Pleased to share a page and explainer for the AI tarpit project Science is Poetry, with legal statement, rationale(s), and a few deployment notes:
-
@balslev One would think so, but they largely seem to be automated. For almost 3 days straight now I have Anthropic, OpenAI and Meta trapped, Google for 2. While some IPs may drop off, crawlers return from another endpoint. The ontent to them seems good to go, as at the present time it ticks their boxes.
Further, each new URL of the proj added to a site is a new hook, so unless they block on IP, they're back in.
If they do blacklist, the VM can be moved to a new host, content regenerated.
@JulianOliver Fun fact: When I ask http://www.chat.dk how many html-pages www.scienceispoetry.net/noodles consists of, it answers "10".
-
@JulianOliver Fun fact: When I ask http://www.chat.dk how many html-pages www.scienceispoetry.net/noodles consists of, it answers "10".
@balslev Hehe. Sweet delusion!
-
R relay@relay.infosec.exchange shared this topic on
-
Pleased to share a page and explainer for the AI tarpit project Science is Poetry, with legal statement, rationale(s), and a few deployment notes:
The page may grow a bit. Just wanted to get it out the door.
@JulianOliver Not that my little blog would really contribute much to the traffic, but how much traffic can you handle? Is there a way to self-host like with Nepenthes? I worry that you are potentially taking on a massive load for the rest of us.
-
@JulianOliver Not that my little blog would really contribute much to the traffic, but how much traffic can you handle? Is there a way to self-host like with Nepenthes? I worry that you are potentially taking on a massive load for the rest of us.
@cybervegan Thank you very much, very kind. It all resides in a Debian GNU/Linux qcow2 image but I may repackage it for Alpine such that it could be more readily distributable, then quick to plug into a reverse proxy (Caddy, Nginx, Apache2).
However right now things are in good shape so far as load. It is on a very powerful server and not even near 15% of all cores, plenty or phys mem available too. So it can take on a lot more before a pressing need for more instances.
-
@cybervegan Thank you very much, very kind. It all resides in a Debian GNU/Linux qcow2 image but I may repackage it for Alpine such that it could be more readily distributable, then quick to plug into a reverse proxy (Caddy, Nginx, Apache2).
However right now things are in good shape so far as load. It is on a very powerful server and not even near 15% of all cores, plenty or phys mem available too. So it can take on a lot more before a pressing need for more instances.
@JulianOliver Ok, appreciate your work, and the resources you are putting into it - bandwidth and electricity is not free even if you had been gifted the hardware. I suspect it will turn into an arms race, but I'm sure you're more than aware of that!
I'm running my static HTML blog on FreeBSD, on a microserver, using nginx, so a qcow2 image probably wouldn't be practical for me - I know that's a bit rare, but web server is a secondary function of this machine. Looks like Qemu is available for FBSD, so maybe I could go that route (when I've managed to get out of autistic burnout that is).
Thanks once again. I'm going to try to add a link to SIP to my blog template and see what happens. My access.log has been filling up with AI bots, so will be interesting to see how much it affects that.
-
@JulianOliver Ok, appreciate your work, and the resources you are putting into it - bandwidth and electricity is not free even if you had been gifted the hardware. I suspect it will turn into an arms race, but I'm sure you're more than aware of that!
I'm running my static HTML blog on FreeBSD, on a microserver, using nginx, so a qcow2 image probably wouldn't be practical for me - I know that's a bit rare, but web server is a secondary function of this machine. Looks like Qemu is available for FBSD, so maybe I could go that route (when I've managed to get out of autistic burnout that is).
Thanks once again. I'm going to try to add a link to SIP to my blog template and see what happens. My access.log has been filling up with AI bots, so will be interesting to see how much it affects that.
@cybervegan FreeBSD is very nice.
Qemu/KVM with host passthrough mode for your hardware would be the first thing to look into.
I'm paying for the server out of my own quiver but may look to take donations for it at some point.
(can relate to the brain-ness)
-
@cybervegan FreeBSD is very nice.
Qemu/KVM with host passthrough mode for your hardware would be the first thing to look into.
I'm paying for the server out of my own quiver but may look to take donations for it at some point.
(can relate to the brain-ness)
@JulianOliver You definitely should!
-
@adam_harvey @asrg@tldr.nettime.org Oh dear. I'll let you guys sort this out!
-
If you're interested in learning more about implementations of resistance in this era of unchecked Big AI, direct action strategies and the techno-politics therein, be sure to check out ASRG's site (https://algorithmic-sabotage.gitlab.io/asrg/) and give them a follow here on Mastodon (@asrg).
They've put a lot of heartbeats and neurons - human stuff - into this area.
A newcomer frantically lost in the Caves of Babble.
----
3.215.221.125 - - [16/Apr/2026:06:14:25 +0200] "GET /noodles/images/primigenous/orchiepididymitis/Lord/havent.png HTTP/1.1" 200 111459 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot) Chrome/119.0.6045.214 Safari/537.36" "-"
---- -
A newcomer frantically lost in the Caves of Babble.
----
3.215.221.125 - - [16/Apr/2026:06:14:25 +0200] "GET /noodles/images/primigenous/orchiepididymitis/Lord/havent.png HTTP/1.1" 200 111459 "-" "Mozilla/5.0 AppleWebKit/537.36 (KHTML, like Gecko; compatible; Amazonbot/0.1; +https://developer.amazon.com/support/amazonbot) Chrome/119.0.6045.214 Safari/537.36" "-"
----Do you have an unused domain that you would be happy to donate to a counter-offensive against unchecked & unregulated AI crawlers that scrape human-made content to simulate & deceive for profit?
If so, pls reply to this post. Your domain would become an entrypoint to the AI tarpit & Poison-as-a-Service project below, allowing concerned public to choose to use it on their sites, helping make the project more resilient to blacklisting.
-
-
Do you have an unused domain that you would be happy to donate to a counter-offensive against unchecked & unregulated AI crawlers that scrape human-made content to simulate & deceive for profit?
If so, pls reply to this post. Your domain would become an entrypoint to the AI tarpit & Poison-as-a-Service project below, allowing concerned public to choose to use it on their sites, helping make the project more resilient to blacklisting.
@JulianOliver do they have to be root domains? Could I set up a subdomain or two for you?
-
@JulianOliver do they have to be root domains? Could I set up a subdomain or two for you?
@narthur A subdomain is just fine, yes!
-
@lightweight Fantastic, thanks Dave!
A: 95.216.76.85
AAAA: 2a01:4f9:2b:c83::2 -
@narthur A subdomain is just fine, yes!
@JulianOliver ok. So what do I do? Just point them at those IPs?
If they all share the same IP, won’t that make them easy to block?
-
@JulianOliver ok. So what do I do? Just point them at those IPs?
If they all share the same IP, won’t that make them easy to block?
@JulianOliver poetry.rainskit.com and poetry.narthur.com. But the 'http' versions just give an nginx welcome page, and the https versions don't have valid SSL. Maybe you want CNAMEs instead? Or reverse proxies? And do you need some other page to link to those domains?
-
@JulianOliver Just added a pair of records for a .eu domain. I'll let it lapse in August; it can do some good work until then.
-
@JulianOliver Just added a pair of records for a .eu domain. I'll let it lapse in August; it can do some good work until then.
@twilliability Great, thanks a lot!
-
@JulianOliver ok. So what do I do? Just point them at those IPs?
If they all share the same IP, won’t that make them easy to block?
@narthur For now they would all share the same IP.
Both domain and IP are naturally able to be filtered at the crawler end, but as numerous sites can be hosted behind one IP, it is my belief that they will drop a domain first.
Further, it's my hope to have instances of the project running on other dedicated hosts down the road.
-
Do you have an unused domain that you would be happy to donate to a counter-offensive against unchecked & unregulated AI crawlers that scrape human-made content to simulate & deceive for profit?
If so, pls reply to this post. Your domain would become an entrypoint to the AI tarpit & Poison-as-a-Service project below, allowing concerned public to choose to use it on their sites, helping make the project more resilient to blacklisting.
@JulianOliver im down. What's the first step?
