, what it also shows you is like one, breaking the pattern sometimes helps. Even when it looks weird or it looks shitty. I think the product was so good that people were quote tweeting, trying to explain why you should check it out. Because in other cases, people just RT. But in this case, I didn't see anybody RT. Everybody was quote tweeting. If you look at the stats on it, you should look at the demo. You should do this. You should do that. And I think that that also creates more engagement. So even though I don't think it was their objective, they drove so much engagement through that. And if the product or what you're launching is incredible or unique, you'll still get the views. Even if you need a shitty launch. That's my other takeaway.

Yeah, it's so interesting. I mean, their profile. There's so many things I don't like about their Twitter profile and stuff. But they got the important thing right. The product is good. They broke a pattern. They got engagement. The thing does, you know, the label on the tin is what's inside the tin. Nice. Good job. Taalas. Taalas. Taalas. Whatever you are.

Yeah. I don't know the name, but I agree. Let's talk a bit about HubCode. I only saw Dharmesh's new tweet, but I didn't see the one that he deleted. So what did the deleted one say? Like, what was it?

Yeah. So Dharmesh, he's the co-founder and CTO, I think, of HubSpot. And he's very AI forward. I think, actually, he famously got equity in OpenAI because he owned chat.com or something. And he didn't sell the domain, but he traded it for equity or something. Anyways, he put up, like, an image that said, like, OpenCode. Or no, it's a fork of OpenCode. And do you know what OpenCode is?

I do. Yes. Can you explain it for the audience? Yeah. For our audience, OpenCode is an opinionated version of Claude Code that is multi-model. So you can hook up any model and they have their own opinions. It's open source and it's built by DAX.

Yes. So Dharmesh launched HubCode. And it's basically a fork of OpenCode, is my understanding. He did give attribution, but it was very, it was a very long tweet. And it was, like, buried in there. Like, you could basically miss it. Or it might have even been threaded. I don't know. Or it might have just been on the, I don't know if he even had a site. He deleted the tweet because DAX and the team, you know, started railing on him. On, like, oh, like, you know, nice product. Where'd you get it? And, you know, they're kind of, uh, kind of razzing him about it, which I'll circle back on. So Dharmesh, he's a nice guy. Like, he deleted that tweet. And this wasn't even a launch. This was like, hey, I'm working on this. And so he deleted that tweet and had this, like, big explainer, gave credit, like, and basically, like, owned up to a few things that he feels he did wrong with his tweet. So a couple things to digest here. The first thing, my first thought when I saw the original tweet and I saw DAX and his crew kind of take it to Dharmesh is I was like, okay, they clearly don't know Dharmesh or his brand. He's like a nice guy. He would never, like, harm intentionally.

Like, what was the problem? He didn't talk about OpenCode? He didn't say that it was like a fork or what was it?

Basically. Yeah. I think they were upset because he wasn't making it clear that it was a fork. Like, that wasn't like the first thing he said. It was buried. And then also like the design, cause it's kind of got the like ASCII art style in the same way that like, you know, Claude, it spells out Claude with the big blocks in ASCII, you know, OpenCode does that. And he just changed the open to hub. So the design looks similar too, but it's, you know, it's a TUI, it's a TUI. Like there's not too much room for design. I'm, you know, maybe I'll get flack for that, but yeah. And it's like early and you could argue that all these, all these like TUI interfaces look the same anyways, and their designs. Anyways, my thought was, man, if these guys understood who Dharmesh was, they'd be looking to like partner and help. Like this is HubSpot, big public company, and this is billionaire Dharmesh. Like this is a guy who could superpower OpenCode where they could say like, hey, like, what are your needs? Whatever. So that was my first thought when I saw it is if you're an open source maintainer, someone's forking you, which also if you're open sourcing something, yeah, people are well within their right to do that. And they can tweet and whatever.

But if they say that's the idea, that's the idea.

Otherwise don't open source it. But if someone does it, there's a need of theirs not being met, which you might be able to meet and form a bigger partnership or customer base or serve a market. You didn't understand. So that was, that was like a miss I thought on their part, which that door is still probably open. And then Dharmesh did a mea culpa. Well, sorry, go on, go on.

No. What I was going to say before the mea culpa is like, I think they like OpenCode blew up because multiple people that are very well known on Twitter talked about OpenCode. Like DHH has been using OpenCode. He added it to Omakub and he's been saying it's the best way to use all of the models for coding. Karpathy talked about OpenCode. And there were like a lot of very known people talking about OpenCode, which is what made them popular. But it made them popular because they were using basically their product, either the open source or the paid version. In this case, they just made a fork and it could have been great publicity, as you said, if they engaged with Dharmesh. But instead of doing that, they just fucked up the relationship. And I think Dharmesh is pretty well known in the engineering world. I think he's better known on LinkedIn than on Twitter because he's more enterprise. So I think it was a fantastic opportunity for OpenCode to actually get access to more enterprise developers. Like DHH, Karpathy have more hipster developers, more of these startup developers always on the edge. And Dharmesh could have been a LinkedIn developer influencer that talks about them, that thanks them. So they could have done, they could have won such a big audience by just helping out. And I think they were just pissed off because it's like we've been open source, but we don't want anybody to fork it where it's just open source because we want to say that it's open source. That's it.

Yeah. Now in his mea culpa, I agree with everything you said there. In his mea culpa, Dharmesh said, okay, well, per my LinkedIn discussion, which is, which is apparently where the productive discussion happened. He's like, I'm not sure I should go the TUI route. And if I do go the TUI route and I also use OpenCode. So now he's saying, I might not even use this after all, then I'll give proper attribution, you know, sorry, my bad, which to me is signaling like, hey, you guys kind of chased me away. Like I got a bunch of love over here from like, and potential customers in partnership. But then here you guys basically chased me out of this. Like I'm not interested. And that feels like the biggest miss for OpenCode here is.

Uh, yeah. I agree. 100%. Like that's mind blown that they fucked it up and weird for open source. But I don't know if they were, if it was like, I think it was honestly, they didn't think about it. They just reacted on it. They forked OpenCode and they didn't talk about it and they were pissed off and they just reacted instead of thinking a bit more about it.

And doesn't this happen all the time with like the founders we work with? Like they react instead of trying to think of what's the opportunity. And what's interesting is especially like most people have good intentions and most people are doing the best they can given their point of view. So Dharmesh's point of view, whatever, hey, an open source thing. I'm going to play with it. I'll just do a couple of tweaks in a social post. If you take that assumption and your reaction, like this is the lesson for, for the audience. Like, okay, if somebody does something with your tech, they don't attribute you, whatever. And you're feeling like your toes are getting stepped on that person probably didn't mean to step on your toes. And therefore like you have an advantage because you can approach them and say like, hey, you probably didn't mean to step on my toes, but maybe instead of forking our tech, you'd love us to work with you and deliver exactly what you need and use our resources for your advantage. And that would change the game probably for Dharmesh because he's, I don't know. Yeah. It's just so common with the people we work with to be reactive and defensive and almost, I don't know. Gatekeepy is maybe the wrong word here, but you know what I mean? So next time you're on Twitter and you see something that pisses you off, just think, just think of a couple extra beats on what's the opportunity here.

And I think asking your founder to wait until the next day is too much because maybe it has already happened. And it like, I tried that in the past, but in Twitter things move so fast. So to me, it's like, don't wait until the other day because that's too much of us, but just go and walk four blocks. Like that's it.

I don't, I don't think it's even that. I think it's no, you should reply, but you have to change your frame. Rauchg has actually gotten fairly good at this. Like he looks for opportunities or feedback more reflexively now. Cause you know, the instinct on a lot of these tweets that come at Rauchg are like, they actually are coming sometimes from a place of, you know, hate and his reaction to them has, has been pretty good. You know, unless Cloudflare is involved, then all bets are off and he gets mad, which is a good segue to our next topic. Cloudflare took a ton of people out. It took out Laravel Cloud. It took out Clerk.

Clerk in this case, this time it wasn't Cloudflare. It wasn't in the past though.

It wasn't Cloudflare this time. Well then great timing because everybody, from my perspective, everybody just thinks it was Cloudflare. And so they're not getting the blame that maybe they should, maybe we should shut up, but.

No. And I think it's not that, but it's like, I think in this new period of downtimes, it's extremely important to be transparent, but take ownership. Like, as you said, like at least for Clerk, it has happened this last year. I think Clerk went down multiple times out of all of the times that went down one, which is this recent one was their fault. But all of the previous one, it was either something that Google Cloud did or something that Cloudflare did or something like that. But I think what Clerk did right is even though they did go down, and even though it was because of an upstream vendor, they still took the ownership and that's what they have to do. Because even if it's Cloudflare that went down and that fucked it up, as a company, you could have multi-cloud. You could have like different clouds. You could have different regions. You could have a lot of things that could help you. But I think in this new world of downtime that is coming from so many different things and so many services that we use that are interconnected, where if one service goes down, the other one goes down and stuff like that, I think taking ownership on, we fucked up, we could have done differently. Being transparent and open into what happened and why, which step by step on it, at what time did we notify that it went down? What did we do in the meantime? What are the things that we learned? And then the third one is, what are the things that we're going to do to fix this for next time? And I think then if something does happen again, where you fuck up, you should be able to at least say like, look, this time, what we fixed last time works, but it's something else because showing what you will do in the future to fix this gives you like, gives accountability on the sense that if it goes down again for the same thing, people should switch services. Like how come you go down for the same thing twice? Like if you go down for different reasons and you explained it and you're working on it and that's your priority, maybe it's okay. Like it still sucks, but I think it's a better optic. And I think it makes more sense in that, in that way.

I think that's true. Yeah. And yeah, like when it's a provider's fault, like it sucks, it's still good to do something nice. At Laravel, you know, the team decided to give everybody like a 10% credit, which is, you know, which is nice. You know, you don't have to give people credit when, well, you don't have to give people credit ever necessarily, unless it's in your contract for uptime, but especially when it's an upstream provider. But yeah, like, you know, now the team's definitely got to talk about, all right, we need a backup network. And yeah, it's kind of, it's kind of weird. Those are bad days, but these have always happened. What's interesting is they just, they just get more and more impactful. But I still remember in like 2017, like an internet backbone thing took out like half the internet in the US for like six hours or something.

I remember that. Um, but, but, but in some of those cases still some companies, I remember like, I remember Mercado Libre was a company here and they went down and they blamed it on Amazon because a whole region, which was AWS East went down, but you should use multi-region. And if you didn't do multi-region, then it's your fault. It's not Amazon. So I think, I don't know, you always take ownership. Like it's always, always your fault because there's always a better way.

Yeah, it's true. If you're not multi-region, you need to go multi-region or.

Or multi-cloud.

And if you're not multi-cloud, you got to go multi-cloud. You got to have backups. Cause even though these things have always happened, like people's businesses are more and more online every year. And more. Exactly. So I don't know. Yeah. So there were some good examples of like what to do, how to apologize to customers. And I feel like a lot of the time people are pretty understanding, but yeah. Interesting.

I agree.

Yeah. Don't have much else to say on that. Yeah. We got anything else. Are we done here?

I think we're done. Happy to be back recording and doing episodes. And you'll hear from us pretty soon.

You say that every week. I feel like whenever you say guys we're back, we're doing it weekly now guaranteed. That's when we're going to miss the next week. So I don't know guys, we'll, we'll see. How about, how about you guys subscribe and like, give us a like or a comment or something. And then we'll see if we do another episode, you know, give us a little engagement, you know, make it, make the juice worth the squeeze.

I like it. We need to get more subscribers. So subscribe. Thank you. And peace.

Send this episode to someone who reacts poorly to others on Twitter. That's it. There you go. All right. Thanks everybody.