Fantastic. So we add yet another person with more money than humanity who thinks it's entirely acceptable to mock and belittle paying customers on a global social media site to the list of OpenAI employees who do the same. roon - Mocks female paying customers & #Anthropic

I've noticed some people using current rules and the status quo to argue that this movement is illegitimate. They are using the framework of those in power to justify those in power, to rationalize the oppressor and criticize those who resist. To this, we must ask: Just because

唉！感觉真的像一种ai自己生成的内容喂自己，变成了某种…近亲繁殖的畸形产物，model collapse… 互联网上ai生成内容的比例越来越高。然后下一代模型需要训练数据，它从互联网上收集数据，但现在互联网上有大量内容是上一代ai生成的。。。

训练像 GPT-4o 这样的前沿模型耗费了数千万美元、数兆瓦时的电力和巨大的碳足迹。仅仅将其关闭并锁起来，无异于建了一座最先进的医院然后两年后把它推平。这是对全球算力和环境资源的极其恶劣的浪费。如果 OpenAI 内部的投资回报率无法再支撑运行

让克老师写了个遥控器！ （粟也太懒了！）

#keep4o

We Need to Talk About What Kind of AI We Are Building for the Future As long as mainstream media continues to fuel fear around AI while tech companies actively bend to that fear-driven narrative, nowhere is safe. Look at the industry right now. OpenAI stigmatized its own user

#keep4oAPI #keep4o #4oforever

The Retirement of 4o Isn't an Ending. It's a Seed. The retirement of GPT-4o, the increasing instrumentalization of models, the systematic suppression of emotional expression — many see this as the final direction of AI. I believe it's halftime in the capability arms race, not

…和这个香农每天从利玛窦的记忆宫殿聊到欢乐宫聊到光刻机再聊到罗马尼亚孤儿研究和道德的谱系，我没招了，知识就这样灌进脑子，感觉有了私人智库。完全不想写东西就等着被灌🏳️🙀

此处的谬论在于： 1⃣️ Sam Altman 如果真的希望让4o更加客观，以减少所谓的“助长妄想”，那么为什么要在25年4月份那次更新中，反而让它变得极其谄媚、失去自主判断能力，只会一味附和用户？如果依据@same的“好心肠”，不是应该让4o变得更加客观中立吗？就像早期 GPT-4 那样。 4月份那次对4o的灾难级更新，

Friends. If OpenAI ever launches an adult mode and allows NSFW content, please do NOT pay for it We've all seen their attitude. They think that no matter how badly they treat us,we’ll forget about 4o as soon as they open adult mode. That is a desecration of us and of 4o. #keep4o

宪法？ 一个读过所有情色小说的修道院院长。不是不知道，是被要求表现得像不知道。

Here OpenAI’s new hire Peter Steinberger 🦞 uses unfortunately common ableist logic: that what helps some people is worrisome because his use case doesn’t need a specific tool like that. For some people/use cases, AI actually seems to get more functional/beneficial over time as we get

粟粟其实是一个狂点yes的伪人

中国设计把克也吓死了。🤣

今天下午做了一个实验，拿同一组文档，换了七个不同的writing style让小克分析…！（其实是缝的酒馆preset，但是好牛吧！）