> Here’s what’s going on. The Internet Archive’s Wayback Machine has been hit by waves of high-volume automated traffic, and we’ve put protections in place to keep the service running.
I'm pretty certain this is scrapers that are trying to workaround blocks on accessing original sites by hitting the Wayback Machine copy instead. Appalling behavior.
In addition to the load it puts on this vital non-profit piece of Internet infrastructure, we've ålso already seen some sites opt out of the Wayback Machine to prevent their content from being scraped via this alternative route.
We used to always "scrape" the wayback machine for any sort of news article we actually paid to consume. I was absolutely shocked by major news sites making very important edits to an article without any sort of editorial notice!
Sadly this sort of thing is probably not really possible anymore, but I can't really blame anyone for making this sort of decision. I can't imagine how much more traffic they get now vs 2021 when we were doing this.
Unfortunately I sometimes have to browbeat Claude into acting like an agent of the user is supposed to. Usually it works, though last time it refused to recognize my moral argument (on the grounds that it's not bound to my interests exclusively and needs to protect the interests of its maker too).
Do you think there should be a way for a site to tell an agent it isn't allowed access? I'm not sure where I land on this exactly tbh, so no judgment cast.
Has it been settled whether robots.txt applies to user-driven chat sessions and if things like the crawl delay should be applied to say an end-user, an ip address, a harness provider, etc? My understanding is robots.txt is more for training exclusions, but less so for agent work.
> I'm pretty certain this is scrapers that are trying to workaround blocks on accessing original sites by hitting the Wayback Machine copy instead. Appalling behavior.
Appalling, yes. But also expected. I'm surprised they haven't been the target of scrapers for years. But sites putting their content behind login walls and other anti-bot mechanisms has certainly exacerbated this. But again, this isn't at all a surprising progression.
> we've also already seen some sites opt out of the Wayback Machine to prevent their content from being scraped via this alternative route.
To be fair, another big motivation was likely users on sites like HN using archive.org (and similar sites) to get around their paywalls. In fact, I'd be surprised if this wasn't a big motivator.
Again, it sucks, but it's not at all surprising to see it progress like this. I wouldn't be surprised to see similar blocks on other archive sites eventually.
What sites would they be targeting? Generic "just give me anything"? Whenever I check regular sites on IA, the coverage is spotty -- they'll have the homepage and a few important pages, but it quickly fizzles out.
Very understandable, you can't store all 15000 pages of any random website and update them etc etc, but that makes them pretty useless for indirect scraping because you usually don't want a tiny taste, you want everything.
It is. They will most likely eventually need to move to a walled model for Wayback due to scraper aggressiveness (like Reddit deprecating anonymous old.reddit.com), or behind Cloudflare for aggressive bot and scraping protection. Hard to defend against abuse of a public resource when its intent is public access with as little restriction as possible.
As someone who operates a large non-profit public data driven website, I have some VERY strong feelings about scrapers. We looked into various commercial solutions (Datadome, HUMAN) and based on our traffic estimates from logs we'd be looking at at least 250k/yr for bot mitigation. Anubis is offering a temporary reprieve, but after reading the recent kernel.org article [0] it's increasingly clear that this is a temporary bandaid.
The cheapest solution is to require a login and rate limit by API key. I also have strong feelings about the tragedy of the commons.
It's shame that the AI arms race causes such collateral damage. Free resources were always exploited, but the stakes ($T) and capabilities around AI allow unprecedented abuse. I wish we could go back... :/
I really don't see any solution to this; the scrapers probably wouldn't even mind destroying sources like IA too much, which would leave them as the only "authorative" source of knowledge in the end. Best way is likely regulation incl. hefty (!) fines, but politics are too slow and too fragmented to be effective. So... Enjoy it while it lasts, I guess.
Mad props to the people at the Archive. You are the heros we need in a formerly open internet that is surrendering to evil big corps and closing down free access.
The Internet Archive is in a really bad spot being attacked from multiple sides at once. But — while service has not been consistent — they have maintained open access. I can still access anonymously from Tor without Cloudflare or some other centralized gatekeeper showing me the middle finger.
If you got some money to spare, consider donating to them. They need it.
Is it established that the scraping scourge of late is primarily driven by AI companies? Anyone aware of any relevant studies?
Beginning to think that the difficulty to browse most websites nowadays due to throttling is yet another negative externality of AI development that society is forced to bear.
I've not been able to access web.archive.org from my work computer - I always get the 429 error.
But I then pull out my phone and can access it just fine. All along I was assuming my company was blocking it. Still weird that it happens every time from my work PC and never from my home one.
I assume that's because the IP range of your company network overlaps with a range used by some scrapers, and if it doesn't happen on your phone even in the corporate network, then IA probably checks some extra signals like the user agent in addition to the IP
Your workplace is probably redirecting traffic through a datacenter IP range. Especially if they have their own datacenters like google, microsoft, oracle, amazon, etc.
Try making a vpn via digital ocean for example and you'll see similar patterns.
That is essentially what LLM vector search results are, but the misappropriated $9Tn worth of FOSS code "AI" scraped and compacted for isomorphic plagiarism tokens is harder to prove now with watermarking skewed outputs. =3
I've been getting this error a lot. Asking users to email them with details of their OS, browser, IP address is just crazy. Their support is supposedly already swamped and they are asking for more!? Changes made by IA shouldn't become my responsibility.
> We’re getting better at telling abusive bots apart from the people who depend on the Wayback Machine every day. If you think you were blocked in error, email [email protected] with your operating system, browser, and IP address, and we’ll look into it.
I really appreciate the Archive team's efforts to make the Wayback Machine more responsive, and have donated a few times to support them.
Unfortunately, the restrictions have been way too strict for the last few months: from my residential IP, simply moving the mouse on the calendar for a specific URL is enough to get stuck on 429 error messages for a while; and from corporate ISPs (for example, on airport WiFi), you often can't access the WM at all. I hope they'll find a way to relax those.
There was a feature on Amazon Web Services for a while, and I wish it was still there...
Downloader pays.
I make some content and upload it. When you want to download it, you pay Amazon the egress fees. And maybe I get to charge just a bit more, to help me with the Ingress, storage, content creation, etc.
I mean, I know that there's going to be problems with rate limiting, etc. And yes, we have those problems with LLM tokens today. But this just feels like such a useful thing that it baffles me that it doesn't exist already.
unrelated: if a website gets hit with "This URL has been excluded from the Wayback Machine", do existing snapshots get purged or may they still be preserved somewhere?
There is a HN article on abusive AI crawlers on the front page almost every week, but we rarely talk about the path forward. Web scraping has been around for as long as the internet, and it was fine because we had established best practices (rate limiting, self-identification, robots.txt, etc.) that the industry agreed upon. Now we have AI labs and their crawlers that don't care about any of this gentlemen's agreement.
So how do we go from here? Is adding more difficult Anubis and Cloudflare bot protection really the solution? How many millions of human hours and billions in infra costs are we willing to spend on this arms race?
Some approaches that I think are promising:
- A robots.txt V2[0] as a standard way for website owners to state how bots and AI crawlers can use their online content and where to go (e.g. distinguish search from AI training use cases, point to a downloadable file instead of crawling everything, etc.).
- Something like Web Bot Auth[1] as a non-centralized standard for self-identifying bots and agents cryptographically. This would allow websites to allow or deny bots very precisely.
Yeah it’s kinda crazy to me that what was once a back alley python script is now accepted as ‘fine, free for all’. The new era of bros really are smth else.
Why not just offer a paid endpoint for the crawlers? It's not like the demand is going to go away anytime soon.
It serves nobody except CloudFlare and hardware companies when one side set up blockers and the other side spend money putting VPN SDKs in consumer TVs.
I am also curious how the (Russian?) paywall bypass mirror archive.is is doing given that they are probably subject to similar amounts of traffic.
Micropayments would solve so many Internet problems. It's not too late to adopt.
Content creators could charge by page instead of depending on malware/ad/surveillance revenue. Spam is cut if there's a charge per mail. Scraping abuse goes away, along with a bunch of DDOS garbage.
The impact is a few cents per page or mail, negligible for a human. But if you're consuming a trillion pages per day, you'd reconsider.
Micropayments would solve all the problems except for the problem that people absolutely loathe micropayments. Like, vein-popping furiously hate them.
Whenever the topic of micropayments for internet content comes up, a bunch of people start talking about payment processors and their floor on prices, and so on. That's not wrong, but it can be designed around and I think it's a scapegoat to avoid confronting the fact that users despise micropayments and we'd rather blame credit card companies for the lack of adoption.
Micropayments would solve so many problems for the internet. And, cryptocurrencies would solve so many problems for micropayments. But, it's a non-starter because any proposal gets flooded with people popping veins about how crypto can't solve anything.
> Why not just offer a paid endpoint for the crawlers?
Because then you're definitely violating US copyright law. There are four prongs of fair use analysis, and one of them is the "nature of the use." In this case, you'd be turning into a commercial use.
What if you're not charging for the content, but as compensation for the network bandwidth / server resources consumed by serving that content? The idea isn't to profit from content (the IA is a nonprofit anyway), just to allow the IA to continue to serve its purpose as an archive of public data without being overwhelmed by bots.
My guess? Because even with a paid endpoint, the type of unscrupulous yahoo that is DDOSing IA today would probably still abuse the free endpoints because they can. The revenue that might come from a paid endpoint could help to scale up, but with how slow IA usually seems, I suspect there is an upper limit to how much traffic they can serve without a LOT more revenue.
This is a major "this is why we can't have nice things" situation in my opinion. IA is one of the most valuable gems of the Internet. The only thing that even comes close to preserving our shared history. The damage being caused (both by the effective DDOSing and by the knock-on impact that abuse has in encouraging publishers to remove their content from the archive) is incredibly serious.
Which AI company is selling access to reliable verbatim copies of websites? I don't mean "it may regurgitate a paragraph", but as a reliable service where you can repeatably get website content snapshots to a reliability level that makes such a use case viable?
Using the information for training purposes is not the same thing. Not legally the same and otherwise.
Internet Archive was almost destroyed by a copyright lawsuit from book publishers within the last few years. I expect they aren't excited to take on any additional risk of similar lawsuits right now.
They need access to sites to archive them. It's already hard to do it as it is, imagine if they start selling access to content. They'd be shooting themselves on the foot, independently of what the law says.
major [citation needed] on that. There are very limited exceptions to the massive power of copyright -- and they're mainly granted to libraries in the form of narrow waivers. And just the cost of fighting the most powerful copyright holders can bankrupt you -- especially if you're a relatively modestly-funded nonprofit.
No one is upset that the AI companies are scraping the web, they’re upset how poorly implemented the scrapers, but the scraping itself is fine. Lots of people and businesses scrape the web.
Then... don't? If it sucks so much why do you need to read the tweets?
Great take: "This private website is owned by a man I don't like, so I refuse to pay for it - or even give it the possibility to monetize my traffic with ads!"
Still quite mainstream take: "... so I'll use an adblocker on it"
Immature take: "This private website that I hate and boycott is also an important part of our culture, but the posts on it are too important and valuable to ignore, so I'll use a proxy to scrape it"
I'm pretty certain this is scrapers that are trying to workaround blocks on accessing original sites by hitting the Wayback Machine copy instead. Appalling behavior.
In addition to the load it puts on this vital non-profit piece of Internet infrastructure, we've ålso already seen some sites opt out of the Wayback Machine to prevent their content from being scraped via this alternative route.
Hence, distinction without a difference.
Sadly this sort of thing is probably not really possible anymore, but I can't really blame anyone for making this sort of decision. I can't imagine how much more traffic they get now vs 2021 when we were doing this.
> Hacker News and the Rails forum are blocking the text fetcher, so I'm using the browser workflow to inspect the pages directly
Unfortunately I sometimes have to browbeat Claude into acting like an agent of the user is supposed to. Usually it works, though last time it refused to recognize my moral argument (on the grounds that it's not bound to my interests exclusively and needs to protect the interests of its maker too).
Appalling, yes. But also expected. I'm surprised they haven't been the target of scrapers for years. But sites putting their content behind login walls and other anti-bot mechanisms has certainly exacerbated this. But again, this isn't at all a surprising progression.
> we've also already seen some sites opt out of the Wayback Machine to prevent their content from being scraped via this alternative route.
To be fair, another big motivation was likely users on sites like HN using archive.org (and similar sites) to get around their paywalls. In fact, I'd be surprised if this wasn't a big motivator.
Again, it sucks, but it's not at all surprising to see it progress like this. I wouldn't be surprised to see similar blocks on other archive sites eventually.
Very understandable, you can't store all 15000 pages of any random website and update them etc etc, but that makes them pretty useless for indirect scraping because you usually don't want a tiny taste, you want everything.
https://en.wikipedia.org/wiki/Tragedy_of_the_commons
(no affiliation)
On the other hand, the Internet Archive is a non-profit offering a free public resource.
The cheapest solution is to require a login and rate limit by API key. I also have strong feelings about the tragedy of the commons.
[0] https://people.kernel.org/monsieuricon/creepy-crawlies
I really don't see any solution to this; the scrapers probably wouldn't even mind destroying sources like IA too much, which would leave them as the only "authorative" source of knowledge in the end. Best way is likely regulation incl. hefty (!) fines, but politics are too slow and too fragmented to be effective. So... Enjoy it while it lasts, I guess.
The Internet Archive is in a really bad spot being attacked from multiple sides at once. But — while service has not been consistent — they have maintained open access. I can still access anonymously from Tor without Cloudflare or some other centralized gatekeeper showing me the middle finger.
If you got some money to spare, consider donating to them. They need it.
The future is bleak :\
Beginning to think that the difficulty to browse most websites nowadays due to throttling is yet another negative externality of AI development that society is forced to bear.
I've not been able to access web.archive.org from my work computer - I always get the 429 error.
But I then pull out my phone and can access it just fine. All along I was assuming my company was blocking it. Still weird that it happens every time from my work PC and never from my home one.
Try making a vpn via digital ocean for example and you'll see similar patterns.
Unfortunately, the restrictions have been way too strict for the last few months: from my residential IP, simply moving the mouse on the calendar for a specific URL is enough to get stuck on 429 error messages for a while; and from corporate ISPs (for example, on airport WiFi), you often can't access the WM at all. I hope they'll find a way to relax those.
Downloader pays.
I make some content and upload it. When you want to download it, you pay Amazon the egress fees. And maybe I get to charge just a bit more, to help me with the Ingress, storage, content creation, etc.
I mean, I know that there's going to be problems with rate limiting, etc. And yes, we have those problems with LLM tokens today. But this just feels like such a useful thing that it baffles me that it doesn't exist already.
Note that Archive Team is separate from the Internet Archive.
Some approaches that I think are promising:
- A robots.txt V2[0] as a standard way for website owners to state how bots and AI crawlers can use their online content and where to go (e.g. distinguish search from AI training use cases, point to a downloadable file instead of crawling everything, etc.).
- Something like Web Bot Auth[1] as a non-centralized standard for self-identifying bots and agents cryptographically. This would allow websites to allow or deny bots very precisely.
- what else?
[0] https://datatracker.ietf.org/doc/draft-vaughan-machine-reada...
[1] https://datatracker.ietf.org/doc/html/draft-meunier-http-mes...
It serves nobody except CloudFlare and hardware companies when one side set up blockers and the other side spend money putting VPN SDKs in consumer TVs.
I am also curious how the (Russian?) paywall bypass mirror archive.is is doing given that they are probably subject to similar amounts of traffic.
Content creators could charge by page instead of depending on malware/ad/surveillance revenue. Spam is cut if there's a charge per mail. Scraping abuse goes away, along with a bunch of DDOS garbage.
The impact is a few cents per page or mail, negligible for a human. But if you're consuming a trillion pages per day, you'd reconsider.
Whenever the topic of micropayments for internet content comes up, a bunch of people start talking about payment processors and their floor on prices, and so on. That's not wrong, but it can be designed around and I think it's a scapegoat to avoid confronting the fact that users despise micropayments and we'd rather blame credit card companies for the lack of adoption.
Because then you're definitely violating US copyright law. There are four prongs of fair use analysis, and one of them is the "nature of the use." In this case, you'd be turning into a commercial use.
This is a major "this is why we can't have nice things" situation in my opinion. IA is one of the most valuable gems of the Internet. The only thing that even comes close to preserving our shared history. The damage being caused (both by the effective DDOSing and by the knock-on impact that abuse has in encouraging publishers to remove their content from the archive) is incredibly serious.
Using the information for training purposes is not the same thing. Not legally the same and otherwise.
Great take: "This private website is owned by a man I don't like, so I refuse to pay for it - or even give it the possibility to monetize my traffic with ads!"
Still quite mainstream take: "... so I'll use an adblocker on it"
Immature take: "This private website that I hate and boycott is also an important part of our culture, but the posts on it are too important and valuable to ignore, so I'll use a proxy to scrape it"
Feels good when you get to dish it out doesn’t it?
Are the abusive bots in the room with us?