@perplexity_ai 1. Resolve this with cloudflare rather than attack them publicly 2. You’re still responsible for third parties you use. You can work with browserbase on fixing this, instead of blaming them too. 3. “more flair than cloud” lmao that was p good Overall shit pr
Such a well-written article, and I think it gets to the root of why Cloudflare did this silly publicity stunt. I have a hard time believing that the CF team didn’t know they were lying in their slam article. They knew exactly what they were doing. They just thought they could bully a startup to hopefully get some publicity juice for themselves. At the end of the day, it seems very clear that Cloudflare can’t actually do the one job they have - tell the difference between malicious traffic and meaningful traffic. Great opportunity for a Cloudflare competitor to come in and do this right. Well done Perplexity team standing up to the bully 💪
You haven’t addressed the core issue of User Agent Identity. Key excerpt: “When Perplexity fetches a webpage, it’s because you asked a specific question requiring current information.” Yes—but who is “you”? What is the identity of the user on whose behalf the agent is acting? And equally important—what is the identity of the agent itself? This lack of clarity around who is acting and on whose behalf isn’t a technical footnote—it’s a foundational gap at the center of growing concerns about LLM-based tools and AI agents. The Web has long been built around identity (e.g., WebIDs, tokens, cookies, logins). Concepts like authenticity, traceability, and consent are not optional—they are essential. Without a verifiable and transparent link between: 1. the user, 2. the agent acting on their behalf, and 3. the systems or services being queried, we risk creating a brittle and unaccountable ecosystem—especially in contexts involving personal data, intellectual property, or sensitive transactions. In short: identity authenticity and agent transparency must be part of the architecture. Fortunately, the Web already provides the infrastructure to support this properly—we just need to use it. /cc @AravSrinivas @eastdakota
I wrote an article on the difference between agents and bots and I agree they are fundamentally different. agentexperience.ax/research/ax-th… That said, like any other delegation of work (e.g. a currier picking up your groceries) if the door is locked, it’s not ok to break in. I don’t see what you’re doing as that, either. Nor do I believe Cloudflare should extort agents at the front door - it’s akin to stopping anyone that might be a door dash currier at the door of a store because the game of paying for shelf placement doesn’t work anymore for physical stores. For the web, the concern of ad rev is real and better models can exist - I don’t believe their approach is actually it and these things are clearly not purely for the betterment of the web alone
@perplexity_ai The sites still want a relationship with the end user. In some cases they want monetization too. We need open protocols that allow agents to express identity and payment (when necessary) for web access Checkout what’s live today with Skyfire: youtu.be/onkJ1LlS7q8?si…
@perplexity_ai Interesting read. It's the natural evolution as "legacy" businesses try to make sense of new tech. It's both ironic and unfortunate, but hopefully it resolves quickly for everyone.
@perplexity_ai Fwiw, Cloudflare also has their own browser automation / scraping solution that anyone can use to scrape web pages that are disallowed by robots.txt developers.cloudflare.com/browser-render…
@perplexity_ai Cool, cool but: - Will the agent become your fan? - Will it join your Patreon? - Will it buy your mug? - Will it join your Discord? - Will it click your ad? - Will it comment on your blog post? - Will it recommend the page to a friend? Yeah… I think it's still a bot.
@perplexity_ai When are you gonna respect the robot.txt file and stop stealing content from creators that don't want their content stolen?
@perplexity_ai So… It seems what Perplexity is trying to say is that - we’ll enter your house without permission, we’ll steal your things, but DO NOT WORRY, we won’t keep them forever. We’ll just use them and then discard them. Sounds about right? An article that could have been a tweet.
@perplexity_ai did yall seriously have an AI write this entire response? this is lazy af lmao
If a website is preventing bots from crawl, it should stop headless agents too. Cause websites expects real users to see their actual contents, more suggestions and ads too but when using headless agents, they dont get any of these. So if the website owner wants you to crawl, they can always allow it and cloudflare wont prevent it but if they are explicitly preventing it then you shouldnt bypass that
I don't want AI training or AI agents to touch, access, or use my work because it completely breaks my business model of providing content online. I should be able to tell you NO about that and you respect the NO. Removing consent and advocating for assaulting people's property is not the "internet evolving." You're just trying to manifest permission for criminal theft on countless victims by justifying exploitation as "helpful."
@perplexity_ai 10-8, arguably a 10-7 with one of the coldest corporate dispute lines I've ever heard > The bluster around this issue also reveals that Cloudflare’s leadership is either dangerously misinformed on the basics of AI, or simply more flair than cloud.
@perplexity_ai Is Perplexity saying: Cloudfare, you closed the door with a new locker, can’t you distinguish between a burglar and Robin Hood? The point is: who owns the house, the things in the house and the door? Not Perplexity. The only thing I learned here: cloudfare seems to be successful
@perplexity_ai Content thieves like you are why I have to log in to every god damned website to see anything anymore. Fuck off.
@perplexity_ai We need to figure out a monetization strategy for headless user agents (which I’m loosely defining user agents that don’t show you the source as-is) before they can be treated similar to browsers
@perplexity_ai These days, it's easy to justify anything if you're making enough money with it. And we thought Google was our worst nightmare!
Well written response distinguishing the use of agents akin to humans asking questions. The compliance aspects based on standards and agreements in the industry, need to be rewritten, I think. Nevertheles, I love @perplexity_ai and @PerplexityComet. Keep going what with you do and make the best.
@perplexity_ai Stop attacking the web Perplexity!
@perplexity_ai 1. get to the point omg I ain’t reading allat 2. Perplexity gave up any claim of being not malicious when they made their bot pretend to be chrome on macOS to bypass the bot protections 💀
@perplexity_ai You take away the content (value) that the site has worked so hard to create and then reorganize it to present it to users on your own platform, thus cutting off all the traffic, attention and potential revenue that would have gone to the original site.
@perplexity_ai Human directed AI is curiosity amplified, not a threat.
@perplexity_ai This post, and the replies to it, truly cast dread in my heart. Maybe humanity is truly beyond fucked. You just put up for discussion the very idea of consent (quoting: "If you can't tell..."), arguing that you should always have open doors by your very nature. What the fuck?
@perplexity_ai @amasad I see a lot of confusion among leadership about AI vs Automation vs Agents vs Bots. Good that it's involving.
@perplexity_ai If it really is human led then why not broker a Captcha deal with @Cloudflare - doesn't have to be every request but could use a number of factors to determine a human is actively driving @perplexity_ai
@perplexity_ai Loved the post. Such a great point about gate keeping in this era. Well done for speaking up.
@perplexity_ai @iruletheworldmo This seems akin to retailers blocking cell access to prevent comparison shopping in the moment. Didn't last long
@perplexity_ai Except every time I ask an llm to check on something that they can fetch they are already using search index data that's from the previous day. Why isn't it fetching in real time?
You're spot on, but it's actually worse. Cloudflare/Matthew Prince are seeking to extend the shelf-life of their Dinosaur company by impeding the new Intelligence Economy rather than pivoting aggressively into it. They are instead pivoting to being the Defenders of an already-dead notion of "intellectual property" and copyright, playing a LARP of champions of a sacred principle. To keep their stock price up, basically. They have nothing to offer. WeWork as a grandpa web co.