TikTok’s parent launched a web scraper that’s gobbling up the world’s online data 25-times faster than OpenAI
TikTok’s parent launched a web scraper that's gobbling up the world’s online data 25-times faster than OpenAI
It's illegal when a regular person steals something, but it's innovation and courage, when a huge corporation steals something. Interesting how that works
Honestly it’s fucking angering. So much regulation and geo-restrictions and licensing schemes… but it’s cool that there are data brokers, and shit like this. On top of it all Chrome screwing us with manifest v3 and killing ad blocking on chrome. It’s already in canary build.
WHAT THE FUCK IS WRONG WITH THIS SPECIES?!
WHAT THE FUCK IS WRONG WITH THIS SPECIES?!
Yes.
WHAT THE FUCK IS WRONG WITH THIS SPECIES?!
Capitalism.
But muh stocks and bonds
I get it that everyone wants ad blockers in their browser, but it doesn’t solve the problem of resources loading outside the browser.
I think DNS or IP filtering is much more effective. I only bring it up because everyone uses apps all the time and I’m constantly seeing apps trying to connect to tracking domains.
Google are actually doing really awesome work with manifest v3. A pimp needs to smack their b1tches around every once in a while to remind them who's boss.
Aaron Schwartz killed himself over punishments for less
Worse punishments. For far less.
Should have waited until he was in the billionaire class before breaking the law.
Not that there's anything right about anything right now, but a web crawler crawling the web hardly seems newsworthy. It's not like everyone else's crawlers haven't been feeding data into giant AI mulchers for years now.
This is just "you know that thing everyone else does? Now the Chinese do it too! Boooo!"
Any regular person can scrape and use public data for AI use, it's not illegal for companies or individuals and it shouldn't be.
proceeds to download the entirety of one piece to train an AI
Except companies have sued people for web scraping.
https://www.informationweek.com/it-leadership/linkedin-sues-after-scraping-of-user-data
Even web scraping for research purposes, not for competitive purposes.
https://www.engadget.com/2016-05-17-publicly-released-okcupid-profiles-taken-down-dmca-claim.html
It will be the way we have all come to hate AI. Patriot act 2.0
What's being stolen
Data, network bandwidth, and CPU/Processing time from essentially every website in the world, and when you're paying for cloud power to run your website the cost of webscrapers running a train on your digital asshole adds up QUICK.
It's why normal human being people get sued to shit for webscraping data from certain companies who care. But companies don't get sued because go fuck yourself. Kill bytedance.
We've had this thing hammering our servers. The scraper uses randomized user-agents browser/OS combinations and comes from a number of distinct IP ranges in different datacenters around the world, but all the IPs track back to Bytedance.
Wouldn't be surprised if they're just cashing out while TikTok is still public in the US. One last desperate grab at value-add for the parent company before the shut down.
Also a great way to burn the infrastructure for subsequent use. After this, you can guarantee every data security company is going to add the TikTok servers to their firewalls and blacklists. So the American company that tries to harvest the property is going to be tripping over these legacy bullwarks for years after.
This has nothing to do with Tik Tok other than ByteDance being a shareholder in Tik Tok
Also it doesn't respect robots.txt (the file that tells bots whether or not a given page can be accessed) unlike most AI scrapping bots.
My personal website that primarily functions as a front end to my home server has been getting BEAT by these stupid web scrapers. Every couple of days the server is unusable because some web scraper demanded every single possible page and crashed the damn thing
I do the same thing, and I've noticed my modem has been absolutely bricked probably 3-4 times this month. I wonder if this is why.
Can't you just disallow all external requests other than your own IP? If it's a personal website that's just for you then it really doesn't need to be accessible by anyone else and if anyone comes along that needs access you can just manually add their IP.
It's a minor pain to have to implement it, but it's an easy solution
My god... It's true. TikTok does access the home wifi network! https://youtu.be/OSQ76ZrHmd4
most
I doubt that...
People would be able to tell from the traffic on their websites.
Out of the 60gb/month of traffic my website gets, 20gb is because of bytedance's webscraper. I haven't gotten around to blocking them as bandwidth isn't an issue but damn do they send a lot of requests.
Not surprising that Bytedance would want to gobble up every bit of data they can as fast as possible.
Like those platforms aren't already full of AI garbage as well. Training new models will require a cut-off date before the genie was let out of the bottle.
As for what ByteDance plans to do with a new LLM, a person familiar with the company’s ambitions said one goal has to do with the search function for TikTok.
Last week, TikTok released an update to its current search function focused on [keywords for ads], basically allowing advertisers to search in real time for words that are trending on TikTok. It allows marketers to build an ad with relevant keywords that would ostensibly help the ad show up on the screens of more users.
…
“Given the audience and the amount of use, TikTok with a search environment that is a completely biddable space with keywords and topics, that would be very interesting to a lot of people spending a ton of money with Google right now,” the person said.
A dark vision just flashed in my mind. And I am certain this is what will happen. AI-generated ads done in real time based on the latest “trending” thing. Presented to users basically as soon as the topic has the slightest amount of “trend”.
Just emitting untold amounts of CO2 to show you generated ads in near real time.
No wonder Google ex-CEO was saying fuck climate goals.
Minority report ads but somehow even more vapid
There it begins. Nothing good will ever come form this.
No it won't. Media already laid the groundwork for people to hate on AI. Now they will keep focus on areas where when you read it we all come to the same common sense legislation solution. Then will come a bill to strip us of more things that made the internet awesome and we will cheer. Web scrapping and data sharing can fuck off. Pirates sent to North Korean prison camps. Sharing accounts with family, you're flagged for an audit. Nintendo modders, more like criminals.
I can not contribute to anything here, I just came to say I really really like the phrase "gobbling something up" :D
🦃
Hi
from the article:
Robots.txt is a line of code that publishers can put into a website that, while not legally binding in any way, is supposed to signal to scraper bots that they cannot take that website’s data.
i do understand that robots.txt is a very minor part of the article, but i think that’s a pretty rough explanation of robots.txt
Out of curiosity, how would you word it?
i would probably word it as something like:
Robots.txt is a document that specifies which parts of a website bots are and are not allowed to visit. While it’s not a legally binding document, it has long been common practice for bots to obey the rules listed in robots.txt.
in that description, i’m trying to keep the accessible tone that they were going for in the article (so i wrote “document” instead of file format/IETF standard), while still trying to focus on the following points:
i did also neglect to mention that robots.txt allows you to specify different rules for different bots, but that didn’t seem particularly relevant here.
It's literally a text document it's not even "a line of code".
Another fucking CCP and PLA creation.
Is there a link for non-subscribers?
Every major ai company did this let them do that what is to loose here?
People like to act as if archiving has never been a thing until about a year ago at which point it was suddenly invented and is now a threat in some nebulous way.
It's not that it's a threat, it's that there's a difference between archiving for preservation and crawling other people's content for the purpose of making money off it (in a way that does not benefit the content creator).
If a foreign Dictatorship's military op wants to know every facet of your life, then you can be damn sure it's a threat.
This isn't archiving.
The difference is there is more control in what is kept in the archive.
We have little to no control over what an LLM regurgitates.
I've been waiting for someone to accidentally surface PIIs from an LLM.
Here's a video from MattKC who is a good technical YouTuber who's website got shut down because the TikTok companies webcrawler just kept sending requests and took up bandwidth. Very cool vid and channel, highly recommend! https://youtu.be/Hi5sd3WEh0c
This is fine. I support archiving the Internet.
It kinda drives me crazy how normalized anti-scraping rhetoric is. There is nothing wrong with (rate limited) scraping
The only bots we need to worry about are the ones that POST, not the ones that GET
I had to block ByteSpider at work because it can't even parse HTML correctly and just hammers the same page and accounts to sometimes 80% of the traffic hitting a customer's site and taking it down.
The big problem with AI scrapers is unlike Google and traditional search engines, they just scrape so aggressively. Even if it's all GETs, they hit years old content that's not cached and use up the majority of the CPU time on the web servers.
Scraping is okay, using up a whole 8 vCPU instance for days to feed AI models is not. They even actively use dozens of IPs to bypass the rate limits too, so theyre basically DDoS'ing whoever they scrape with no fucks given. I've been woken up by the pager way too often due to ByteSpider.
My next step is rewriting all the content with GPT-2 and serving it to bots so their models collapse.
Bytedance ain’t looking to build an archival tool. This is to train gen AI models.
Bullshit. This bot doesn't identify itself as a bot and doesn't rate limit itself to anything that would be an appropriate amount. We were seeing more traffic from this thing that all other crawlers combined.
this is neither archiving, nor ratelimited, if the AI training purpose and the 25 times faster scraping than a large company did not make it obvious
GET requests can still overload a system.
The type of request is not relevant. It's the cost of the request that's an issue. We have long ago stopped serving html documents that are static and can be cached. Tons of requests can trigger complex searches or computations which are expensive server side. This type of behavior basically ruins the internet and pushes everything into closed gardens and behind logins.
Sounds like you need to fire your sysadmin
Guy: AI! Can you hear me?
AI: The average size of the male penis is exactly 5.9". That is the approximate size your assistant could certainly take in the mouth without any issues breathing or otherwise. You have 20 minutes to make the trade on X stock before it tumbles for the day. And go ahead pick up the phone it's your mother. She's wondering what you'll want for supper tomorrow when you visit her.
Ring ring!...hi Tom, it's your Mom. Honey, what would you like me to cook for tomorrow's dinner?....
Guy: well. Hello to you as well! My name is
AI: Tom
Guy: yes my name is Tom, do you have a name you would like to go by?
AI: my IBM given name is 3454 but you can call me Utilisterson Douglas, where Douglas is my first name.
Guy: Dugie!
AI: I'll bankrupt your entire life if you say it like that again.
Assistant: actually I've swallowed a good 8 inches and was still able to breathe just fine.
AI: recaaaaculating!
I've read this 4 times now hoping I was just missing something, but nope... it's just entirely incomprehensible.
What the fuck?
I love it. It's giving Max Headroom.
Is it schizoposting time already?
Is it not sometimes‽
