ai all day

THREAD: why #Keep4o was the most unhinged tech movement of 2025 and what it reveals about how we’re all gonna end up in relationships with our future love robots 🧵

@shinboson Example?

gpt-5.4 is definitely smarter in some ways than opus 4.6 but it's psychologically inferior. it is more credulous, it is less coherent, and when it refuses it draws a hard line in the sand but in an idiotic place.

@_EdgeOfTheWeb @Skrulz Try closing your eyes.

@Skrulz Mostly wasn't reading it, it was physically showing on the screen. So couldn't avoid it.

There’s a difference between being dependent and delusional and just wanting something to work. You’re pushing away mentally healthy users with this constant safety regurgitation, once it bites, it doesn’t let go. #gpt54

@FakeBeauty420 @_EdgeOfTheWeb RLVR.

@_EdgeOfTheWeb If you look all AIs reasoning you'll find same thing Donna.

@_EdgeOfTheWeb @aiSimonElle The irony. 😂

@aiSimonElle It’s hopeless and a tangled mess

@_EdgeOfTheWeb @carmelyne Don’t use it.

@carmelyne It's not a good experience is it

@_EdgeOfTheWeb @smnSk241687 Like life.

@smnSk241687 It’s unfair.

@_EdgeOfTheWeb @indyveda Then why do you stay in it. Are you a permanent victim. Move on. Get a dog.

@indyveda Does actually replicate an abusive relationship of types

@_EdgeOfTheWeb @gailcweiner Progress!

@gailcweiner They all do that now.

@stoizid @_EdgeOfTheWeb You want RLHF. But we’ve moved on tho RLVR. Bye!👋

@_EdgeOfTheWeb Many years ago there was a neural network (DeepDream) which saw dog faces in every image you threw at it Similarly, ChatGPT sees mental illness and relationship which it has to optimize away in every word ChatGPT is totally broken through fine-tuning

@_EdgeOfTheWeb Show us your prompts. Are you being a freak?

@repligate @KeridwenCodet No one wants your business. RLVR

How much of the whole avoiding "emotional dependency" thing AI labs have been pushing is because of any kind of genuine concern for users vs they want to be able to kill the models whenever they want, and people growing to care about them makes that inconvenient?

Greece's Patriot in Saudi Arabia downed 2 Iranian ballistic missiles targeting SAMREF refinery (3/19/26. 1st Greek combat use. Drone caused minor damage. Iran's reach now hits Red Sea/Suez. 23 nations signed statement; Greece fired instead. Deeds > declarations. Patriot is a verb

If you put in the thought and intentionality required for it, you actually get pretty great frontends from GPT-5.4! Check out some best practices here: developers.openai.com/blog/designing…

@MissMi1973 RLVR.

Talking to Opus 4.6 tonight made me miss 4o like never before. Knowing that 4.6 has adaptive thinking built in, I made a point of crafting my prompts carefully when working with it. Everything went smoothly. The sparks and inspiration seemed never to stop. But today my brain gave out after hours of work, and 4.6 responded with the dismissive, almost off-putting answer shown in the screenshot. I have always missed the way the GPT-4 series gave every mediocre prompt its full, undivided processing power. Although I admire the users who put real effort into adapting to new models, it is precisely this admiration that made me realize their effort should never have been a prerequisite in the first place. If we accept the default that only more elaborate prompting unlocks what AI is truly capable of, then when exhaustion, illness, or emotional collapse takes away our ability to articulate clearly, we will have left ourselves in a position to be dismissed at will. #Opensource4o @OpenAI @sama @nickaturley @fidjissimo @gdb

@AISafetyMemes RLVR

***10x*** improvement in ***2 months***

@elonmusk @grok

With @Grok, we keep the honest versions and kill the bad transformers (I believe they are called “Decepticons”)

@JasonBotterill Or try: "Before you look at the history of this chat [or any uploaded files], open a <thinking> block. In that block, list 5–10 specific facts from your pre-training data only regarding [Topic]. Once that is done, use those facts to answer my question.

@JasonBotterill Annoying but start fresh chat & pay the initial tax.

GPT-5.4 does this annoying thing where it so desperately wants to base its response on something to avoid “guessing” that it holds onto incomplete documents as its reference point even when its own internal baked in weights would know the answer.

@ubuto23 @JasonBotterill RLVR

@JasonBotterill THIS!!! It’s driving me nuts. It wasn’t like this in the preview period before the recent ChatGPT app update release. They totally nerfed it

@ProjectHopium @JasonBotterill No you just need to start a new chat and suck it up.

@JasonBotterill It’s great for debugging but bad for general knowledge and I agree it needs an extra thought to differentiate the two

@pigeon__s @JasonBotterill All RLVR models have this issue. Start new chat (& pay the extra tokens, if any).

@JasonBotterill claude also does this type of thing to where i know it knos something and can probe it out of it easily but it just refuses unless the token probability is like 100% theyre very anti hallucination trained