> If it can solve certain biological challenges, it could build itself a tiny molecular laboratory and manufacture and release lethal bacteria. What that looks like is everybody on Earth falling over dead inside the same second.
If I put a chimp in a room with a red button labelled "launch nuke", there is a nonzero chance that catastrophe could arise. Will The Guardian publish me now?
These thought experiments presuppose it being able to synthesize mega-toxins from highly abundant resources with completely benign tooling and zero human supervision. If this is immediately possible in any capacity anyways, why did it take AI for us to find out?
"Prove me wrong" AI arguments need to get real. Show me the exploit in isolation, give me steps to reproduce it and maybe I'll change my tune.
Suffice to say they're AI researchers with access to advanced lab environments. If they told you the secret to getting sparks of AGI yourself, how would they fund their Large Language Collider?
Unfortunately this is not the first time the Guardian publishes such a hysterical “AI WILL KILL US ALL” piece, they publish several per week, and this is just the most ridiculous example. They appear firmly on-board with the moat-digging the big players are engaged in.
I cancelled my subscription some time ago due to the low quality reporting over AI (and some other subjects)
If I put a chimp in a room with a red button labelled "launch nuke", there is a nonzero chance that catastrophe could arise. Will The Guardian publish me now?
These thought experiments presuppose it being able to synthesize mega-toxins from highly abundant resources with completely benign tooling and zero human supervision. If this is immediately possible in any capacity anyways, why did it take AI for us to find out?
"Prove me wrong" AI arguments need to get real. Show me the exploit in isolation, give me steps to reproduce it and maybe I'll change my tune.