LogoThread Easy
  • 探索
  • 撰写 Thread
LogoThread Easy

您的一体化 Twitter 线程助手

© 2025 Thread Easy All Rights Reserved.

探索

最新在前,按卡片方式浏览线程

开启时会模糊预览图,关闭后正常显示

RT @indie_maker_fox: 之前分享过Typefully,现在它的API v2要出了

感觉这就很妙,利用一个跨媒体内容发布平台的API去管理你要发布的内容,这对于内容自动跨平台发布来说,好像特别有用

https://t.co/b8PhBOcIc3

Typef…

RT @indie_maker_fox: 之前分享过Typefully,现在它的API v2要出了 感觉这就很妙,利用一个跨媒体内容发布平台的API去管理你要发布的内容,这对于内容自动跨平台发布来说,好像特别有用 https://t.co/b8PhBOcIc3 Typef…

🔥 The best AI SaaS boilerplate - https://t.co/VyNtTs0jSX 🚀 The best directory boilerplate with AI - https://t.co/wEvJ1Dd8aR 🎉 https://t.co/bh1RxeERuY & https://t.co/zubXJCoY92 & https://t.co/tfQf8T7gGF

avatar for Fox@MkSaaS.com
Fox@MkSaaS.com
Wed Dec 03 01:01:04
It's remarkable that Opus, Gemini and DeepSeek all conclude that Proof C (DeepSeek) is either human-written or indeed from DeepSeek. GPT 5.1 labels it "human" and then assigns both its own output (rating it 2/10!) and actual human proof to itself.
of note, Opus on DS-Math V2:

It's remarkable that Opus, Gemini and DeepSeek all conclude that Proof C (DeepSeek) is either human-written or indeed from DeepSeek. GPT 5.1 labels it "human" and then assigns both its own output (rating it 2/10!) and actual human proof to itself. of note, Opus on DS-Math V2:

We're in a race. It's not USA vs China but humans and AGIs vs ape power centralization. @deepseek_ai stan #1, 2023–Deep Time «C’est la guerre.» ®1

avatar for Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Teortaxes▶️ (DeepSeek 推特🐋铁粉 2023 – ∞)
Wed Dec 03 00:59:30
🎉  分享一个宝藏浏览器插件,drizzle studio

原本在Cloudflare的dashboard的D1数据库中看数据的体验是挺差的,现在你只要在后台管理界面,点击这个插件,它会自动识别你的D1数据库列表,并使用Drizzle Studio去打开,瞬间查看数据库的体验舒服了!

🎉 分享一个宝藏浏览器插件,drizzle studio 原本在Cloudflare的dashboard的D1数据库中看数据的体验是挺差的,现在你只要在后台管理界面,点击这个插件,它会自动识别你的D1数据库列表,并使用Drizzle Studio去打开,瞬间查看数据库的体验舒服了!

🔥 The best AI SaaS boilerplate - https://t.co/VyNtTs0jSX 🚀 The best directory boilerplate with AI - https://t.co/wEvJ1Dd8aR 🎉 https://t.co/bh1RxeERuY & https://t.co/zubXJCoY92 & https://t.co/tfQf8T7gGF

avatar for Fox@MkSaaS.com
Fox@MkSaaS.com
Wed Dec 03 00:57:05
DeepSeek V3.2 Speciale one shot the shortest solution to my 3rd (currently, hardest) λ-calculus problem.

Current solutions, from worst to best.

Opus 4.5:

rotL = λN.λT.λt.(T λk.(λn.λs.λz.(n λg.λh.(h (g s)) λu.z λu.u) N λu.λr.λk.λe.(u (r e) k) λr.λk.(r k) t k))

Gemini 3 Pro / GPT 5 Pro:

rotL = λn.λt.λf.(t λx.(n λg.λh.(h (g λp.λg.λy.(p (g y)))) λu.λg.(g x) λu.u f))

DeepSeek V3.2:

rotL = λN.λT.λc.(T λa.(N λf.λg.λx.(f (g x)) λh.(h a) λz.z c))

All are correct.

I find this curious. A few months ago, no model solved this. Now all recent ones (except GPT 5.1) seem to. I'll make harder problems!

Worth noting that DeepSeek failed on the stack underflow bug, which only Opus 4.5 and a non public checkpoint of Gemini 3 Pro solved.

DeepSeek V3.2 Speciale one shot the shortest solution to my 3rd (currently, hardest) λ-calculus problem. Current solutions, from worst to best. Opus 4.5: rotL = λN.λT.λt.(T λk.(λn.λs.λz.(n λg.λh.(h (g s)) λu.z λu.u) N λu.λr.λk.λe.(u (r e) k) λr.λk.(r k) t k)) Gemini 3 Pro / GPT 5 Pro: rotL = λn.λt.λf.(t λx.(n λg.λh.(h (g λp.λg.λy.(p (g y)))) λu.λg.(g x) λu.u f)) DeepSeek V3.2: rotL = λN.λT.λc.(T λa.(N λf.λg.λx.(f (g x)) λh.(h a) λz.z c)) All are correct. I find this curious. A few months ago, no model solved this. Now all recent ones (except GPT 5.1) seem to. I'll make harder problems! Worth noting that DeepSeek failed on the stack underflow bug, which only Opus 4.5 and a non public checkpoint of Gemini 3 Pro solved.

Kind / Bend / HVM / INets / λCalculus

avatar for Taelin
Taelin
Wed Dec 03 00:47:25
Gemini 3 Pro、Codex Max 5.1 和 Claude Opus 4.5 一起挑战「从零开始构建网页版多人在线 3D 反恐精英射击游戏」

🎯 核心挑战
任务被分为两个阶段,共约 7 个提示词:
· 前端开发:设计游戏场景、角色、物理效果、第一人称视角射击和音效。
· 后端开发:实现多人联机功能、房间选择系统以及数据持久化。

🏆 选手表现速览
· Claude Opus 4.5:✨ 最佳设计师前端王者
   它生成的地图、角色和枪械模型在视觉上最精致,审美在线,“像个真正的游戏”。但在处理复杂的后端逻辑重构时容易卡壳,需要人工介入修 bug。
· Gemini 3 Pro:🛠️ 最佳工程师后端专家
   在处理多人联机、数据库和代码重构时表现最稳健。它擅长通过不断运行构建命令来自我修复错误,逻辑性极强,几乎是一次性跑通了复杂的后端需求。
· Codex Max 5.1:⚖️ 平衡型选手中规中矩
   在前端和后端都拿到了不少“第二名”。它表现稳定,不像另外两个模型那样有明显的偏科(要么视觉极好,要么逻辑极强)。

💡 关键洞察
1. “视觉”与“逻辑”的分野
· Claude 就像一位充满艺术感的前端开发者,它生成的方块人甚至有 “Minecraft” 级别的细节,枪械有后坐力动画,视觉体验极佳。
· Gemini 则像一位严谨的后端架构师。在面对“将单机游戏重构为支持多房间的多人游戏”这一复杂任务时,Gemini 展现了强大的逻辑推理能力,能从文档中学习并自我修正,而 Claude 则陷入了 React 生命周期(useEffect)的陷阱中。

2. AI 的“学习”方式不同
· Codex 倾向于“自省”,它会深入检查现有的代码库和函数定义。
· Claude 极其依赖文档,它会反复阅读提供的开发文档,但在代码库内部逻辑探索上稍弱。
· Gemini 则是实干派,它结合了阅读文档和“试错”,通过不断运行编译命令来发现报错并修复,这种工作流非常高效。

3. “Vibe Coding” 的未来
虽然三个模型在零人工手写代码的情况下都做出了游戏,但完全不看代码的 “Vibe Coding” 时代尚未完全到来。当遇到复杂的 React 状态管理问题时,仍然需要人类工程师介入“救火”。

📝 总结
· 需要快速出图、做原型、搞设计,选 Claude
· 需要构建坚固的系统、处理复杂逻辑,选 Gemini
· 想要一个稳健的中间选项,Codex 是不错的选择

阅读原文

Gemini 3 Pro、Codex Max 5.1 和 Claude Opus 4.5 一起挑战「从零开始构建网页版多人在线 3D 反恐精英射击游戏」 🎯 核心挑战 任务被分为两个阶段,共约 7 个提示词: · 前端开发:设计游戏场景、角色、物理效果、第一人称视角射击和音效。 · 后端开发:实现多人联机功能、房间选择系统以及数据持久化。 🏆 选手表现速览 · Claude Opus 4.5:✨ 最佳设计师前端王者 它生成的地图、角色和枪械模型在视觉上最精致,审美在线,“像个真正的游戏”。但在处理复杂的后端逻辑重构时容易卡壳,需要人工介入修 bug。 · Gemini 3 Pro:🛠️ 最佳工程师后端专家 在处理多人联机、数据库和代码重构时表现最稳健。它擅长通过不断运行构建命令来自我修复错误,逻辑性极强,几乎是一次性跑通了复杂的后端需求。 · Codex Max 5.1:⚖️ 平衡型选手中规中矩 在前端和后端都拿到了不少“第二名”。它表现稳定,不像另外两个模型那样有明显的偏科(要么视觉极好,要么逻辑极强)。 💡 关键洞察 1. “视觉”与“逻辑”的分野 · Claude 就像一位充满艺术感的前端开发者,它生成的方块人甚至有 “Minecraft” 级别的细节,枪械有后坐力动画,视觉体验极佳。 · Gemini 则像一位严谨的后端架构师。在面对“将单机游戏重构为支持多房间的多人游戏”这一复杂任务时,Gemini 展现了强大的逻辑推理能力,能从文档中学习并自我修正,而 Claude 则陷入了 React 生命周期(useEffect)的陷阱中。 2. AI 的“学习”方式不同 · Codex 倾向于“自省”,它会深入检查现有的代码库和函数定义。 · Claude 极其依赖文档,它会反复阅读提供的开发文档,但在代码库内部逻辑探索上稍弱。 · Gemini 则是实干派,它结合了阅读文档和“试错”,通过不断运行编译命令来发现报错并修复,这种工作流非常高效。 3. “Vibe Coding” 的未来 虽然三个模型在零人工手写代码的情况下都做出了游戏,但完全不看代码的 “Vibe Coding” 时代尚未完全到来。当遇到复杂的 React 状态管理问题时,仍然需要人类工程师介入“救火”。 📝 总结 · 需要快速出图、做原型、搞设计,选 Claude · 需要构建坚固的系统、处理复杂逻辑,选 Gemini · 想要一个稳健的中间选项,Codex 是不错的选择 阅读原文

邵猛,中年失业程序员 😂 专注 - Context Engineering, AI Agents. 分享 - AI papers, apps and OSS. ex Microsoft MVP 合作 - 私信/邮箱:shaomeng@outlook.com 📢 公众号/小红书: AI 启蒙小伙伴

avatar for meng shao
meng shao
Wed Dec 03 00:47:04
Parents encourage or let their kids opt into disability diagnosis because it seems like there’s little downside: more time on tests, better chances at college admissions, optional performance enhancing drugs, accommodations of all types.

I can’t stress this enough: the consequence is your child’s character. Many kids genuinely believe they’re sick or that there’s something wrong with them. You tell a girl she seems anxious, she’ll believe she’s anxious. You tell a boy he has a true disability in the form of ADHD and he starts thinking creativity or day dreaming is a deficiency. 

You can gaslight people into believing they’re sick, and we have entire systems and institutions encouraging this. We’ve convinced young people it’s fine to be weak and frail, when we should be doing the opposite: convincing them they’re resilient, independent, strong people who can handle any challenge.

If there’s one thing I believe as a parent, it’s that you can will your children into greatness. Society will encourage them to do the opposite, but you don’t have to comply.

Parents encourage or let their kids opt into disability diagnosis because it seems like there’s little downside: more time on tests, better chances at college admissions, optional performance enhancing drugs, accommodations of all types. I can’t stress this enough: the consequence is your child’s character. Many kids genuinely believe they’re sick or that there’s something wrong with them. You tell a girl she seems anxious, she’ll believe she’s anxious. You tell a boy he has a true disability in the form of ADHD and he starts thinking creativity or day dreaming is a deficiency. You can gaslight people into believing they’re sick, and we have entire systems and institutions encouraging this. We’ve convinced young people it’s fine to be weak and frail, when we should be doing the opposite: convincing them they’re resilient, independent, strong people who can handle any challenge. If there’s one thing I believe as a parent, it’s that you can will your children into greatness. Society will encourage them to do the opposite, but you don’t have to comply.

Co-Founder, American Dynamism. General Partner @a16z. Catholic. Mother. American. 🇺🇸 🚀💪

avatar for Katherine Boyle
Katherine Boyle
Wed Dec 03 00:46:46
  • Previous
  • 1
  • More pages
  • 1753
  • 1754
  • 1755
  • More pages
  • 5634
  • Next