🚨 someone just dropped a full 10-stage academic research pipeline for Claude Code.
It doesn’t write your paper for you, it hunts references, formats citations, verifies data, and even runs a "devil's advocate" agent to attack your own thesis.
Here's why it's a massive deal for academics:
→ Anti-AI Voice: Learns your specific writing style.
→Integrity Gates: Actively hunts down fabricated citations and statistical errors.
→Simulated Peer Review: Runs your draft through a 7-agent panel (including a Devil’s Advocate).
→Cheap: A full 15k-word paper costs ~$4–$6 in API credits.
Best part?
It's 100% free and open-source.
Install in 30s: `/plugin install academic-research-skills`
repo in 🧵↓
因为太多人写过 Andrej Karpathy 的 LLM Wiki,我就没写,其实在我心中比 Auto Research 更有创意,Auto Research 本身不新鲜,早就有相关理论,但 LLM WIKI 倒是让我眼前一亮。
我们每个人或多或少都在做信息收集的工作,比如 X 上看到好的文章点赞或者收藏,看到一篇好的技术文章添加到浏览器收藏夹,微信上有人分享了篇好文章点收藏,还有更多的是惊鸿一瞥再也找不到然后想起来根据关键词去 Google ……
其实绝大部分收藏后再也不会打开,一方面是因为收藏即看过的心理暗示,一方面是因为散落各地找起来太麻烦。
所以第一个问题其实是中心化的信息收集整理,把散落在各处的信息汇聚在一处。
已经有很多工具了,我自己也有写小工具/agent 帮助收集信息,因为我除了收集外还有一些二次加工的需要,比如翻译、总结。
但还存在问题就是信息是点状的,最多人工打个tag、加个分类。
但 Karpathy 的更进一步,让 LLM 帮你把信息整理成结构化的。这一步是我之前没考虑过的,也没见过有其他产品做的。
这里面的差别在于以前整理是要人做的,你自己建分类,自己打 tag,对于勤劳的爱整理的人当然没问题,但对于我这种懒人来说是不会做的,所以找信息是比较麻烦的。
但如果这种事情让 Agent 做,那就省事多了,毕竟它不知疲倦,而且极擅长处理内容。
只要稍加调教,它就能帮你把信息整理得井井有条,编程成你自己喜欢的格式,就像你的秘书一样,你只要去看看 WIKI 就可以方便的找到需要的信息,不需要以前那样去各个地方用关键字找。
这里面最核心是思路的转变,信息的收集和整理,不再是人主动的行为,而是 AI Agent 在帮你做这些事情,你所要做的就是每天去看属于自己的 WIKI。
Boris Cherny, the creator of Claude Code, shared his entire setup.
He runs 5-10 Claudes in parallel. Half his coding happens from his phone.
Here's his 3-part formula for better results:
Use the smartest model available
— Counterintuitive: it's actually cheaper
— Smarter model = fewer tokens = lower total cost
— "Once the plan is good, the code is good"
Invest in your Claude MD
— Plain text file. No special format.
— Whole team contributes multiple times a week
— Every mistake Claude makes gets added so it never happens again
Give Claude a way to verify its own output
— Let it run the code. Let it see the browser.
— "Imagine you're a painter wearing a blindfold"
— Same thing for an AI that can never check its work
His morning routine: wake up, kick off 3 sessions from his phone, check in later.
His workflow: start in plan mode
→ lock the plan
→ auto-accept edits
→ done.
No fancy setup.
No complex tooling.
Just multiple Claudes, a good plan, and a shared knowledge base.