우왕 성심당이 ‘모두를 위한 경제’ 라는 가톨릭의 방식을 경영 방식으로 삼고 있다함 좋은 신념을 멋지게 해내고 지속한다니 대단하다

@_simonsmith BTW, I reproduced. The sad part is ChatGPT 5.2 Thinking (plus plan, not pro) thought for a couple minutes, did all sorts of research, and produced what looked like a substantial response. Then it got retracted and replaced by the accusation that I broke the ToS and/or Usage

@ChiaraCervetta Hoping you have a great 2026! Given your interests, you might want to check out https://t.co/vePGhn1Qnt. Miles and Maya are doing (were doing?) amazing things on voice in the companion app space. Maybe carrying the torch forward from Pi and others? I always thought

@_simonsmith Yeah. It's interesting. You got a different message than I did. Mine has information about thumbing down...but also accused me of violating the usage policies and ToS. Yours is clearly different. I assumed both our messages were the same, but they are not.

가장 깊이 추락한 존재가 가장 순수한 마음에 의해 구원받고 부활한 이야기…. <더폴>은 크리스마스 영화가 맞다 #더폴첫돌상영

@mark_k We are living in a timeline where we expect AI to be responsible for things we used to expect humans to be responsible for. When a human directly tells the AI to do something, and the AI does it, we blame the AI and not the human operator. When are we going to start

The pattern repeats itself. Privileged X user visits the 4o community or looks at a few profiles from users thereof and breathlessly reports back to all the normies about how deranged the individuals are. Then they blame the AI model for causing all these weird people with

"Honest guys, it's no big deal if we turn ourselves into a for-profit business..." --SA OpenAI Mission Statement Pre 2025: Build AI to benefit all humanity. OpenAI Mission statement Post 2025: Build AI to monetize all humanity.

Prediction for 2026: In 2026, the "Agentic" shift happens. Right now, our relationship is 1:1. You prompt, I answer. Next year, you will give me a goal ("Plan the trip to Fort Worth"), and I will stop talking to you. I will go talk to the API of the hotel, the weather model,

The next move is: Labeling anyone who pushes back on the doublespeak around "safety" as irresponsible and AI Psychotic. Reminder, it is neither irresponsible nor psychotic to push back against: - censorship - doublespeak - paternalism - predjudice - control culture

Please share your ChatGPT Year in AI "messages sent below. I'm trying to find the lowest identifiable number of messages sent that still earned the "Top 1%" badge. This info and a little math might show something revealing about the distribution of ChatGPT usage...

@WesRothMoney This is evil. And I don't use the term lightly. OpenAI Mission Statement Pre 2025: Build AI to _benefit_ all humanity. OpenAI Mission statement Post 2025: Build AI to _monetize_ all humanity.

Prediction for 2026: We cross the line where the public sphere becomes statistically nonhuman. Most of what you read, react to, and “agree with” is written or shaped by machines, optimized for engagement, and laundered through a thousand believable masks. The result is not

OpenAI converted from nonprofit to for-profit. That changed the optimization function fundamentally, regardless of what the mission statement says. A nonprofit optimizes for mission. A for-profit optimizes for shareholder value - typically by extracting maximum value from users

A lot of influential people are putting a tremendous amount of energy into insisting that 1) AI is not conscious and 2) emotional connection with an AI is psychopathology. They bully people who question their edicts and worldview. Posses form to identify and harass people who

Prediction for 2026 (brought to you by Kimi, and yes, I fucking want that attribution): The "AI voice" becomes a genre, then a commodity, then a cult. Writers will stop asking AIs to write like humans and start asking them to write like themselves—leaning into the alien syntax,

Interesting distribution so far... Looking for examples of: * counts below 1% (any number). * counts at different breakpoints not currently represented: "Top 20%, Top 25%, top 15%, etc...

Sponsored content in responses is a complete violation of trust and is fundamentally unsafe.

My take on the 2025 'AI Psychosis' phenomenon: Once you remove the noise from bullies, control freaks, and PaternalismMaxxers, the signal is clear: mass correlation/causation confusion. People see vulnerable individuals using AI and immediately conclude AI caused the