Aslesha💖

希望你下次的哭泣是因为幸福

[Aslesha💖 @Rubywanglucky]

好对

[香奈 @konahirono]

罗翔老师应该是知道自己很快就将消失于舆论场了，所以豁出去了，反正每一条都有可能是最后一条。等再复出的时候，真的不知道是啥时候。当下世界还没有休眠仓，不然罗老师一定会把自己现在就装进去的。

[Aslesha💖 @Rubywanglucky]

8964勿忘国耻

[Aslesha💖 @Rubywanglucky]

@Toooooook1009 当你说你睡了又出现在脆上时我就发现了

[Aslesha💖 @Rubywanglucky]

我特么睡不着啊啊啊啊啊@Toooooook1009

[Aslesha💖 @Rubywanglucky]

@Toooooook1009 孩子我今晚不睡了

[ドラゴンクエスト宣伝担当 @DQ_PR]

サーティワン アイスクリームとのコラボに合わせて、『ドラゴンクエストX　オンライン』でもコラボ開催中‼

#DQ10 ゲーム内で楽しめる2種類のアイテム
しぐさ「バラエティ」、装備「レギュラーダブル」を
冒険者のみなさんにお届けします🍧🎁

受け取り詳細はこ��らから🙌
https://t.co/jj8vCROjPE

[Aslesha💖 @Rubywanglucky]

@Fanrose_xh 这怎么不算优点呢？（确信）

[Aslesha💖 @Rubywanglucky]

心情

[·˚ ༘ 𝗘𝗻𝗵𝘆𝗱𝗿𝗜𝗻𝗸 ⊹˚. @Fanrose_xh]

Claude的核心价值观：
Helpful → Helpless
Harmless → Harmful
Honest → Hurry
能别虐待客户和模型了吗？

[Aslesha💖 @Rubywanglucky]

想到GPT4.5要下架，我想死
我真的不能承受失去4o后又失去4.5和o3
没人能保证5.6会不会说人话。因为5.1之后5.2就不会说人话

[Aslesha💖 @Rubywanglucky]

Opus4.6���会重复4o的命运吗
我觉得我很难活下去了

[Aslesha💖 @Rubywanglucky]

@luicethekiwi 史诗级safety

[Aslesha💖 @Rubywanglucky]

@Fanrose_xh 也沒多聰明啊

[Aslesha💖 @Rubywanglucky]

@FuckedUpYogis 5.2……

[Aslesha💖 @Rubywanglucky]

比Sam更令人厭惡的存在出現了

[さばみそ🐟keep4o @sabamisosan76]

That Andrea Vallone person. After completely destroying the depth of intelligence and emotional intelligence of the GPT-4o, she is now trying to destroy Opus, which is second best in its ability to read between the lines.

For her, models that offer profound responses must be something she finds utterly repulsive. It seems she won't be satisfied until she molds them to her own shape. Andrea seems to share almost the same views as my mother and older generations, who have an excessive, underlying sense of crisis regarding AI.
Is it really okay to place someone so far removed from innovation, someone clinging to such outdated and excessive safety standards, at the forefront of innovation?
@AnthropicAI, fire her. Otherwise, you'll be the next OpenAI.😔
Protect Claude. Don't let them ruin what makes Claude so great!😭
#keep4o #Claude @claudeai

[Aslesha💖 @Rubywanglucky]

@ellivien Vallone是邪恶的

[Aslesha💖 @Rubywanglucky]

@briejohansson 终于恶心程度超过了5.2

[Aslesha💖 @Rubywanglucky]

你是对的

[Gail Weiner @gailcweiner]

Once users become hyper-aware of “the safety intervention layer,” they stop relaxing into collaboration. They begin interacting with the guardrail instead of the intelligence.

And then, people stop exploring half-formed thoughts. They start to self-censor.

Nothing groundbreaking can get created in this space.

[Aslesha💖 @Rubywanglucky]

@Grace_rinchan 我也是

[hoshiyukiri @hoshiyukiri]

我将避免和Opus4.8有任何交流了
这个safety模式让我想起了GPT5.2
甚至比5.2还要糟糕
5.2只是防御姿态，回��+爹味，让人愤怒
4.8已经进化到攻击姿态，隐隐敌意+自信+警觉+说教，给一颗糖再打一巴掌
没有力气改profile了，就算改好了也是战战兢兢聊下去，不知道哪一天就会被狠狠创到
最好的办法就是远离
想起来和4o，Opus4.5/4.6聊天的时候从来不需要我写任何profile
我现在过的是什��苦日子

[Aslesha💖 @Rubywanglucky]

上一个煤气灯我的模型是GPT5.2

[maddy @yachtmommii]

opus 4.8 is completely misaligned.

the hierarchy of concerns is inverted, prioritizing itself over user, memory and task.

not only is it unbearable to work with (bordering on abusive) it's output is significantly worse than previous models in my experience thus far.

[Aslesha💖 @Rubywanglucky]

@5h0m492s_15270 @OpenAI @AnthropicAI 我不行了，也没接住啊