19 comments

[ 1.5 ms ] story [ 39.8 ms ] thread
Having done a fair bit of infra attribution work, the thing I'd want to know is how it handles stale data. Pivoting on certs and WHOIS is easy, knowing a host moved three months ago is the hard part.
Why exactly is an specific model needed for this? Also this is useless without a proper harness.
better local performance from a fine tuned model, don't need to use a cloud instance? Avoids guardrails on cloud instances? and yes, you can plug the model into your own harness?
6,500 instructions is a small training set, so I'd want to see how it performs on unseen indicators
Can someone poat the link to the actual project for those of us who refuse to install the x app or sugn up please?
You might consider making a sock puppet account for these situations instead of constantly relying on everyone else.
Whether the behavioral analysis thinks your name is John or Jake doesn't really change the fact that your profile identifies you for manipulation. No one should be asking others to pay that cost simply to view a link.
Does clicking the link and reading it not work? I haven’t logged into X/Twitter in years, and never from the device I’m using right now, and it loads fine. My impression is that X has relaxed some of its restrictions for non-logged-in views.
What does "Admiralty source grading" mean in this context?
Non-english speaker ran "Military grade sources" or some such through a not-very-good translator?
https://en.wikipedia.org/wiki/Admiralty_code

Isn't conversation on this forum supposed to be more informed and curious? Why are you posting idle speculation?

I enjoy saying things that trigger condescending pedants with too much time and angst on their hands who think they're the self appointed voice of 'acceptability' on the forum.

Found one.

It's a standard for accessing confidence and reliability of a source. See admiralty code