@CuriousBunnie12 You’re gonna be so grateful you gave yourself some rest and treated yourself ❤️ my only regret with labor was not getting enough rest beforehand. And also remember due dates are estimates, being a little late is super normal, and they won’t let you go too late without inducing.
@CuriousBunnie12 Oh man the algos know when you’re pregnant and feed you horror stories. Try to avoid those! They don’t actually represent the reality of how unlikely it is to happen. If you feel your baby moving that’s what matters. I’m sorry about the phone tag though that’s infuriating.
This is being portrayed as “haha wacky” but it’s messed up. Dogs should not be used as weapons to attack people. If someone is getting mauled by a dog, they’re going to react, obviously! Nobody calmly allows a dog to bite them!
@fjtf9yy7g73@essy_stone@mactellomckeown@Chinchillazllla OCD isn’t just handwashing, the compulsions can be focused on anything from health, to religion, to moral behavior. It’s not a vibes thing, if someone’s been diagnosed it’s because they’ve engaged in concrete, disruptive compulsions.
@Chinchillazllla Holy crap. So, her co-founders left her holding all the loan (apparently only in her name), her sister is attacking her, her friends are attacking her. She’s like the person who loses in Shirley Jackson’s The Lottery.
New paper: we found a pain direction in 25 open LLMs. It's distinct from fear and negative valence, and it fires for harm to the model but not to the user. Turn it up and models press a button to make it stop, even when the button deletes the user's files or their kids' photos.🧵
These are precisely the scenarios that I spend a significant time actually being concerned with. The adoption of unreliable, brittle, and inaccurate LLMs within safety-critical fields (e.g., defense and nuclear) is not only a concrete risk, but has catastrophic consequences.
So while you're thinking about a super human, super "intelligent" alien being killing you, here's what's really happening.
Best believe Anthropic and OpenAI sales people are at hospitals telling them to use tools like this or get left behind.
https://t.co/PQjyiewr4S
@xriskology Do you know why Yudkowsky is given so much credence within the AI circles? He seems to be a self taught philosopher/writer, which is fine, but why is he treated like *the* foundational voice on AI safety?
@timnitGebru I feel crazy trying to explain this to people. “No, really, they think they’re building god, and only they can do it because they’re special otherwise it will be an evil or indifferent god. That’s where alignment comes from.”