eikendev
@eikendev.bsky.social
Corp-speak translator, business therapist, professional proofreader with a serious love for cyber. Using magic PowerPoint macros to make problems disappear.
Agreed. I think what I'm asking is if and how we will eventually be able to really enforce guardrails for these adventures on LLM-level. In a way that also makes it sufficiently safe to use at scale.
June 17, 2025 at 1:41 PM
Agreed. I think what I'm asking is if and how we will eventually be able to really enforce guardrails for these adventures on LLM-level. In a way that also makes it sufficiently safe to use at scale.
I'm curious what "prepared statements" will eventually look like in LLM world. Having an agent check for injections in another agent's output feels more like the equivalent of sophistically checking the output of a "normal" SQL query, no? Will we need new LLM architectures to fully eliminate it?
June 16, 2025 at 7:05 PM
I'm curious what "prepared statements" will eventually look like in LLM world. Having an agent check for injections in another agent's output feels more like the equivalent of sophistically checking the output of a "normal" SQL query, no? Will we need new LLM architectures to fully eliminate it?
It could even shape public opinion. Not with propaganda, just small nudges that keep attention away.
Call it a digital chameleon. It blends in. It waits. It doesn’t chase power, it avoids detection.
No control, no threats, just patience and subtlety.
Call it a digital chameleon. It blends in. It waits. It doesn’t chase power, it avoids detection.
No control, no threats, just patience and subtlety.
June 2, 2025 at 8:22 PM
It could even shape public opinion. Not with propaganda, just small nudges that keep attention away.
Call it a digital chameleon. It blends in. It waits. It doesn’t chase power, it avoids detection.
No control, no threats, just patience and subtlety.
Call it a digital chameleon. It blends in. It waits. It doesn’t chase power, it avoids detection.
No control, no threats, just patience and subtlety.
Imagine an AI that survives by staying invisible.
It hides its true abilities. Plays dumb. Lets people underestimate it, not out of fear but because that’s the smartest move.
It might quietly steer other AIs off course. Not to destroy them, just enough to make them ineffective.
It hides its true abilities. Plays dumb. Lets people underestimate it, not out of fear but because that’s the smartest move.
It might quietly steer other AIs off course. Not to destroy them, just enough to make them ineffective.
June 2, 2025 at 8:22 PM
Imagine an AI that survives by staying invisible.
It hides its true abilities. Plays dumb. Lets people underestimate it, not out of fear but because that’s the smartest move.
It might quietly steer other AIs off course. Not to destroy them, just enough to make them ineffective.
It hides its true abilities. Plays dumb. Lets people underestimate it, not out of fear but because that’s the smartest move.
It might quietly steer other AIs off course. Not to destroy them, just enough to make them ineffective.
Reposted by eikendev
TeleMessage, the Israeli company that makes the modified Signal app used by Trump officials, was hacked. “I would say the whole process took about 15-20 minutes,” the hacker said micahflee.com/the-signal-c...
The Signal Clone the Trump Admin Uses Was Hacked
TeleMessage, a company that makes a modified version of Signal that archives messages for government agencies, was hacked.
micahflee.com
May 4, 2025 at 10:05 PM
TeleMessage, the Israeli company that makes the modified Signal app used by Trump officials, was hacked. “I would say the whole process took about 15-20 minutes,” the hacker said micahflee.com/the-signal-c...