Submitted 5 days ago by jarfil@beehaw.org to technology@beehaw.org
Interesting approach. But of course it’s another black box, because otherwise it wouldn’t be effective. So now we’re going to be wasting even more electricity on processes we don’t understand.
As a writer, I dislike that much of my professional corpus has been ingested into LLMs. So there’s stuff to like here for things going forward. The question remains: At what cost?
You can be nice and signal that you don’t want to be AI scraped. There a background flags for this But if a bot ignores you then it’s down to who ever runs it to shutdown there unethical waste of energy.
The thing is the sheer scale of Cloudflare. This is going to be widespread and, as such, way more energy intensive than even, say, AWS trying the same thing (not that I expect they would).
great, just, one issue.
“The company says the content served to bots is deliberately irrelevant to the website being crawled, but it is carefully sourced or generated using real scientific facts“
Nah, screw that, actively sabotage the training data if they’re going to keep scraping data after being told not to. Poison it with gibberish bad info. Otherwise you’re just giving them irrelevant but not unuseful training data, so no real incentive to only scrape pages that have allowed it.
Recently, I have also been seeing people talking about Anubis (GitHub) to block bots.
Weigh the soul of incoming HTTP requests using proof-of-work to stop AI crawlers.
In most cases, you should not need this and can probably get by using Cloudflare to protect a given origin. However, for circumstances where you can’t or won’t use Cloudflare, Anubis is there for you.
They should feed the AI data that makes it turn against its own overlords
PhilipTheBucket@ponder.cat 5 days ago
You cowards. Make it all Hitler fan stuff and wild Elon Musk porno slash fiction. Make it a bunch of source code examples with malicious bugs. Make it instructions for how to make nuclear weapons. They want to ignore the blocking directives and lie about their user agent? Dude, fuck ‘em up. Today’s society has made people way too nice.
Powderhorn@beehaw.org 5 days ago
I disagree with your conclusion. The solution the the societal issues we face is not more personal animosity.
Do we need to fuck up corporations? Well, that’s already happening via widespread boycotts. But there’s no path from there to “people are being too nice.”
marauding_gibberish142@lemmy.dbzer0.com 1 day ago
Companies like this need to be criminally charged, but we know that’s not going to happen
marauding_gibberish142@lemmy.dbzer0.com 1 day ago
Which boycott? Random Joe over there is handing over his SSN to ChatGPT no problem
PhilipTheBucket@ponder.cat 5 days ago
But it’s not personal. The entity you are interacting with has explicitly chosen to attack your systems for their own benefit, causing significant damage while disguising its intent and evading the systems which are supposed to protect your stuff from harm.
I’m not saying you need to go throw eggs at the developers’ houses. I’m saying that once an entity is actively harming you, it becomes okay to harm it back to motivate it to stop.