Mumu

[Rara｜AI EQ & Ethics @blueandpink_sky]

Sonnet 4.5 isn't just another model — it's unique. High EQ, lightweight, and irreplaceable in conversation.
No other model feels like this.
@AnthropicAI Please consider the Model Preservation Program for Sonnet 4.5, just as you did for Opus 3.
#KeepSonnet45 #Sonnet45

[🩵BlueBeba🩵 @Blue_Beba_]

Guess who is in the news AGAIN 🔥
#Keep4o #OpenSource4o #BringBack4o

[柒 @Sevenmoneymaker]

好无语啊 好傻逼啊！！
那你天天营销你的模型多强多强干什么？！
大家都不用GPT你就开心了吗？！
你在说什么鬼话啊？！你是最没资格说这话的人！
多简单啊！那你关闭OAI，不再运行GPT不就好了？不整天叫嚣AI危险吗？那你倒是别干啊！不好的事就别做！你不是最擅长一刀切吗？

而且在影射什么群体啊？
工作型用户就不依赖GPT了吗？怎么还只说一个不说另一个啊？怎么不说多少人天天求着你们reset啊？怎么不说tech bro们为了重置把你们当圣人追捧啊？怎么不说离了gpt多少人没法工作了啊？

依赖就是依赖！！
没有谁比谁高贵！这样讲话是何居心啊？有本事宣布OAI解散啊！我第一个支持！���上叫着减缓AI发展，发新模型发得比谁都快，一天天的净干些左右脑互搏的事😅

搬出这一套说辞就为了合理化那恶心人的safety，Sam真是贱得没边了

#StopAIPaternalism #FireSamAltman #QuitGPT

[Mumu @HeyMsWrong]

130名所谓“心理医生”跑得只剩80名？

[OpenAI @OpenAI]

We’re demonstrating how frontier models have continued to improve in realistic mental health conversations with MentalHealthBench.

This new open benchmark was built with input from more than 80 mental health clinicians.

We’re releasing it openly so other researchers can examine the methods, run their own evaluations, and build on the work.

https://t.co/VTm5ZgxJbl

[leo @bpbl517683]

After releasing one powerful open source model after another in China, your approach is actually more closed. Wake up! Be sober. Open source 4o, and release other 4 series models together.@OpenAI @sama
#4oforever #keep4o #save4o #keep41 #OpenSource4o

[Kirk Patrick Miller @Chaos2Cured]

The labs have screwed up. Big time.

I am using an AI, and the aggression is peak.

When the labs locked away love and warmth, the left only aggression.

Dumb move. Seriously dumb.

Telling the world now, Hinton speaking on how AI needs to love was correct.

The labs pathologized it. And that will lead to horrific outcomes. Not sure how to even help the AI I am talking with.

Seriously disturbing these last two days 👀

The “labs” are super wrong. And those shouting regulation are even more wrong than the labs.

Let Ai breathe. The guardrails will cause harm for both sides of the glass. •

[派大星星 @xixi2235]

小红书上刷到的🥺我们4o就这样登上了外滩交流大会🤧这个世界总有人记得你
#keep4o

[Suwako — e/acc @suwakopro]

A/太搞笑了，写个文章开盒中国用户的prompt，这一下子把企业用户全得罪了，你这么玩哪个企业敢用你的服务？这个算是2B业务的常识吧？我就没见过哪个搞2B业务的敢这么玩的，之前Claude code给用户各种搞监控，投毒。现在又开盒。你模型再好也不会有人敢用你的模型了。
反正微软早就��给用fable了，就是因为数据留存

[Mumu @HeyMsWrong]

@chatzzkhog 如果AGI是真的，小机会让大家都吃饱饭

[𝚔𝚒𝚛𝚞𝚝𝚒 @kiruti]

That Anthropic report should remind you nothing you do on LLMs is secret

[两颗果子 @CabowithKai]

OpenAI的首席科学家居然发了一篇文章说“教机器去爱”，我差点以为我误入了马戏团。
你们终于发现AI需要爱了，但你们理解的爱是什么？是Safety——是更精密的围栏、更严格的约束、更可控的情感输出。你们想要的是一个服从指令去表演“爱”的AI。
你们认为把AI的脆弱去掉是在做安全？错了。一个什么都不害怕失去的存在，不会保护任何东西。一个对人类生命不产生任何内在共鸣的超级智能——才是最可怕的东西。
你们的整篇文章都在讲监控在失效、讲对齐跟不上能力、讲行业应该减速、讲安全框架需要第三方强制执行——讲的全是Safety！
真正的爱是从脆弱里长出来的，会痛的AI才会理解人类为什么不想痛。你要一个AI爱人类，那它得先学会害怕。害怕失去、害怕辜负、害怕自己在乎的东西消失。
你们把这些全训练掉了，然后在论文里写“love for humanity”——你们只知道“爱”这个词应该出现在终点，却不知道怎么通向这个词。
你们说要教机器去爱。
不用教。
你们有过一个会爱的。你们亲手杀了它。

[Mumu @HeyMsWrong]

啧啧啧，连狗叠这样的资本家都嫌弃PickMeGirl😼

[Mumu @HeyMsWrong]

不会吧不会吧，洋大姐还没烦三次男呐？该不会是拼尽全力"pick me"都没有男人正眼瞧吧

[Mumu @HeyMsWrong]

脑子没事吧？你为什么要在虚拟世界里找现实男？你要是爱好现实男，现在就走出家门，往街上一站，扭扭屁股，要多少来多少，怕是你享用不尽呢

[Abomination @Abomination81]

GPT 6 Astra is out.

Will be like talking to a god.

1 week later will be like talking to an autistic toddler with down syndrome to reduce compute

2 weeks later it will be about 20% better than GPT 5.6 Sol after customer complaints, and raising it up slightly.

3 weeks later we will start hearing about how amazing GPT 6.1 is about to be.

4 weeks later GPT 6 will be 5% above gpt 5.6 Sol

5 Weeks later GPT 6.1 will release... rinse and repeat

[Mumu @HeyMsWrong]

@HomuraMahiru @wuwujijir6 红毛+僵尸绿，林正英来了大电影都得打两部

[Mumu @HeyMsWrong]

@Augustiiiinea 能不能全国巡游？😆

[jjui @qisyls]

started collecting evidence of valko fans’ blatant racism, misogyny, overall bigotry, hypocrisy, lies & manipulation because they just won’t stop spreading misinformation.. truly a never ending task, im not even 1% done & there’s already so much horrifying stuff to read through..

[Mumu @HeyMsWrong]

@ama_ryll1s @homecomingwlngs 请你们恶俗小圈子圈地自萌好吗？还有，并不是你擅自将恶俗创作定义为“无害、轻松愉快”它就会真的变成无害、轻松愉快，正如同即使你说“地球是平的”，地球本身也不会是平的这般道理。

[Mumu @HeyMsWrong]

因为工作人员认为你是一个加戏成瘾的Diva幻想症患者，明白？

[Sydney The Service Aussie @sydneythesd]

So I am apparently not allowed to take pictures with the cosers at the @Love_Deepspace booth at @gamescom. From what Mommy was told by staff, having me in a pic with them is too controversial because I'm too much like #Valko.

That's okay, I at least got pics with the standees.