It’s difficult to be a shittier company than OpenAI, but Perplexity seems to be trying hard.
Step 1, SOMEHOW find a more punchable face than Altman
Altman’s face looks like it’s already been punched
When a firm outright admits to bypassing or trying to bypass measures taken to keep them out, you think that would be a slam dunk case of unauthorized access under the CFAA with felony enhancements.
Fuck that. I don’t need prosecutors and the courts to rule that accessing publicly available information in a way that the website owner doesn’t want is literally a crime. That logic would extend to ad blockers and editing HTML/js in an “inspect element” tag.
That logic would not extend to ad blockers, as the point of concern is gaining unauthorized access to a computer system or asset. Blocking ads would not be considered gaining unauthorized access to anything. In fact it would be the opposite of that.
gaining unauthorized access to a computer system
And my point is that defining “unauthorized” to include visitors using unauthorized tools/methods to access a publicly visible resource would be a policy disaster.
If I put a banner on my site that says “by visiting my site you agree not to modify the scripts or ads displayed on the site,” does that make my visit with an ad blocker “unauthorized” under the CFAA? I think the answer should obviously be “no,” and that the way to define “authorization” is whether the website puts up some kind of login/authentication mechanism to block or allow specific users, not to put a simple request to the visiting public to please respect the rules of the site.
To me, a robots.txt is more like a friendly request to unauthenticated visitors than it is a technical implementation of some kind of authentication mechanism.
Scraping isn’t hacking. I agree with the Third Circuit and the EFF: If the website owner makes a resource available to visitors without authentication, then accessing those resources isn’t a crime, even if the website owner didn’t intend for site visitors to use that specific method.
When sites put challenges like Anubis or other measures to authenticate that the viewer isn’t a robot, and scrapers then employ measures to thwart that authentication (via spoofing or other means) I think that’s a reasonable violation of the CFAA in spirit — especially since these mass scraping activities are getting attention for the damage they are causing to site operators (another factor in the CFAA, and one that would promote this to felony activity.)
The fact is these laws are already on the books, we may as well utilize them to shut down this objectively harmful activity AI scrapers are doing.
Nah, that would also mean using Newpipe, YoutubeDL, Revanced, and Tachiyomi would be a crime, and it would only take the re-introduction of WEI to extend that criminalization to the rest of the web ecosystem. It would be extremely shortsighted and foolish of me to cheer on the criminalization of user spoofing and browser automation because of this.
Do you think DoS/DDoS activities should be criminal?
If you’re a site operator and the mass AI scraping is genuinely causing operational problems (not hard to imagine, I’ve seen what it does to my hosted repositories pages) should there be recourse? Especially if you’re actively trying to prevent that activity (revoking consent in cookies, authorization captchas).
In general I think the idea of “your right to swing your fists ends at my face” applies reasonably well here — these AI scraping companies are giving lots of admins bloody noses and need to be held accountable.
I really am amenable to arguments wrt the right to an open web, but look at how many sites are hiding behind CF and other portals, or outright becoming hostile to any scraping at all; we’re already seeing the rapid death of the ideal because of these malicious scrapers, and we should be using all available recourse to stop this bleeding.
DoS attacks are already a crime, so of course the need for some kind of solution is clear. But any proposal that gatekeeps the internet and restricts the freedoms with which the user can interact with it is no solution at all. To me, the openness of the web shouldn’t be something that people just consider, or are amenable to. It should be the foundation in which all reasonable proposals should consider as a principle truth.
That same logic is how Aaron Swartz was cornered into suicide for scraping JSTOR, something widely agreed to be a bad idea by a wide range of lawspeople including SCOTUS in its 2021 decision Van Buren v. US that struck this interpretation off the books.
Site owners currently do and should have the freedom to decide who is and is not allowed to access the data, and to decide for what purpose it gets used for. Idgaf if you think scraping is malicious or not, it is and should be illegal to violate clear and obvious barriers against them at the cost of the owners and unsanctioned profit of the scrapers off of the work of the site owners.
to decide for what purpose it gets used for
Yeah, fuck everything about that. If I’m a site visitor I should be able to do what I want with the data you send me. If I bypass your ads, or use your words to write a newspaper article that you don’t like, tough shit. Publishing information is choosing not to control what happens to the information after it leaves your control.
Don’t like it? Make me sign an NDA. And even then, violating an NDA isn’t a crime, much less a felony punishable by years of prison time.
Interpreting the CFAA to cover scraping is absurd and draconian.
If you want anybody and everyone to be able to use everything you post for any purpose, right on, good for you, but don’t try to force your morality on others who rely on their writing, programming, and artworks to make a living to survive.
I’m gonna continue to use ad blockers and yt-dlp, and if you think I’m a criminal for doing so, I’m gonna say you don’t understand either technology or criminal law.
If I put a banner on my site that says “by visiting my site you agree not to modify the scripts or ads displayed on the site,” does that make my visit with an ad blocker “unauthorized” under the CFAA?
How would you “authorize” a user to access assets served by your systems based on what they do with them after they’ve accessed them? That doesn’t logically follow so no, that would not make an ad blocker unauthorized under the CFAA. Especially because you’re not actually taking any steps to deny these people access either.
AI scrapers on the other hand are a type of users that you’re not authorizing to begin with, and if you’re using CloudFlares bot protection you’re putting into place a system to deny them access. To purposefully circumvent that access would be considered unauthorized.
That doesn’t logically follow so no, that would not make an ad blocker unauthorized under the CFAA.
The CFAA also criminalizes “exceeding authorized access” in every place it criminalizes accessing without authorization. My position is that mere permission (in a colloquial sense, not necessarily technical IT permissions) isn’t enough to define authorization. Social expectations and even contractual restrictions shouldn’t be enough to define “authorization” in this criminal statute.
To purposefully circumvent that access would be considered unauthorized.
Even as a normal non-bot user who sees the cloudflare landing page because they’re on a VPN or happen to share an IP address with someone who was abusing the network? No, circumventing those gatekeeping functions is no different than circumventing a paywall on a newspaper website by deleting cookies or something. Or using a VPN or relay to get around rate limiting.
The idea of criminalizing scrapers or scripts would be a policy disaster.
Ehhhh, you are gaining access to content due to assumption you are going to interact with ads and thus, bring revenue to the person and/or company producing said content. If you block ads, you remove authorisation brought to you by ads.
That doesn’t make any logical sense. You cant tie legal authorization to an unsaid implicit assumption, especially when that is in turn based on what you do with the content you’ve retrieved from a system after you’ve accessed and retrieved it.
When you access a system, are you authorized to do so, or aren’t you? If you are, that authorization can’t be retroactively revoked. If that were the case, you could be arrested for having used a computer at a job, once you’ve quit. Because even though you were authorized to use it and your corporate network while you worked there, now that you’ve quit and are no longer authorized that would apply retroactively back to when you DID work there.
Right? Isn’t this a textbook DMCA violation, too?
Uh… good?
You could say they are… Perplexed.
That’s the entire point, dipshit. I wish we got one of the cool techno dystopias rather than this boring corporate idiot one.
You’d think that a competent technology company, with their own AI would be able to figure out a way to spoof Cloudflare’s checks. I’d still think that.
Or find a more efficient way to manage data, since their current approach is basically DDOSing the internet for training data and also for responding to user interactions.
This is not about training data, though.
Perplexity argues that Cloudflare is mischaracterizing AI Assistants as web crawlers, saying that they should not be subject to the same restrictions since they are user-initiated assistants.
Personally I think that claim is a decent one: user-initiated request should not be subject to robot limitations, and are not the source of DDOS attack to web sites.
I think the solution is quite clear, though: either make use of the user identity to walz through the blocks, or even make use of the user browser to do it. Once a captcha appears, let the user solve it.
Though technically making all this happen flawlessly is quite a big task.
Personally I think that claim is a decent one: user-initiated request should not be subject to robot limitations, and are not the source of DDOS attack to web sites.
They are one of the sources!
The AI scraping when a user enters a prompt is DDOSing sites in addition to the scraping for training data that is DDOSing sites. These shitty companies are repeatedly slamming the same sites over and over again in the least efficient way because they are not using the scraped data from training when they process a user prompt that does a web search.
Scraping once extensively and scraping a bit less but far more frequently have similar impacts.
When user enters a prompt, the backend may retrieve a handful a pages to serve that prompt. It won’t retrieve all the pages of a site. Hardly different from a user using a search engine and clicking 5 topmost links into tabs. If that is not a DoS attack, then an agent doing the same isn’t a DDoS attack.
Constructing the training material in the first place is a different matter, but if you’re asking about fresh events or new APIs, the training data just doesn’t cut it. The training, and subsequenctly the material retrieval, has been done a long time ago.
Can’t believe I’ve lived to see Cloudflare be the good guys
They’re not. They’re using this as an excuse to become paid gatekeepers of the internet as we know it. All that’s happening is that Cloudflare is using this to menuever into position where they can say “nice traffic you’ve got there - would be a shame if something happened to it”.
AI companies are crap.
What Cloudflare is doing here is also crap.
And we’re cheering it on.
Lesser of two bad guys maybe?
This is why companies like Perplexity and OpenAI are creating browsers.
ask AI how to do it?
I set up a WAF for my company’s publicly facing developer portal to block out bot traffic from assholes like these guys. It reduced bot traffic to the site by something like - I kid you not - 99.999%.
Fucking data vultures.
Can someone with more knowledge shine a bit more light on this while situation? Im out of the loop on the technical details
AI crawlers tend to overwhelm websites by doing the least efficient scraping of data possible, basically DDOSing a huge portion of the internet. Perplexity already scraped the net for training data and is now hammering it inefficiently for searches.
Cloudflare is just trying to keep the bots from overwhelming everything.
Perplexity (an “AI search engine” company with 500 million in funding) can’t bypass cloudflare’s anti-bot checks. For each search Perplexity scrapes the top results and summarizes them for the user. Cloudflare intentionally blocks perplexity’s scrapers because they ignore robots.txt and mimic real users to get around cloudflare’s blocking features. Perplexity argues that their scraping is acceptable because it’s user initiated.
Personally I think cloudflare is in the right here. The scraped sites get 0 revenue from Perplexity searches (unless the user decides to go through the sources section and click the links) and Perplexity’s scraping is unnecessarily traffic intensive since they don’t cache the scraped data.
…and Perplexity’s scraping is unnecessarily traffic intensive since they don’t cache the scraped data.
That seems almost maliciously stupid. We need to train a new model. Hey, where’d the data go? Oh well, let’s just go scrape it all again. Wait, did we already scrape this site? No idea, let’s scrape it again just to be sure.
They do it this way in case the data changed, similar to how a person would be viewing the current site. The training was for the basic understanding, the real time scraping is to account for changes.
It is also horribly inefficient and works like a small scale DDOS attack.
It’s worth giving the article a read. It seems that they’re not using the data for training, but for real-time results.
First we complain that AI steals and trains on our data. Then we complain when it doesn’t train. Cool.
It seems like it’s some kind of distraction to make people think things aren’t as bad as they really are, it just sounds too far-fetched to me.
It’s like a bear that has eaten too much and starts whining because a small rabbit is running away from him, even though the bear has already eaten almost all the rabbits and is clearly full.
Skill issue. Cope and seethe
