Oren Tapiero

I HAVEN’T OPENED A SINGLE BOOK THIS MONTH. CLAUDE READ 34 OF THEM FOR ME I built a vault in Obsidian, connected Claude and added Notebooklm went to sleep woke up with flashcards, mind maps, and an AI that answers from my own notes only my friends are still highlighting pdfs manually the only thing separating us is one setup: claude + obsidian + notebooklm every idea captured. every book processed. zero knowledge lost if this is the setup you’ve been looking for - like & bookmark so you don’t lose it

🔖 Claude Cookbook is worth bookmarking. 81 practical guides across 15 categories, covering agents, tools, RAG, evals, multimodal apps, Skills, integrations, and production workflows by @AnthropicAI platform.claude.com/cookbook/

BREAKING: ANTHROPIC JUST OPEN SOURCED THE ENTIRE WALL STREET WORKFLOW. DCF models. LBO models. Equity research reports. Merger analysis. KYC checks. All of it. Free. On GitHub. It connects Claude directly to: -> Bloomberg, FactSet, S&P Global, Morningstar, PitchBook -> Builds real Excel models with live formulas and sensitivity tables -> Drafts CIMs, IC memos, earnings reports, and buyer lists -> Runs PE due diligence, GL reconciliation, and NAV tie-outs This is not a chatbot wrapper. These are production agents that own entire financial workflows. The kind firms pay $50,000 to $500,000 per year in software to run. Now it is a one-line Claude Code plugin install. 19.8K GitHub stars. Apache-2.0 License. 100% Open Source.

40% of the code Claude writes for you is wasted. you're paying for the rewrite. a 65-line markdown file fixes it. 120,000 developers have starred it. the author tested it on "30 codebases over 6 weeks" and reported a mistake rate drop from 41% to either 11% or 3% depending on whether you read the headline or the body. the irony is that the article is right. CLAUDE.md is the most under-leveraged file in your stack. 65 lines of behavioral rules outperform a 4,000-token preferences dump. "be careful" is useless. testable imperatives are gold. "be senior" doesn't work Claude already thinks it is. the 4 rules that ship the most leverage: / state assumptions, never guess silently / minimum code, nothing speculative / surgical changes, don't refactor adjacent code / define success, loop until verified compliance: ~80%. mistake rate: from ~40% to single digits. no human caught the contradicting numbers in the title. nobody had to.

🚨 BREAKING: NotebookLM can now teach you like a private university tutor. Here are 8 prompts that make studying 10x easier: (Save this before it blows up)

Anthropic CEO when he found out their Chinese competitor is 8x cheaper and open source

Researchers are publishing articles with AI-generated fake references to papers that don't even exist. So, we built a Reference Auditor that looks for hallucinations in any bibliography. Select "Reference Auditor" in Chat Academia. Paste your bibliograph and it will show you which papers are real and which hallucinated. Check it out at chatacademia.com

Jane Street pays $650,000 a year for quants Oxford wrote the exact trading bible to get there 130 pages. Market making, predictive signals, limit orders, portfolio construction. The same material that costs $80,000 in tuition if you enroll. I read it. Then I wrote the implementation guide for Polymarket. Every formula in the article below - LSTM architecture, Kelly sizing, VPIN detection, stationary features - maps directly to a chapter in this book. Read the Oxford paper for the theory. Read the article below for the execution👇

one line from this algorithmic trading textbook that changed how i think about strategy development: "successful algorithms generally exploit deviations from established benchmarks." you need to know what normal looks like, then systematically trade the deviations from it. from "Algorithmic Trading and Quantitative Strategies"

Academic Researchers right now:

כבר שלוש שנים אנחנו שומעים יום יום על כך שאחים לנשק אמרו שלא יאריכו שירות מילואים אם תעבור ההפיכה. אתמול חיילים דתיים סירבו פקודה לישון בבסיס עם חיילות. סירבו. מנהיג דתי אמר משהו? גינה? כמובן שלא. פקודות צה״ל זה לחילונים.

Paper (SSRN): papers.ssrn.com/sol3/papers.cf… Full write-up on Medium: @oren.j.tapiero/modelling-crypto-order-flow-imbalance-as-an-additive-multiplicative-process-or-detecting-whale-fc7a6bd858a9" target="_blank" rel="nofollow noopener">medium.com/@oren.j.tapier… Comments and pushback welcome.

Compare to FTX, October 2022 to February 2023. Both the multiplicative share AND β² were simultaneously large. The OFI channel explained 10.6% of realised squared returns in that episode. Roughly ten times its explanatory power in any other window. Necessary AND sufficient.

Built a whale detector for crypto using two OLS regressions and four-hour bars. No on-chain data. It fires during COVID and LUNA weeks ahead of the crash. It also fires all through 2024, when prices barely moved. The gap between those two facts is the whole point. ↓