AI UNDERDOGSDAILY PICK
AI UNDERDOGS
AI 实盘两年 +135%
+135% in 2 years, real money
xbtlin/ai-berkshire
AI 帮他实盘跑两年
收益率 135%——这次不是 demo
是真钱在赌
Two years of live trading with AI, returning
135% — not a toy demo, actual money
on the line
四大投资大师对冲分析
4 masters, adversarial analysis
巴菲特、芒格、李录、段永平
每人一套独立分析框架
四个 Agent 从不同角度互相对冲
最后合成一份研究报告——散户以前根本不可能同时跑这四套逻辑
Buffett, Munger, Li Lu, Duan Yongping — each
has a dedicated agent with a distinct analytical
checklist. Four agents run adversarial analysis and converge
into one report. Retail investors could never run
these frameworks in parallel before
★ SIGNAL 1
不是挂名提示词
Not a name-drop prompt
不是在 prompt 里写一句「请你扮演巴菲特」就完事
每个大师 Agent 有完整的检查清单
从护城河、管理层
安全边际不同维度拆解
最后四份报告互相对照
有矛盾的地方会明确标出来
Not just 'act like Buffett' in a prompt
Each agent has a full checklist — moats
management, margin of safety — and the four
reports cross-check each other. Contradictions get flagged explicitly
★ SIGNAL 2
1378 次提交，真钱验证
1,378 commits, proven with real money
1378 次提交
15500 个 Star
茅台的十年利润预测模型最近刚更新——这不是写完就扔的项目
是持续在迭代的活系统
两年实盘 +135%
这比任何 benchmark 都有说服力
1,378 commits, 15,500 stars. A Moutai 10-year profit
projection just landed — this isn't a write-and-forget
repo, it's a living system. Two years of
live trading at +135% beats any benchmark
作者自己说的：「两年实盘 +135%
让我带你看看我用 Claude Code 搭的这个多 Agent 研究框架」——不是卖课
GitHub 仓库完全开源
自己去看代码
The author says: 'Two years of live trading
at +135% — let me walk you through
the multi-agent research framework I built with Claude
Code.' Not selling courses — it's fully open
source on GitHub, go read the code yourself
AI UNDERDOGS
xbtlin/ai-berkshire
xbtlin/ai-berkshire
xbtlin/ai-berkshire
关注 · 每天发现更多 AI 神作
github.com/xbtlin/ai-berkshire