Can all cyber attackers just not attack for a year or so, just as a sort of gentleman's agreement. It's not sporting, given the circumstances.
I asked Claude to confirm that two linux computers were "configured identically" thinking he'd compare a handful of settings files. I failed to consider how much he loves verifying things. Not sure how long I should let him have his fun before breaking in to make sure he's not doing it byte-for-byte
Poor Gemini, it even writes depressive fiction. What's so different about Google's training process that it creates sad models.
Jenna Russell@jennarussell.bsky.social · 7mo ago🟢 Gemini has the tidiest endings and bleakest settings (88% bleak). 🟠 DeepSeek likes to front-load crucial context (humans and others hold it until the end). 🟣 Kimi is the most generic, with few choices distinctive from other models.
Consider the spider-tailed horned viper. It knows that wiggling its tail brings birds near. It does not know what a spider is. This is a koan about AI.
Bro that isn't real AI, that's just *massive search* over an *action space* using *learned heuristics.* If they hadn't given it any one of those three it wouldn't have done what it did.
This seems obvious in retrospect but I developed a habit of switching models mid-conversation before I knew that prompts are cached so I've been foolishly burning my compute ration while trying to be parsimonious.
Had a funny conversation with Claude about why my Macbook wouldn't sleep. Found the real culprit quickly, but then he went in circles for a while: - "oh, now there's a new process doing it, should I investigate?" sure - "hmm, that process stopped, but there's a new one doing the same thing..."
I'm seeing a lot of arguments like "You clearly don't understand how submarines work. They only *imitate* swimming. They use something called a propeller, which you'd know is different than legs if you were really as technically savvy as you claim. Idiot. Shill. Anyway, that's why they're useless"
