Gonto: Why do you always have to mention LinkedIn in every podcast?

Hank: LinkedIn is great because you're underestimating the power of LinkedIn. 

Okay, Cloudflare fired the first shot or at least the first public shot. Perplexity is firing back, Browserbase is getting in the mix.

This is an interesting development. We're kicking off with follow-up from our topic from our last episode on how Cloudflare called out Perplexity for stealth scraping, they called it. They put Perplexity in a really bad light in order to put themselves in a good light as defenders of their customers' websites, making sure that bots weren't scraping immorally.

Perplexity fired back and said, hey, these aren't bots, these are agents. It's a new, different thing. We have to think about them differently.

And then Browserbase put themselves in the middle. What do you think of all this? What are your takes?

What are the points I'm missing?

Gonto: The first thing I'm proud of is that we found this fight before it started. It was appealing to us last week before it actually exploded. So I'm proud of us actually finding an interesting topic on Twitter.

After the humble brag, a couple of things on it. I think Perplexity did a really good job on relying on this idea of agents are based on a user command. A bot that is scrapping the web is not based on a user command.

So anything that's based on a user command, you should allow us to do, the others you shouldn't. And what they were saying is that they were doing those fake things because otherwise they would get locked. I'm pretty sure that Perplexity also uses this tactic for the bots as well.

So what I'm saying is, I think the reply is really good. I probably don't think it's entirely true. And I think most people will see through it as well.

So if I would have been Perplexity, I would actually take a bit of mea culpa because just on how it looks like, it's obvious that they're not being honest. Like I would do something like, hey, it's true. We've used it for bots and that will stop.

But we also used it for agents, which are based on user commands for comments and stuff like that. And for those, we think we need to keep on doing it. But they actually fucked it up, I think, on that.

And I think most people that I talked to about it didn't believe them. So sometimes taking a bit of mea culpa, but then fighting back, I think is a better answer. What was your take on the Perplexity reply?

Hank: I think you're right on that, especially because some of the replies we're seeing, like there was the reply about Garry Tan and Y Combinator traffic being blocked because Cloudflare wants to monetize the traffic. And people don't trust Cloudflare. Cloudflare has this reputation of being very obstinate, very self-righteous, and their CEO never apologizes.

So Perplexity had an opportunity to counter position and be like, hey, here's the nuanced take. They tried to have the nuanced take, but they didn't have the nuanced morality take as well, which would have given them some interesting high ground and won them some favors. And by the way, people are already suspicious of them because their CEO has been very transparent about like, we're going to have ads and all sorts, like we're going to monetize the crap out of Perplexity.

People are like, whoa, that's old news now. But I agree with you, they missed that opportunity for counter positioning in a really good way here.

Gonto: And I think what's interesting is that, as you said, there was a lot of pushback towards Cloudflare. Like I saw the tweet from Suhail saying like he never says, sorry, as you said, the Garry Tan. And I saw others as well saying like Cloudflare by default now is stopping to show our content.

So it's interesting that it's so clear for most people in tech, how Cloudflare is doing it, not in the benefit of their customers as the article first themes, but rather for their own benefit. And they fucked it up because it could have been a great marketing, but by disallowing AI bots to crawl by default all websites, I think they fucked it up. They went too extreme instead of giving people a way to opt in.

So it's interesting because like two monsters that are fighting, and you mentioned Browserbase. I think Browserbase did a fantastic thing where they injected themselves into a fight that they had nothing to do with, but because they are proud of it and they are big, like they look bigger than before because they are fighting with giants.

Hank: Well, you might have missed it, but in the Twitter article that Perplexity posted, they did mention Browserbase and they said, oh, we think Cloudflare is confusing some of our stuff for Browserbase…

Gonto: Oh, I didn’t see that.

Hank: Which has some bot-like behavior and some agentic-like behavior. And by the way, quick aside, I thought it was clever of, because this fight is taking place on Twitter, Perplexity did a Twitter article, which is really good. Like you're meeting the user, so it's more likely that people will read deeper into Perplexity's side because they're using the channel where the fight is happening.

Gonto: And it's more than that because links in Twitter get deprioritized, but if you have an internal article, it doesn't get deprioritized. So they chose really well, not just in the same platform, but also on a way so that they could get more virality than Cloudflare.

Hank: Yeah, exactly. And they need that because Cloudflare has the history and the clout, like they're very well established. I just checked, Cloudflare's reach is definitely bigger and more significant and they got more likes and all that stuff.

So Perplexity's kind of the underdog here.

Gonto: But still, both are giants compared to Browserbase. And by then, I didn't know they were in the article, but even if they were, they injected themselves on the fight and they used it also to announce that they are getting into a new space. Like they hired somebody to own identity for agents and they mentioned how they hired that person and the launch for that person was actually writing this blog post together with Paul.

So I think they did a fantastic, really, really good job on injecting it. They had a good argument, they had good thoughts, and they weren't as spicy as Cloudflare or Perplexity. So they tried to have a more neutral way so they don't get that many haters or lovers.

I don't like neutrality as much, but I think they did it on purpose.

Hank: 

Well, and I think it's authentic. I mean, I know Paul of Browserbase. Oh, you do too.

We've got a special episode with him coming up. But I remember I once like pinged Paul, I was like, hey, could Browserbase do this? And I sent him some idea.

He's like, “we try to prevent it being used that way because that is immoral by our standards.” And so I think his nuance take here is very authentic. And I think he has some compass here that's ruling him, which maybe to your point is holding back him being more inflammatory or getting more.

But he's also the underdog here. So just getting the mention and the word out, it's plenty. Of course, his reach on this was far, far less than the others.

But I'm sure there was some spike of interest in Browserbase just based on the Perplexity mentioned.

Gonto: Exactly. At least it was mentioned, it was showing on the Twitter timeline. I'll switch topics a bit.

I read a tweet last week that actually blew my mind. I'm a millennial, so I still use Instagram, but I only use Instagram to see what my friends are doing. I basically use stories.

And what I realized after reading this tweet is that I'm one of the few that actually do that. Like most people use TikTok and they don't see as much their friends. They see more influencers.

And in Instagram, the stat is that 7% of the people see something related to their friends. The other 93% sees nothing. That means that 7% sees friends and potentially also other things.

But that means that people now don't care as much anymore about what their friends are doing and why, which was true before on Instagram, TikTok and others. Now it's all about what are influencers saying. So I think the game of influencers is the number one thing that matters now for any B2B SaaS.

And we'll talk a bit about what OpenAI did there, but what's your take on the influencers game?

Hank: Yeah. I mean, I don't have a stat handy or anything, but I do know that LinkedIn's focus is similarly more and more on content and the algorithms that control that, as opposed to the recruiter and sales experience. They're trying to-

Gonto: Why do you always have to mention LinkedIn in every podcast?

Hank: LinkedIn's great because you're underestimating the power of LinkedIn. I'm going to start bringing LinkedIn links instead of tweets to this. You're going to have to do your research.

Gonto: Don't bring LinkedIn links one month after they happen.

Hank: Fair. This whole thing with influencers, it's something we've talked about at the start of the year because one of our goals, at least one of my goals, was to figure out and start doing more influencer marketing in the B2B space. And it's clear that we're primed for this.

We've got to figure this out. I think we'll do a follow-up on how we're doing on our goals in another week or two. But I wrote down kind of a counterpoint here because you wrote that nobody cares about their friends anymore, which I think is true, but there is a craving for this.

So I know that also Gen Z is leaning far into in-person, both working in-person and in-person events. I just did a huge in-person event. And it's very clear that when people are together, they can actually focus and they're talking and they're not on social media as much.

So there's kind of an interesting thing here of people crave it. It's not how they behave when they're on the social media. They don't seek out their friends, but they still have this craving and this desire for it.

And I wonder if there are ways to tap into it. And I think it's in-person events and dinners and conferences.

Gonto: Yeah. And I do see also private chats. So maybe it's more about private chats online and then offline, it's more conference stuff like that.

That is friends. But all of the rest of social media is not friends at all, which is very interesting.

Hank: That's true. There's the whole concept of dark social. That was kind of a term that was thrown around maybe a year or two.

It hasn't really picked up, but dark social is the idea of like, okay, what are all the Slack workspaces and Discord channels and Telegram groups and WhatsApp groups that people are in, especially as professionals that you can tap into to do some sort of marketing, right? And an example is we have a multi-hundred person unofficial Telegram group for Laracon that's been going for years. So it just kind of reactivates the week before Laracon every time and people get together and we have a huge Discord and all that.

But there's something to that. It's not as visible in these type of stats, but maybe that's where people are going for this type of stuff.

Gonto: Exactly. I have an example on this. In Argentina, people were using Uber and recently there was a new ride app called Didi.

Didi is the Chinese Uber. And what they did to get people was, there's a lot of like WhatsApp groups for Uber drivers where they help each other, they talk, et cetera. And Didi, what they did was once they got one driver in that group, if they helped other people, they will give extra referrals if they would inject Didi into some of those dark groups.

And that's how they grew in Argentina, for example. They leveraged these small groups, at scale, to get people to change, which I think is fascinating as a way to inject yourself. And that's what you need to do. I think when you're the underdog. 

This week, OpenAI also tweeted about a really good video professionally made with seven influencers. And that's the other thing that blew my mind on how good it was.

They invited seven very well-known influencers like Theo, like Sean and a few others. They gave them access to GPT-5 and they asked them like, hey, go try it out. And one of the folks actually tried things that they tried in the past, like Theo, for example, tried the ball game that he always tries every time.

He was like, oh, this is really good. And I personally hated the GPT-5 launch. I was complaining on Twitter about it, but I saw how these six, seven people that were invited, including Theo, were like GPT-5 is the best model.

And I don't think they got paid to say that, but because they were included before, the team helped them, et cetera, they unconsciously owed it to them to be nice and be good at GPT-5. So they are basically getting influencers special access. It's good for them because now they can say, oh, OpenAI gave me special access.

And they basically create this feeling on these folks on, you cannot say bad things about GPT-5. So that to me was a masterful move. Of course, OpenAI can do it because it's OpenAI.

Like I think that's hard for less known startups, for sure.

Hank: True. It's the law of reciprocation, right? It's the reason that salespeople, you know, give you swag.

It's the reason you get early access to stuff and so on, because there is kind of a social implied rule that you're going to say the good things about it. And sometimes NDAs even specify that, that, you know, hey, you can't say anything negative about this. And also when you're OpenAI and you're cutting the video, you can cut it so the good stuff is in there. I think…

Gonto: But it's not only a video. They were all of them. I follow them on Twitter to see what they said.

They all said only good things on Twitter about GPT-5 when there were so many people complaining.

Hank: Yeah. Well, and it's interesting. Let's talk about that aspect of it.

This is less about the influencer game and more about like the product launch game. GPT-5 has been hyped and people saw the difference between 3.5 and 4 and 4 and these other iterative models that really mostly just added breadth. It's really hard to have a big hype launch that's going to just be marginally better.

And OpenAI has known that it's only going to be marginally better. There's nothing like new or revelatory about it. They kind of put themselves in a corner with this, I feel like.

And I don't know if there's a way that they could have gotten out of it or maybe downplayed it or set expectations appropriately. But people who are saying it's bad, they didn't actually mean that it's bad. They just mean, oh, I was hoping for another big leap and there's not a big leap here.

Gonto: Exactly. To me, the big leap was from GPT-4 to O3. O3 was to me a big leap and they didn't make as much noise on that.

GPT-5 was very similar to my usage on O3 and I actually preferred O3 a bit more. But I agree that they hyped it too much knowing that it wasn't going to be there. Maybe they thought they were in the future.

I don't know. But I think they fucked up. You don't create high expectations if you're not going to deliver.

Hank: Yeah. And I'm kind of relating this to myself because at Laravel in several weeks from now, like a couple of months, we're going to release a huge update to our main product, Forge, which we announced at Laracon. And so I've had these thoughts of like, okay, where are people going to be underwhelmed?

And where is the hype appropriate? Because I've got to do this in the campaigns. So it feels very relevant to me and to anyone launching a major update.

Gonto: But that's where, as you work inside the company, it's hard to know if it's going to be really good or not. And that's where I would actually include influencers not one week before, but months before. “Hey, try it out. Give me your honest thoughts. I'm going to try to improve it based on your thoughts, if you can.” 

But I think getting people from the outside who are known influencers from before instead of one week before could even help even more.

I don't know if they get it. Claire tweeted at me the other day saying they got access only five days before.

Hank: Oh, yeah. That's not a lot. And really, was it clear to you if they got access, it looked like they got access while they were sitting on that couch?

Gonto: No, they got it earlier as well. They got it in their account. It's not just that.

Hank: I was like, because when I saw that, I was like, if this is true, if they did just get access right then, that's a disaster. How are they going to create the content? So that's good. Because clearly some of them had put some thought into what they would run and all that.

It's also fun. There's kind of a hack here of when you have multiple influencers at once, they're almost competing to give you the good clips.

They all want to be in the video. So they're all trying to think of what's the cleverest thing I can do? What's the cleverest test I can run?

What's the best thing I can say? So that my face is shown the most in the video because they all want the spotlight. It's part of their job.

Gonto: I agree. Let's switch to our latest small topic. Do you want to talk about it?

Hank: Yeah. Small topic. I saw this tweet.

David Kramer is always throwing out spicy tweets. He's a fun follow. He tweeted out just an email he got.

He's like, this is like the easiest way to get me to unsubscribe. It was about the Gartner Magic Quadrant. You know, Vercel got in it and they sent out an email announcing this.

This is standard, you know, AR analyst relations. It's kind of a part of PR. It's a standard practice in AR.

And I hate it. I always tell my teams like, we got to be very careful on self-congratulatory tweets, emails, anything. Blog posts are always okay because those are very like opt-in.

People have to see it. But for something like this, I would have tucked this at the bottom of a newsletter. This would have been an item number three at the highest.

I also always tell my team there are two laws of email. It has to be valuable and it has to be interesting. This to me, I was just like, oh, that misses the mark on both.

Most people are not interested in this and it's also not providing any value. What you want is you want people who follow Gartner to see that you're on a Magic Quadrant and then come to you. And what you want is for this to be mid-funnel content, where if you're talking to some enterprise that has a little like, hey, can you guys actually handle our volume or whatever?

You can say, hey, look, Gartner's evaluated us, Forrester, whatever, blah, blah, blah. Like, where do you agree, disagree with this? How's my take?

Gonto: I agree. My note actually, I literally wrote, it's okay if it's hidden. Like, it can be part of the newsletter, but just hidden with other stuff.

Like, I like the idea of like, I will mention it, but it's not the main thing that we mentioned. It's just part of a newsletter, part of something that we send. And in that case, I think it's okay.

The other way that I would do it is actually being funny on how we're bragging about it. Like, I would tweet like, oh, we're now visionaries, something like that. And then like, hashtag humblebrag and potentially like connecting with VCs for congratulating themselves or something like that.

I think if you congratulate yourself, but you're actually funny about how that's a fucked up thing to do, then I think it can work.

Hank: Yeah. That's a careful line to walk. I would not recommend that to most people.

I would just, just tuck it in the footer of a newsletter, put it in your blog and you're good. But you do these things so that traffic flows to you. People used to do it because it was something to talk about and brag about, and it was appreciated more.

But since every company can effectively pay to play with Gartner and Forrester and the other analysts.

Gonto: It's not that easy, I'd say. Like, of course you need to pay, but it takes years to become a visionary by working with them, showing them and stuff like that. So I don't think it's that easy.

I would still not show it because the only people that will care are the enterprise folks. And for them, you either have a special page for that or you send it to them directly or something like that.

Hank: But the only people who know the distinctions between being on the quadrant and being a visionary on the quadrant and that stuff, it's people like you and me who have had to pay at previous companies, so much money to get on those quadrants and done the years of work. We're talking dev tool marketing, 0 developers understand the distinctions on these different squares and charts.

Gonto: The only people that do know besides us are, I always think about Gartner. Gartner is the way of you converting into IBM. This idea of like, nobody got fired for bringing or buying IBM.

Being a Gartner visionary leader basically puts you in that category. It's like, okay, if Gartner says you're a visionary, that means they are stable enough that you're not going to get fired if you hire them. So it gives more broad enterprises to use it.

But as you said, like developers or startups, they don't give a shit about it.

Hank: Yeah. And Gartner and Forrester and the like, they have services where they will generate leads for you because companies like IBM actually pay them and ask them like, hey, what tech should we be looking at? What should we be investigating?

I think that's it for today. That's it. Good stuff. Thanks, everybody.

Gonto: Thank you all.