steven

@steven12c

I am an AI enthusiast, like new technology, like mac, technology sharing, love life

Tham gia vào Tháng 7 2014

189Bài đăng 476Người theo dõi 423Đang theo dõi

steven đã đăng lại

池建强

@sagacity

14 thg 12

周末看了个 OpenAI 论坛的内部分享，叫 Vibe Engineering。内容不少，我整理了分享的全文放到墨问里，差不多 9100 字，还有实操项目的 github 仓库地址，可以先收藏再看： Web 版本：note.mowen.cn/detail/6gMtxCP… 什么是 Vibe Engineering？AI…

steven đã đăng lại

meng shao

@shao__meng

13 thg 12

[开源推荐] Claude-Mem: 为 Claude Code 设计的持久化记忆压缩系统来自 @Claude_Memory，通过智能压缩和本地存储，能够记住之前的编程操作、决策和上下文，赋予 AI 跨会话的长期记忆能力并降低 Token 消耗，基于 @trychroma 向量数据库。核心价值：解决“失忆”问题通常情况下，当你结束与 AI…

shao__meng's tweet image. [开源推荐] Claude-Mem: 为 Claude Code 设计的持久化记忆压缩系统

来自 @Claude_Memory，通过智能压缩和本地存储，能够记住之前的编程操作、决策和上下文，赋予 AI 跨会话的长期记忆能力并降低 Token 消耗，基于 @trychroma 向量数据库。

核心价值：解决“失忆”问题
通常情况下，当你结束与 AI…

elvis

@omarsar0

11 thg 12

Claude Code plugin to persist memory across sessions. 3.7K⭐️ It's built with hooks, uses SQLite for session storage, and supports both semantic and keyword search.

omarsar0's tweet image. Claude Code plugin to persist memory across sessions.

3.7K⭐️

It's built with hooks, uses SQLite for session storage, and supports both semantic and keyword search.

steven đã đăng lại

huangserva

@servasyy_ai

12 thg 12

两个月推特，粉丝 0→13.9k，收入 1w+ 很多人问我：“这2月你到底成长了啥？” 数据只是表面的，没去拿x工资，小钱也不可惜，我早就不在意了。我来推的本意就是学习，而这俩月学到的东西，值回票价太多： 1. 从只会 Cursor 敲点玩具代码 → 到日常狂用 Claude Code、Codex、Droid、Devin 等 2.…

huangserva

@servasyy_ai

11 thg 12

过去的就让它过去，成长的是我，不是这个账号本身！

凯林

@kaylyn_0x

11 thg 12

今天发现自己莫名掉了 100 多粉去查了一下才知是 X 又主动封了一批账号，其中大多都是我 4 月多蓝 V 互关认识的，他们透露被冻结的主要原因是违反“真实性” 这可太熟悉了，之前我小号就是这么被封的。事后汲取教训，本号运营至今都安然无恙，分享些经验给大家： 1. 最好使用固定 VPN 和…

kaylyn_0x's tweet image. 今天发现自己莫名掉了 100 多粉

去查了一下才知是 X 又主动封了一批账号，其中大多都是我 4 月多蓝 V 互关认识的，他们透露被冻结的主要原因是违反“真实性”

这可太熟悉了，之前我小号就是这么被封的。事后汲取教训，本号运营至今都安然无恙，分享些经验给大家：

1. 最好使用固定 VPN 和…

steven đã đăng lại

🚀 Qwen3-Omni-Flash just got a massive upgrade (2025-12-01 version) ! What's improved: 🎙️ Enhanced multi-turn video/audio understanding - conversations flow naturally ✨ Customize your AI's personality through system prompts (think roleplay scenarios!) 🗣️ Smarter language…

Alibaba_Qwen's tweet image. 🚀 Qwen3-Omni-Flash just got a massive upgrade (2025-12-01 version) !

What's improved:

🎙️ Enhanced multi-turn video/audio understanding - conversations flow naturally
✨ Customize your AI's personality through system prompts (think roleplay scenarios!)
🗣️ Smarter language…

steven đã đăng lại

FastAPI

@FastAPI

8 thg 12

The full-stack-fastapi-template now uses @shadcn / ui (with @tailwindcss) for the frontend 🚀🎉 There's still dark mode too, and all the nice things 🦇 Thanks @alejsdev ! 🙌 github.com/fastapi/full-s…

FastAPI's tweet image. The full-stack-fastapi-template now uses @shadcn / ui (with @tailwindcss) for the frontend 🚀🎉

There's still dark mode too, and all the nice things 🦇

Thanks @alejsdev ! 🙌

github.com/fastapi/full-s…

steven

@steven12c

8 thg 12

Nano Banana Pro 提示词收藏

Jackywine

@Jackywine

4 thg 12

这个网站牛啊 Nano Banana Pro 500 多个玩法提示词全收集： youmind.com/zh-CN/nano-ban…

steven

@steven12c

8 thg 12

glm的新模型来了

Z.ai

@Zai_org

8 thg 12

GLM-4.6V Series is here🚀 - GLM-4.6V (106B): flagship vision-language model with 128K context - GLM-4.6V-Flash (9B): ultra-fast, lightweight version for local and low-latency workloads First-ever native Function Calling in the GLM vision model family Weights:…

Zai_org's tweet image. GLM-4.6V Series is here🚀

- GLM-4.6V (106B): flagship vision-language model with 128K context
- GLM-4.6V-Flash (9B): ultra-fast, lightweight version for local and low-latency workloads

First-ever native Function Calling in the GLM vision model family

Weights:…

steven

@steven12c

8 thg 12

今天直接升级到16版本了

weijunext - nexty.dev

@weijunext

6 thg 12

提醒大家，一定要重视前两天 React 和 Next.js 爆出的漏洞，今天群里越来越多人中招了。不要心存侥幸，只要你的项目是今年启动的，基本上都处在风险中，因为 Next.js 15 和 16 所有版本都有问题。修复方法其实很简单，只要把你使用的 Next.js 和 React 版本升级到任意修复版本就可以。…

weijunext's tweet image. 提醒大家，一定要重视前两天 React 和 Next.js 爆出的漏洞，今天群里越来越多人中招了。

不要心存侥幸，只要你的项目是今年启动的，基本上都处在风险中，因为 Next.js 15 和 16 所有版本都有问题。

修复方法其实很简单，只要把你使用的 Next.js 和 React 版本升级到任意修复版本就可以。…

steven

@steven12c

8 thg 12

泊舟

@bozhou_ai

7 thg 12

发现了一个宝藏网站，不得不说信息源这块比我自己做的好得多，包含文章，视频，播客，推文，最重要的是还免费，简直赛博菩萨

steven

@steven12c

8 thg 12

学习黄老师的新作

泊舟

@bozhou_ai

7 thg 12

发现了一个宝藏网站，不得不说信息源这块比我自己做的好得多，包含文章，视频，播客，推文，最重要的是还免费，简直赛博菩萨

steven

@steven12c

6 thg 12

喔，可惜体验不了

G3mini

@GeminiApp

4 thg 12

Gemini 3 Deep Think is here. Deep Think is our most advanced reasoning mode that explores multiple hypotheses simultaneously to give you an even more sophisticated output.

steven

@steven12c

4 thg 12

还不快去试试可灵 2.6 模型

G3mini

@GeminiApp

4 thg 12

Gemini 3 Deep Think is here. Deep Think is our most advanced reasoning mode that explores multiple hypotheses simultaneously to give you an even more sophisticated output.

steven

@steven12c

3 thg 12

电脑配置够用的真的可以试试

G3mini

@GeminiApp

4 thg 12

Gemini 3 Deep Think is here. Deep Think is our most advanced reasoning mode that explores multiple hypotheses simultaneously to give you an even more sophisticated output.

steven

@steven12c

2 thg 12

黄老板的新作，很不错

G3mini

@GeminiApp

4 thg 12

Gemini 3 Deep Think is here. Deep Think is our most advanced reasoning mode that explores multiple hypotheses simultaneously to give you an even more sophisticated output.

steven

@steven12c

1 thg 12

很漂亮的风格

G3mini

@GeminiApp

4 thg 12

Gemini 3 Deep Think is here. Deep Think is our most advanced reasoning mode that explores multiple hypotheses simultaneously to give you an even more sophisticated output.

steven

@steven12c

30 thg 11

电脑内存够用的，真的可以试一下

Gorden Sun

@Gorden_Sun

30 thg 11

本地部署也特别简单，直接用ComfyUI就行，我的配置是M3 Max 36G，内存占用30G，大概25秒出一张1024分辨率的图。官方教程：docs.comfy.org/tutorials/imag… 简单讲，就3步： 1）下载ComfyUI 2）分别下载模型： qwen_3_4b.safetensors，放到ComfyUI/models/text_encoders文件夹里…