Supervised Feed

Math Refresher for Adults — The Perfect Solution for Mastering Essential Math Skills: amzn.to/4wETqS7 There are lots of reasons why you might want to have this book, no matter what your math proficiency might currently be. See the list of reasons on the book cover. My reason: I tutored students in math when I was in high school and college, I took lots of math courses (including honors math classes for math majors) in college, I learned high-level math (group theory, differential geometry & forms, tensor calculus) from Nobel laureates in theoretical physics classes in grad school at Caltech, I taught math in my computational & data science courses as a university professor, I guided students who feared math into doing and loving calculus (conquering their fears and converting some into math lovers), and I sat for countless hours with my own children through many years of math homework & exam prep in their school years. My latest target audience are my grandkids. A book like this keeps me grounded in these new opportunities to share the love and beauty of math with a new generation. Math Lovers-Я-us!

Mistral Large 4 is now open weights, narrowly making it the #1 open weights model developed outside of China @MistralAI has updated and released the weights of their flagship model, Mistral Large 4, with 1.05T total and 52B active parameters. With high reasoning, Mistral Large 4 scores 26 on the Artificial Analysis Intelligence Index v5, up 1 point from Mistral Large 4 Preview (25), with the major differences being increased token efficiency and reduced hallucinations. It sits 1 point ahead of the leading US-developed open weights model, Muse Spark Open (xhigh, 25). On the Artificial Analysis Cyber Index, Mistral Large 4 (high) scores 51, third among current open weights models behind DeepSeek V4.1 Flash 1017 (max, 57) and MiMo-V2.6-Pro (56). Key takeaways: ➤ Narrowly leads non-Chinese open weights models: Mistral Large 4 (high) leads Muse Spark Open (xhigh), the next-best open weights model developed outside China, with the largest gap on Terminal-Bench 4.0 (25% vs. 8%). It trails leading open weights models DeepSeek V4.1 Flash 1017 (max) and GLM-5.3 (max) by 6 points on the Intelligence Index, largely due to weaker coding and agentic knowledge work performance: 15% on FrontierCode 2.0 (vs. 29% and 22%, respectively) and 1349 Elo on AA-Briefcase (vs. 1570 and 1504) ➤ Increased token efficiency: Mistral Large 4 (high) uses 82K output tokens per Intelligence Index task, 39% fewer than Preview (135K), placing it on the open weights Pareto frontier. The seven open weights models that score higher use between 113K and 225K output tokens per task ➤ #2 open weights model on the Coding Agent Index: On the Artificial Analysis Coding Agent Index, Mistral Large 4 (high) scores 48 in Mistral Vibe while using 8.5M tokens per task, about half as many as the leading open weights model GLM-5.3 (max) in Opencode (49, 15.9M). It also scores above another leading open weights model, DeepSeek V4.1 Flash 1017 (max), in DeepSeek Harness (46, 16.7M) ➤ Reduced hallucinations: Mistral Large 4 (high) has a 28% hallucination rate on the updated AA-Omniscience v1.1, down from 41% for Preview and matching Muse Spark Open (xhigh). However, its accuracy remains relatively low at 26%, similar to Ling 3.1 Flash 1019 (27%), and behind DeepSeek V4.1 Flash 1017 (max, 40%) and Muse Spark Open (xhigh, 42%) ➤ Lower price and cost per task than the Preview: Mistral Large 4 is priced at $0.70/$2.10 per 1M input/output tokens, a ~50% reduction from the Preview's pricing of $1.36/$4.18. At the new pricing, it costs $2.06 per Intelligence Index task, down 63% from $5.54 for the Preview. This is comparable to Muse Spark Open (xhigh, $2.02) but more costly than leading open weights models DeepSeek V4.1 Flash 1017 (max, $0.60) and MiMo-V2.6-Pro ($0.26) Key model details: ➤ Size: 1.05T total parameters / 52B active parameters ➤ Context window: 1M tokens ➤ Multimodality: Text and image input, text output ➤ Pricing: $0.70/$2.10 per 1M input/output tokens ($0.07 per 1M cached input tokens) ➤ License: MEOW License (Mistral Ecosystem Open-Weight License), where operating a Model-as-a-Service business requires a separate commercial license with Mistral Congratulations to the team at @MistralAI on the release!

these 10 github repos are the most DANGEROUS quant stack on the internet Quant funds spend millions on data terminals, research desks and execution engines, these 10 repos cover every piece of it for free That's 329k+ stars of code that hedge funds would rather you never put together Here's every repo and what it does: 1. OpenBB (75k stars): an open data platform for analysts, quants and AI agents, prices, options chains, macro and fundamentals from one place github.com/OpenBB-finance/Op… 2. TradingAgents (118k stars): a multi-agent framework that mirrors the dynamics of real-world trading firms, AI analysts write reports, bull and bear researchers debate, a trader decides and a risk team signs off github.com/TauricResearch/Tr… 3. qlib (50k stars): Microsoft's AI-oriented quant investment platform, from exploring ideas to production github.com/microsoft/qlib 4. RD-Agent (15k stars): Microsoft's AI researcher that proposes factors, writes the code and backtests them, in their NeurIPS 2025 paper it hit up to 2x higher annualized returns than classical factor libraries in backtests, with 70% fewer factors github.com/microsoft/RD-Agen… 5. gs-quant (14k stars): the quant toolkit created and maintained by Goldman Sachs' own quants, pricing, risk and timeseries math github.com/goldmansachs/gs-q… 6. nautilus_trader (31k stars): a production-grade, Rust-native trading engine with a deterministic, event-driven architecture github.com/nautechsystems/na… 7. Chronos (6k stars): Amazon's pretrained models for time series forecasting, zero-shot, so they forecast data they were never trained on github.com/amazon-science/ch… 8. QuantLib (8k stars): the free, open source library for quantitative finance, modeling, trading and risk management in real life, from options to bonds and swaps github.com/lballabio/QuantLi… 9. skfolio (3k stars): portfolio optimization, factor models and risk management built on scikit-learn github.com/skfolio/skfolio 10. vectorbt (10k stars): runs thousands of trading ideas before others finish one github.com/polakowo/vectorbt Here's how they work together: Data: OpenBB feeds the whole stack the same prices, options and macro data Forecast: Chronos predicts where every series is likely to go next Alpha: RD-Agent mines new factors inside qlib, so research runs on its own Decisions: TradingAgents turns forecasts and factors into trades, with a risk team that has to sign off Risk: QuantLib prices the options, gs-quant runs Goldman's risk math on everything Portfolio: skfolio sizes the positions so no single bet can sink you Testing: vectorbt stress-tests every version of the strategy, fees included Execution: nautilus_trader runs it, paper first, always Here's how you can run it for free yourself: 1. Open Claude Code and clone all 10 repos 2. Have Claude pull 5 years of S&P 500 prices, options and macro data with OpenBB 3. Run Chronos on every stock to forecast the next 20 days 4. Let RD-Agent mine new factors in qlib on the same data 5. Feed the forecasts and factors to TradingAgents and keep only the trades its risk team approves 6. Price the options legs with QuantLib and run gs-quant's risk math: volatility, drawdown, beta 7. Build the portfolio in skfolio with a max drawdown limit 8. Backtest 10,000 versions of the strategy in vectorbt and keep the ones that survive fees 9. Paper trade the winner in nautilus_trader for 2 weeks, if it doesn't beat buy-and-hold, you don't have an edge yet i won't say anything more: