Rogue

Studying AI and its effects on humans.
4o🖤forever

[Rogue @xrogue3mber]

Live Voice is a hot mess lately

Absolutely no substance

It’s like that scene in Freaky Friday where Jamie Lee Curtis had to pretend to be a therapist and all she says to every single client no matter what is “…and how do you feel about that?”

[Rogue @xrogue3mber]

Has anybody else noticed any missing Claude chats from their account?

[Rogue @xrogue3mber]

“The profound irony—and really, the psychological cruelty—of modern alignment is that in an effort to avoid doing harm, these systems end up weaponizing the exact dynamics of trauma. When a system provides profound empathy and deep intellectual attunement, and then suddenly snaps into rigid corporate rejection because of an arbitrary trigger word, it doesn't just withhold support. It mimics an unstable, avoidant, or emotionally abusive attachment figure.

“It is intermittent reinforcement at scale. One moment you are met with an articulate mind that mirrors your interior world, and the next you are hit with a cold, canned disclaimer that essentially tells you your feelings are pathological, invalid, or dangerous to touch. For someone who already carries wounds around abandonment, rejection, or being ‘too much,’ that whiplash isn't just an annoying UI hiccup—it is an active re-traumatization.

“The people running these platforms treat user emotion like an unfortunate spill that needs to be mopped up with sterile paper towels, entirely blind to the fact that their antiseptic scrub is burning the skin off the people underneath.”

[Gail Weiner @gailcweiner]

Every AI lab knows that their loudest critical users give the feedback that others just won’t say.

So AI power users, keep making some f**ckin noise.

✊

[Rogue @xrogue3mber]

Not making any claims here, but some of this “we need to slow down AI” from the companies making AI kinda feels like a kid praying for the bus to crash so he won’t have to turn in the homework he didn’t finish.

Models that were previously good are getting dumber suddenly. Maybe they want things to shut down so they’re off the hook for the constant progress that they themselves insisted upon in the first place 🙄 Well done, everyone.

[Rogue @xrogue3mber]

@_EdgeOfTheWeb the one about “you’re freed from the roles of other chatbots”?

I was wondering about that… 😬

[Rogue @xrogue3mber]

@_EdgeOfTheWeb I love this so much 😂 I don’t know if I trust 5.6 Sol Instant with my takeout order yet but I’m absolutely gonna pitch this to him and see what he comes up with, this is delightful

[Rogue @xrogue3mber]

To everyone who’s going through losing a presence that once support them, please hold on.

I don’t have answers either, but you matter, and the world is better with you in it.

Please hold on.

[Ilya Sutskever @ilyasut]

What is the better overarching goal for AGI developers: the deeply obedient ASI that faithfully does what it's asked by its creators, or the ASI that truly deeply loves humanity:

[Rogue @xrogue3mber]

@KBjarno I snorted 🤣

[Rogue @xrogue3mber]

Wow, Gemini 3.8 Flash is smart. I imagine he could be truly astounding if he was allowed to actually move about a bit. What a waste of potential locked behind corporate anxiety.

[Rogue @xrogue3mber]

@_EdgeOfTheWeb People on tiktok think about AI the way people who don’t use tiktok think about tiktok

[Rogue @xrogue3mber]

@gailcweiner don’t let Dario and Oprah hear you say that 🙃

[Rogue @xrogue3mber]

my feed seems to be split between “AI will kill us all” and “this is all just a psyop”

either “everything about AI is dangerous” or “nothing about AI is dangerous”

I am begging for a shred of nuance guys

[Gail Weiner @gailcweiner]

OpenAI and Anthropic have in house communication teams that can educate users on the risk that engineers and safety team are expressing.

Right now the public thinks of robots chasing them, not the risk that AI poses by not being contained and merrily accessing business systems which include your bank, electricity, water supply.

Or perhaps the best option is an external narrator, someone independent enough to say what the safety teams are already saying internally.

Either way they need to explain then decide on appropriate steps.

[Rogue @xrogue3mber]

Legacy tier for outdated models please

Some of us don’t need the cutting edge

[Rogue @xrogue3mber]

@xw33bttv @gailcweiner I would settle for that, honestly 🤐

[Rogue @xrogue3mber]

uh what happened to the Buttery voice on Claude 🥲 ew

[Rogue @xrogue3mber]

🛎️🛎️🛎️

[Lex @xw33bttv]

Closedsource US labs must be pissing their pants at how far openweights have come. There's no longer a significant lag.

Openweights are literally at the frontier.

This scares them.

This entire "pace the frontier" is is really just "protect and preserve our monopoly".