何年派

🌟清华大学深研院在读 ing。
📚探索学习的本质，AI 的狂热爱好者。
🌐科技、经济、个人思考。
🚀Input 要尽快变成 Output。

[何年派 @Luke_Mathblue]

@vista8 20 刀的用了一年半，没啥事

[何年派 @Luke_Mathblue]

日落山谷

[何年派 @Luke_Mathblue]

grok bot 细节做的很好

[何年派 @Luke_Mathblue]

真不错，我决定做一门考研数学的版本。

[铜匠・十点睡觉 @CopperForgeAI]

我决定将我做了几个月的3d互动教育课程开源，每个月就那么零散的几个订阅单子，太浪费了。
由于我之前做的AI Saas是北美版，所以在陆续改成国内的版本，目前已经可以开源55门。

🔗 在线用：https://t.co/O4Pm54GHsc
🔗 GitHub：https://t.co/jwidnvL5ny

覆盖物理、化学、生物、地理四科，按中国高中课标和人教版教材做：合成氨、接触法制硫酸、氯碱工业、高炉炼铁、电流的磁场、光电效应、有丝分裂、神经冲动、三圈环流、厄尔尼诺……

每门课是什么样：
· 一个 3D 世界，能拖、能转、能剖开看塔里面
· 右边一台中控台，滑杆和开关每一步都能动，读数和曲线实时变
· 一条步骤线，老师一路按「下一步」就能讲完一节课，学生没动手的步骤会自动演示
· 最后一步是考核，有目标值

几个我比较在意的点：
1. 每门课就是一个 HTML 文件，约 1 MB。拷进 U 盘，双击就能打开，断网、投影、电子白板都能用。
2. 数字可追溯。每门课先写事实表：教材里的数值带页码，为了教学设的数值单独标「课件设定」并写明理由。课件上出现的每个数都要能在事实表里找到。
3. 可以下载、可以嵌入学校网站，课程页一键复制嵌入代码。

怎么做出来的：
为了节约token，我大量使用了DeepSeek，所以效果也没有我做的Sass站那么好，但是整体还可以的。

· DeepSeek 批量查教材、写事实表、搭课件
· Claude 设计统一的课件外壳和视觉规范，负责审核
· 图像模型出参考图和封面
· 每门课在三种屏幕尺寸下逐步跑自动检查（数字、遮挡、对比度、标注有没有压在一起），再由人一步一步看截图，过了才上线

AI 能把速度拉上来，但「讲得对、看得懂」还是要靠一遍遍地审。即便这样，55 门里肯定还有错。如果你是老师或学生，发现科学错误或者画面看不懂的地方，请直接在课程页点「报告错误」，或者在 GitHub 提 Issue，我会一条条修，如果有其他的课程需求，我也会慢慢开发，反正也是Agent在做，我做好了sop，工作量也不大。

接下来：
· 四科继续往上做，目标全部做完，之后做初中和小学
· 出英文和其他语言版本
· 把完整的制作方法写成长文

如果觉得有用，帮忙转给身边的老师，或者给仓库点个 ⭐。也欢迎关注我，我会在这个线程持续分享课程。

[何年派 @Luke_Mathblue]

我的 bot 大军又回来了。

[何年派 @Luke_Mathblue]

@Khazix0918 哈哈哈

[何年派 @Luke_Mathblue]

@Argona0x 闷声发财为什么告诉大家？

[何年派 @Luke_Mathblue]

@oran_ge 保护伞公司逐渐成型。

[何年派 @Luke_Mathblue]

很多人都没搞明白，学习之所以是学习，就是神经元连接的改变，不可能很舒服的。与其找一堆工具和方法，不如踏踏实实地遵循底层逻辑去做吧。

[何年派 @Luke_Mathblue]

牛🐮

[James Camp 🛠,🛠 @JamesonCamp]

Fastest way to grow a small account on X right now is quote tweets

you dont need followers. youre borrowing the audience of the post you quote

my emirates QT was one line and did over 1M impressions

the hard part is finding the right post while its still climbing.

thats the whole setup below

[何年派 @Luke_Mathblue]

@xingbugengming 酷

[何年派 @Luke_Mathblue]

@bozhou_ai 写作依然垃圾，不如在聊天中。

[何年派 @Luke_Mathblue]

还是一坨，没有聊天好用

[Tibo @thsottiaux]

Day 22/ Terra is back!

Our first model that is significantly better at writing, personality and nuanced understanding of requests.

Packs a punch across the board, very good at computer use, and gives 2X more usage than 6.1 Sol across API and subscriptions.

[何年派 @Luke_Mathblue]

突发 ，A 社即将上市！

[Anthropic @AnthropicAI]

Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO.

Read more: https://t.co/ofhv0DK3Zp

Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.

[Anthropic @AnthropicAI]

Anthropic has filed an S-1 registration statement with the SEC related to its proposed IPO.

Read more: https://t.co/ofhv0DK3Zp

Legal disclaimer: A registration statement relating to these securities has been filed with the SEC but has not yet become effective. These securities may not be sold, nor may offers to buy be accepted, prior to the time the registration statement becomes effective. This post does not constitute an offer to sell or the solicitation of an offer to buy these securities, nor shall there be any sale of these securities in any state or jurisdiction in which such offer, solicitation or sale would be unlawful prior to registration or qualification under the securities laws of any such state or jurisdiction.

[何年派 @Luke_Mathblue]

@NoraX2026 公认的，经过科学试验的，其实没有那么多。乱七八糟的。

[何年派 @Luke_Mathblue]

@NoraX2026 主动回忆、间隔重复。

[何年派 @Luke_Mathblue]

@yupi996 因为Fable5.5 要出了

[何年派 @Luke_Mathblue]

@kk_shinkai 现在感觉御三家还是各有千秋。

[何年派 @Luke_Mathblue]

@xingbugengming 主要是你做的质量很顶啊