ashten,

today i learned, i can use cloudflare to block AI scrapers

so, woem.men users, rest peacefully. your content (hopefully) wont be scraped by AI!

18+ niko,
@niko@furry.engineer avatar

@ashten i went a lil bit more fun

set $redir_to_gz 1;<br></br>if ($host = gz.niko.lgbt) {<br></br>    set $redir_to_gz 0;<br></br>}<br></br>if ($user_agent !~* (claudebot|ZoominfoBot|GPTBot|SeznamBot|DotBot|Amazonbot|DataForSeoBot|2ip) {<br></br>    set $redir_to_gz 0;<br></br>}<br></br>if ($redir_to_gz) {<br></br>    return 301 https://gz.niko.lgbt/;<br></br>}<br></br>

basically tosses all the AI bots i've seen into gz.niko.lgbt (click at own risk, decompresses 100GB file in memory :3) with a compression type of gzip so it has to be decompressed to see any "content" (there is none)

notfire,
@notfire@fedi.notfire.cc avatar

how could one go about hosting this

edit: saw it before the edit.

niko,
@niko@furry.engineer avatar

@notfire i just edited it with the two major setup things behind gz.niko.lgbt (don't forget the gunzip off if you host it yourself that way nginx doesn't blow its foot off if the client doesn't support gzip)
outside of those two it's just a completely normal nginx server block

notfire,
@notfire@fedi.notfire.cc avatar

@niko i wonder if you could use fuse or something to generate an infinitely big file on the fly?

niko,
@niko@furry.engineer avatar

@notfire depending on the HTTP client the other end uses the client might either decompress on the fly (not a single one i've tested does that) or download the entire body and then decompress (curl, firefox, and chromium do this)
it would work for the former but the latter would just eventually end up having a connection closed thanks to some other reason likely

notfire,
@notfire@fedi.notfire.cc avatar

@niko yeah i meant for either of them really, just serve infinite data till the connection gets broken to waste as much bandwidth and time as possible for companies' ai bots

niko,
@niko@furry.engineer avatar

@notfire i would think an OOM condition on a server thanks to uncontrolled gzip decompression would have slightly more negative impacts which honestly is half the point with this otherwise i'd do what you suggested yeah

18+ mstar,

@niko @ashten could also redirect to mstar.evilthings.de/cat/awawawa. Just responds with a permanent awawa stream that only ends once you cut the connection on your end

18+ niko,
@niko@furry.engineer avatar

@mstar @notfire see post above for your infinite data option :)

as for why i didn't do this i figure a tiny innocent 100MB payload that decompresses in memory to 100GB is slightly more explosive

  • All
  • Subscribed
  • Moderated
  • Favorites
  • random
  • rosin
  • osvaldo12
  • modclub
  • DreamBathrooms
  • khanakhh
  • magazineikmin
  • thenastyranch
  • Youngstown
  • everett
  • InstantRegret
  • slotface
  • ngwrru68w68
  • kavyap
  • tacticalgear
  • JUstTest
  • GTA5RPClips
  • Leos
  • Durango
  • mdbf
  • cisconetworking
  • normalnudes
  • ethstaker
  • cubers
  • provamag3
  • tester
  • anitta
  • megavids
  • lostlight
  • All magazines