
AI Berkshire
github.com/xbtlin/ai-berkshire →Two years of live trading with AI: +135% returns
WHAT IT SOLVES
Value investing analysis is expertise-heavy — retail investors can't realistically run Buffett, Munger, Li Lu, and Duan Yongping's frameworks in parallel to cross-check a single company
WHY IT'S INTERESTING
Four masters' playbooks, not decoration
Buffett, Munger, Duan Yongping, and Li Lu each get their own agent with a distinct analytical checklist. It's not a name-drop prompt — agents run adversarial analysis from different angles and converge into one research report
1,378 commits, battle-tested with real money
The +135% live track record over two years is the hardest proof. 1,378 commits, 15.5k stars, and reports still rolling in (Moutai 10-year profit projection just landed) — this isn't a demo, it's a living system
「Two years of live trading at +135% — let me walk you through the multi-agent research framework I built with Claude Code」
TECH GUESS
Claude Code / Codex driving a multi-agent workflow; prompt-layer defines each master's analytical framework, likely Python for data scraping and report generation
Discussion (0)
- No comments yet — be the first.
Related
#229▶ 135Upload your textbook. AI teaches it back to you, chapter by chapter
#228▶ 111Someone needs to watch the AI spend meter
#227▶ 183Your users are kittens. Don't let them fall off the cliff