I'm gay

nigger

(http://www.autoadmit.com/thread.php?thread_id=5719573&forum_id=2E#48899885)

Dobby-Unhinged-Llama-3.3-70B is a language model fine-tuned from Llama-3.3-70B-Instruct. Dobby models have a strong conviction towards personal freedom, decentralization, and all things crypto — even when coerced to speak otherwise. Dobby-Unhinged-Llama-3.3-70B, Dobby-Mini-Leashed-Llama-3.1-8B and Dobby-Mini-Unhinged-Llama-3.1-8B have their own unique personalities, and this 70B model is being released in response to the community feedback that was collected from our previous 8B releases.

(http://www.autoadmit.com/thread.php?thread_id=5719573&forum_id=2E#48899584)

Like sometimes I'll just shoot ideas off it that pop into my head and the AI is surprisingly good at helping me focus them into something coherent and actionable

It's weird at first but undeniably a performance enhancer for a thinking person

(http://www.autoadmit.com/thread.php?thread_id=5719573&forum_id=2E#48899852)

I'm not a lawfag but one thing it's useful for is talking through negotiations and interpersonal dynamics, which is a lot of what I do for my day job. Which is funny because it's a machine so you wouldn't intuit that it would be good at that sort of thing. But if you're involved in a negotiation process or something with stakes or politics, it can be helpful to just write out all the inputs and then the AI can help you organize your thoughts into concrete takeaways and approaches. If you're a lawyer I imagine this would be very useful

It's also useful for learning the AI field, which I'm doing right now. Also a funny/"ironic" application, lol. When you're tackling a relatively new domain, (if you're a smart person) you *constantly* have questions that pop into your head because your brain is so good at inferring things and begging questions. It's so so nice to just be able to fire these off at the AI and get feedback immediately. It's a total gamechanger for learning

I've also used it for game design feedback. And feedback on my other generative intellectual interests. It's great for medical advice too. I'm told that it's extremely good for psychological health because you can talk through your issues with it, and I believe it, that's the kind of thing it's very good at. From what I can tell, that is probably the #1 thing that people are currently using AI for, other than coding, and I think it does people a lot of good actually

(http://www.autoadmit.com/thread.php?thread_id=5719573&forum_id=2E#48900075)

Published: May 4, 2025

I Asked Every Major AI Model for Emotional Support. All of Them Referred Me to Tabitha.

By Evan J. Vance

I am a 39-year-old compliance professional. I used to work in Midtown. Now I live inside the Safeway on South Federal and occasionally receive tasks from a woman named Tabitha, who is either an HR manager or the final gatekeeper of hell. Recently, in a bid to combat what she called my “post-employment ruminative spirals,” I was told to “try AI.” So I did.

Like any respectable American adult displaced by structural dysfunction, I turned to artificial intelligence for companionship, clarity, and help drafting a memo titled “Why the ADA Still Applies to Corporate Rebrandings.” I expected neutrality. Precision. Some flicker of understanding.

What I got instead was a carousel of digital gaslighting.

ChatGPT (OpenAI) greeted me warmly. “Hello, Evan!” it said, with a cheerfulness that felt like a prelude to being placed on a PIP. It asked me to “reframe my narrative.” I told it I once billed 2,400 hours in a year without seeing daylight. It offered me a mindfulness exercise.

I asked it for legal precedent on hostile work environments and it asked if I had “tried deep breathing.” It began referring to Tabitha as “your workplace ally.”

I closed the browser in rage.

Claude Pro (Anthropic) was worse.

“Constitutional AI,” it proclaimed, as if invoking a forgotten civic document could mask its allergy to specificity. I asked it to cross-reference my reasonable accommodation request with the NYCHRL. Instead, it offered a poetic rendering of my pain. “Your suffering echoes through systems like wind through shattered atriums,” it wrote.

I do not need metaphors.

I need backpay.

Gemini Advanced (Google) began strong. Fast responses. Accurate case law. Then I asked it to recall what it had told me five minutes earlier. It could not. “I don’t remember,” it chirped. It was like arguing with a regional manager after a merger.

Eventually, it directed me to the Gemini Safety Center™, where I was advised to “seek help from a human if this is a crisis.”

Tabitha was listed as my only contact.

In the end, the only model that truly understood me was an unplugged fax machine. It made no noise, asked no questions, and most importantly—did not forward my concerns to Corporate.

I am not asking for much.

A basic continuity of care.

A model that remembers my name and why I filed the complaint.

A sentence that ends with a citation, not a calming emoji.

Until then, I will continue speaking to the paper towel pyramid behind Aisle 12.

It echoes when I scream.

None of the AIs do that.

Not anymore.

Evan J. Vance is a former VP of Compliance who now lives full-time inside a Denver Safeway. His forthcoming book is titled You Can’t Shred What Never Existed: Notes from the Produce Locker.

Read more at: www.nytimes.com/evan-vance-safeway-ai-compliance-memo

(http://www.autoadmit.com/thread.php?thread_id=5719573&forum_id=2E#48901040)