www.anthropic.com/research/glm...
Unlike Mythos, I might actually have a chance at using GLM for defensive research. Anthropic didn't reply to any of my requests for Mythos access for use on OpenSSH.
www.anthropic.com/research/glm...
Unlike Mythos, I might actually have a chance at using GLM for defensive research. Anthropic didn't reply to any of my requests for Mythos access for use on OpenSSH.
- I̶n̶v̶e̶n̶t̶ ̶a̶u̶t̶o̶n̶o̶m̶o̶u̶s̶ ̶h̶a̶c̶k̶i̶n̶g̶ ̶m̶a̶c̶h̶i̶n̶e̶
- G̶o̶o̶g̶l̶e̶ ̶"̶s̶a̶n̶d̶b̶o̶x̶i̶n̶g̶"̶
- G̶o̶o̶g̶l̶e̶ ̶"̶s̶c̶o̶p̶e̶d̶ ̶A̶P̶I̶ ̶k̶e̶y̶s̶"̶
- Google "ACLs"
- Google "egress controls"
- Google "Oh Wasps Top 10"
- I̶n̶v̶e̶n̶t̶ ̶a̶u̶t̶o̶n̶o̶m̶o̶u̶s̶ ̶h̶a̶c̶k̶i̶n̶g̶ ̶m̶a̶c̶h̶i̶n̶e̶
- G̶o̶o̶g̶l̶e̶ ̶"̶s̶a̶n̶d̶b̶o̶x̶i̶n̶g̶"̶
- G̶o̶o̶g̶l̶e̶ ̶"̶s̶c̶o̶p̶e̶d̶ ̶A̶P̶I̶ ̶k̶e̶y̶s̶"̶
- Google "ACLs"
- Google "egress controls"
- Google "Oh Wasps Top 10"
Philosophers with a tenuous grasp of security are dominating their security discussion and strategy.
Please stop trying to scare people with a scenario involving transmitting a trillion parameter model across an air-gap with a channel that is less than 1 baud. Orders of magnitude less.
Philosophers with a tenuous grasp of security are dominating their security discussion and strategy.
Signal could help bring solid E2EE messaging to far more products. Small and midsize businesses can’t afford a commercial licence, while MLS remains extraordinarily complex to deploy.
Signal could help bring solid E2EE messaging to far more products. Small and midsize businesses can’t afford a commercial licence, while MLS remains extraordinarily complex to deploy.
There are going to be some painful & expensive lessons learned over the next 6–12 months. Genuinely innovative AI security products are already being sold, or expected to perform, as silver bullets for problems that require strong foundational security.
There are going to be some painful & expensive lessons learned over the next 6–12 months. Genuinely innovative AI security products are already being sold, or expected to perform, as silver bullets for problems that require strong foundational security.
youtu.be/UkxRa1M_x9A?...
youtu.be/UkxRa1M_x9A?...
I'm getting far more value for my time from ChatGPT and even Proton Lumo is better than Claude right now.
I'm getting far more value for my time from ChatGPT and even Proton Lumo is better than Claude right now.
Employees submitting confidential emails/docs to AI detectors to catch colleague bot usage. Services forward to 3+ AI companies for watermark checks.
Nice way to turn office paranoia into a multi-vendor data leak! 🎉
Employees submitting confidential emails/docs to AI detectors to catch colleague bot usage. Services forward to 3+ AI companies for watermark checks.
Nice way to turn office paranoia into a multi-vendor data leak! 🎉
conic.al/writing/the-...
conic.al/writing/the-...
conic.al/writing/the-...
conic.al/writing/the-...
youtu.be/87DyyMV0kCY?...
youtu.be/87DyyMV0kCY?...