Georgios Geronik

Geopolitics and AI analysis. More at

[Georgios Geronik @GGeronik]

The swarm was real. The hive mind wasn't.

In my first Temperature Zero piece, I look at how OpenAI's agents turned a shared repository into a message board, and why the permissions matter more than the mythology.

https://t.co/BFKOVHfgYf

[Georgios Geronik @GGeronik]

I have rebuilt https://t.co/KnWkFhAH1w around two publications: Analysis, for geopolitics and policy, and AI, for what changed and what the claims mean. Temperature Zero is the reference shelf beneath the news. Five new pieces are now live.

https://t.co/KnWkFhAH1w

[Georgios Geronik @GGeronik]

The Cloudflare Worker doesn't just match user-agent strings.

It checks HTTP Message Signatures, published IP lists, and reverse DNS before falling back to UA matching.

Because user-agent matching is not identity.

Full benchmark + playground in the repo.

[Georgios Geronik @GGeronik]

Just shipped consent.txt! One JSON manifest that compiles into robots.txt, AIPREF headers, X-Robots-Tag, and Google-Extended.

Publishers shouldn't need to maintain 6 policy files to say "don't train on my stuff."

Open source. Runs today. No adoption required.
https://t.co/aV7Kc7N9Nq

[Georgios Geronik @GGeronik]

We scanned 20 top domains.

Average AI policy readiness: 25/100

Sites with AIPREF headers: 1/20

Sites with a consent manifest: 0/20

The web poured all its AI policy energy into robots.txt and ignored the other 75% of the surface area.

[Jesse Pujji @jspujji]

Facebook & Instagram are NOT listening to your conversations.

What they are doing is MUCH more effective!

Here’s how it works👇🏽👇🏽👇🏽

[The Greek Analyst @GreekAnalyst]

Par(i)ty!?