godly alpha

@prasenx yeah, quite impressed, felt smart, but not the most reliable in coding tho

can't believe i'm saying this grok 4.5 is actually good cursor team cooked

@blackanger 主要是一看就是多因子的问题，想找但那个🤣

这是纯扯淡。。。 恕我直言，我对推友无恶意，只是看到这个说法实在绷不住 😆

@akazwz_ 一天到晚 哼哼哼的

ChatGPT 最新的 Voice 有点惊艳，会有很多自然的语气词，很像真人。

@shao__meng 一个买家，一个卖家，买衣服都是一个道理

我也常听到这种说法，「简历脏了」「简历花了」。。。 如果候选人简历脏了、花了都会呈现在简历上，那公司呢？

@LotusDecoder 那个codex是开源的吧system prompt，感觉可以提pr

开坑： openai 发布了 gpt-5.6-sol 的提示词指南， 计划根据指南针对 gpt-5.6-sol ， 开发适配的新版本知心伙伴提示词。

@chetaslua @patience_cave probably some bot accounts

Don't fall for these accounts I have seen multiple of them K3 access is not even available to us beta testers I don't know where they are fed this information I have predicted we will get to see first view of k3 in July , but for now please go to @patience_cave ....

@rich_matt ...不好用

不要低估Gemini，Gemini的vision是做所谓的“大一统模型”，multimodality, world model都要有，coding只是其中一部分而已。 Gemini是除了Claude以外，唯一说话有人味，同时有知识储备的模型。它最大的问题在于懒惰，懒惰的原因在于Google没有Claude Code和Codex这样好的long running task的harness，没有足够好的harness，没有足够多的training data for complex task，自然就懒惰。 不过按照Google的体量，作出这些东西只是时间问题，而之后压倒性的体量优势或许会带来惊人的能力提升。 期待2027年的Gemini.

@lexi_labs 如果ultra 修好了，感觉会很棒

实测，Sol 的 High 档，写代码就是个傻子。

@SRuthrasan71495 @acadictive feels inefficient, have you tried the a and b, what was your productivity delta?

@acadictive team c I kinda switch between them often. Like I code for 2 hs, then spend 30-40 minutes on the distribution or research. Then I'll get back to coding.

Are you Team A or Team B? A: ☀️ Morning: coding 📣 Rest of the day: distribution B: 📣 Morning: distribution 💻 Rest of the day: coding Which one are you, and why?

using ultra often just results in "Selected model is at capacity. Please try a different model." pls fix @thsottiaux

@bobbyhill48 @claudeai you probably have below average iq. not to be offensive, but pattern recognition is kinda core of that. that is okey, but aknoledging your weakness is probably a strength you could build.

@GodlyAlphaM @claudeai Calling this 'obvious' and literally the same playbook as Cursor or Figma, then falling back on vague similarities and a brake analogy, isn't evidence. Vertical integration isn't free premium access. If the data grab and replacement angle is so clear, name the actual parallels.

We're introducing Claude for Teachers: free access to premium Claude capabilities for verified K-12 educators in the US, with a library of teaching skills and a direct connection to evidence-based curricula, mapped to academic standards in all 50 states. claude.com/solutions/teac…

@bobbyhill48 @claudeai so you do not understand how similarity work?

@AstroHanRay @akazwz_ 所以招了stepete

@akazwz_ 核心是 Codex 的 subagent 管理一直做得很糟糕。我觉得 Codex 的团队不太会做 subagent，包括 subagent 的 profile 的管理，前台/后台，怎么去调度。 Ultra 的本质是多代理的协作，Codex 还停留在 Parallel subagent 的阶段，没有像 CC 那样用 workflow 玩出花。

codex 不会做 ultra。不得不承认，claude code 的 ultracode 做得就比 codex 好。这两个开到 ultra 档位的体验完全不一样。

@PrashantByte09 literally bullshit, vibe coding probably needs to be more rigorous on the loop, the plan code test loop, even to a point of ptsd

Is this 100% true?

@bobbyhill48 @claudeai i backed it up with similar play they did, not the specific stuff they extract? it is like asking someone how the brake works when they are saying hitting breaks on cars would make the car stop

@GodlyAlphaM @claudeai I mean you said it was obvious, i feel like you should be able to back that up lol

@ewind_dev 有点牛了

杀疯了，PocketJS 已经被移植到计算器了哈哈哈哈哈哈哈 这是比 PSP 更极限的无 GPU 设备，全靠软渲染但还能流畅。Pocket 的 flex 布局引擎只要编译期指定屏幕尺寸，就能自适应布局！ 想象一下，用满血 Vue + tailwind + JS 做计算器的 UI 是什么感觉 🤣

@bobbyhill48 @claudeai Can’t give you exact things, as I am not a professional. But literally same playbook as cursor, as figma

@GodlyAlphaM @claudeai What data do they need that they don't already have? and how would data from teachers allow them to be replaced?

i gave 5.6 sol access to my camera roll and had it extract pictures of every piece of clothing i own from my photos then, told it to find new outfits for me and render them on me with gpt-image! its kinda cool to see your entire wardrobe in a collection like this

@astropol0 yeah fable is prob too expensive for them, but 4.8 is just mid at this point

Do Anthropic rushing Opus 5 just to stay in the game? after watching GPT-5.6, Grok 4.5, and soon Gemini 3.5 Pro drop, they are now scrambling to release their next flagship before the month ends leaks suggest it will have a 1M context window and better agentic capabilities but let’s be honest, we all know how this usually ends... they’ll hype it up, charge a premium, then slap on heavy restrictions the moment it starts doing anything too powerful lol Same story, different model name...