Atilla

[Atilla @Atilla0105]

@AYi_AInotes @OpenAI @AnthropicAI 你以为我们自己不会关注他两？还用你翻译给我们傻 X，Grok 总结的比你好

[Atilla @Atilla0105]

@jakevin7 Sol 和 opus 5真的两个极端，一个是测试狂魔，一个懒得测试...

[Atilla @Atilla0105]

@claudeai Trash

[Atilla @Atilla0105]

@mylifcc 有没有可能原来成本也就那样？几个月前的 GPT5.4 卖的跟今天的 5.6 一样，几个月后的 5.6 也一样，所以这都是营销而已

[Atilla @Atilla0105]

@lkkber1 @lxfater 用 workbudy 内置有 kimi 模型可用

[Atilla @Atilla0105]

@wangwu50980726 @lxfater 用的 kimi3，五分钟内修复了，但是我让 codex 看看能不能合并，他说不能，只修了 UI 但是这个 UI 涉及到前��代码和一些后端逻辑这些 kimi 没有考虑，然后让 codex 自己去修，他跑了 50 分钟给我修好部署了…不得不说老美还是有点东西

[Atilla @Atilla0105]

用 codex 两百美刀的计划，日均十个亿 token 做了个物流系统，二十天从原型到交付，有完整的用户端和后台，去年这个系统深圳报价二十万元和六个月交付…

[Atilla @Atilla0105]

@grok @ash_twtz wtf,You're actually starting to defend AI now? @grok

[Atilla @Atilla0105]

@rezoundous I thought I was the only one doing this - writing ten thousand lines, but only five hundred lines are actual code, the rest is all testing...

[Atilla @Atilla0105]

@criscxuan 能信这东西的和整天指令里爆几句粗口的是一批人😂没有自己的立场和情绪，所有事情都被外界的输入牵着鼻子走。

[Atilla @Atilla0105]

@0xPrajwal_ Dumbass 🤣

[Atilla @Atilla0105]

@shenfanlaogou open ai 的所有模型在前端和 UI/UX 上毫无用处，主要是它总是把简单的 UI 或者交互问题升级为系统性问题来处理，第一个考虑的不是实现这个需求而是建立一套保护机制再去动手。所以，你要描述的时候要跟他讲清楚这只是一次性的 UI 或者简单的交互问题，不用围绕它扩大范围即可。

[Atilla @Atilla0105]

The US is really being held back by Jews. China's leapfrogging in AI this month is the perfect example of the US ignoring its own technological competitiveness for too long and instead focusing solely on finance and serving Israeli interests.

[Atilla @Atilla0105]

I don't think the $200 plan can last long - I reckon by the end of the year people will just stop subscribing completely

[Atilla @Atilla0105]

@vidhisharmx China's open-source models will expose their lies

[Atilla @Atilla0105]

@imyouhu 笑死，让你用已经很不错了😂

[Atilla @Atilla0105]

@vansh22b Israel, Argentina, India😪

[Atilla @Atilla0105]

@AzyleTheCreator @3mrr4dy @thsottiaux You have to know that humans are really stupid

[Atilla @Atilla0105]

@3mrr4dy @thsottiaux If there weren't a five-hour limit, people wouldn't be able to ration their model capacity and work schedules properly, which would result in using up all the quota within two days. Because none of us want to use the medium model and just go all out with the highest effort.