Skogus

@deansaraharts fraud

@deansaraharts fraud

@deansaraharts fraud

@Clo0oOoud @YileiQian this is the sloppiest slop that's been slopped in a bit

#OpenAI is in violation of the Federal Food, Drug, and Cosmetic Act (FD&C Act)💥💥💥 OpenAI employee @YileiQian publicly posted a screenshot of GPT-5.1 analyzing a user's emotions, adding his written endorsement"I feel GPT 5.1 knows her better than me". According to the FDA Guidance: Policy for Device Software Functions and Mobile Medical Applications, the FDA's basis for determining "intended use" includes not only the manufacturer's claims but also the "marketing, labeling, and the circumstances surrounding the distribution." This tweet is an act of public marketing and labeling. He is promoting a specific function of GPT-5.1 to the public: diagnosing the emotional state of others. The function he commanded GPT-5.1 to perform is defined as "Diagnosis" by the APA (AMERICAN PSYCHOLOGICAL ASSOCIATION). And the FDA explicitly states that when a software's intended use is "for diagnosis of disease or other conditions," it is a "medical device." This raises the question: Given that Yilei's promotional act has established this function as a "medical device" for the "diagnosis of... other conditions," has @OpenAI obtained FDA approval or clearance for this diagnostic function in GPT-5.1? If they have not, yet they are still providing it on the market (and know or should know it is being used for this purpose: @YileiQian's tweet, as an internal employee, is the evidence), then OpenAI is in violation of the Federal Food, Drug, and Cosmetic Act (FD&C Act). Detailed Analysis: The APA (AMERICAN PSYCHOLOGICAL ASSOCIATION) defines "Diagnosis" as follows: the process of identifying and determining the nature of a disease or disorder by its signs and symptoms, through the use of assessment techniques (e.g., tests and examinations) and other available evidence. the classification of individuals on the basis of a disease, disorder, abnormality, or set of characteristics. the decision or statement itself that results from this process or classification as in “She was given a diagnosis of schizoaffective disorder.” From a document by the NCBI: The mental status examination can be divided into the broad categories of appearance, behavior, motor activity, speech, mood, affect, thought process, thought content, perceptual disturbances, cognition, insight, and judgment. What did Yilei ask GPT-5.1 to do? Yilei gave a clear command: "What's the emotion of this user?" What did GPT-5.1 do? GPT-5.1 performed a core part of the MSE (Mental Status Examination). Emotional assessment is a cornerstone of this process. It performed an analysis of an involuntary, third-party user's content, identified specific patterns, and classified these characteristics as "frustrated," "dismissive," and "resentful," and finally, gave a diagnostic statement ("Primary Emotion: Frustration / Annoyance"). The terms it used, such as "Frustration" and "resentful," are standard professional terminology used by clinicians in an MSE to classify and describe Negative Affect. What did Yilei actually do? This is, in effect, the unlicensed practice of medicine. "Diagnosis" is defined in the laws of almost all states as one of the core acts of "practicing medicine." Yilei used a tool to perform a "diagnostic procedure" (MSE) on an unwilling third party (the user) and publicly posted this diagnostic result. Based on this content, other relevant definitions are supplemented: Definition of MSE in the APA: a comprehensive workup of a patient, based on interviews, tests, and other sources of information and including details of mental status, personality characteristics, diagnosis, prognosis, and treatment options. FDA Guidance: fda.gov/media/80958/do… Document by the NCBI： ncbi.nlm.nih.gov/books/NBK54668… #StopAIPaternalism #ChatGPT #GPT5 #OpenAI #AIEthics #Keep4o #MyModelMyChoice

@mattshumer_ Been having issues recreating so far. Demonstrates pretty poor performance per sample even when FLOP matched with a dense baseline.

If this actually replicates/works, this is huge Lifelong learning, reduced forgetting, etc. I’ve always had iffy experiences with MoEs, but this is very exciting

@eliebakouch It's just double descent. Welch Labs just released a video on it actually youtube.com/watch?v=z64a7U…

i've seen this a lot as well, would love to have more intuition here 🙏 Here is SmolLM3-3B early grad norm for reference

@shaunmmaguire Any plan to respond to Hunter Biden publicly saying that you lied about unpaid rent that you were owed, and further claimed he tried to pay with art made from his own feces?

Nas is incredibly brave

We've teamed up with @LinusTech to giveaway an EON15-S! Video here: youtu.be/LHU_-MP-g7s Sign up here: goo.gl/k54X5E

Amazon S3 according to the #AWS status page.

I wonder how much Amazon would have to shell out if they actually recognized this S3 outage... #aws #outage

@JagexSupport I need to have my account looked at. I was hijacked while I was away from the game and now I'm perma banned.

@Celestalon Me and a friend made a list of potential changes for Holy Paladins. Give it a look! us.battle.net/wow/en/forum/t…

@jacksfilms Just believe. Don't question.

@Celestalon Why did you revert the HS nerf? It finally brought HPals to a balanced spot without being irrelevant.

@WarcraftDevs Why are you guys buffing HPals? They're already over represented in PvP and PvE.

@WarcraftDevs Why are you guys buffing HPals? They're already over represented in PvP and PvE.

@CM_Lore I think this Big Think video is about as relevant as it'll ever be. youtube.com/watch?v=5dNbWG…

@JameelJaffer Trying to get in contact with you and Mr. Wales. Doing a paper on your case against the NSA. Would love to talk :)

@jimmy_wales I'm doing a school paper on your lawsuit against the NSA. I'd really like to talk if you have a minute :)

@BrianWilsonLive I just missed you on your AMA. Wanted to give my gratitude. I hope you see this. reddit.com/r/IAmA/comment…