Posit PBC

Many of our LLM tools can execute code and access your files. What does this mean for privacy? As AI assistants like Positron Assistant and Databot become more integrated into our daily work, they naturally raise questions about where data goes and who can access it. Simon P. Couch and Sara Altman share thoughts for data science teams and leaders on navigating the critical intersection of data privacy and AI tools. Their blog post addresses these concerns by shifting the focus from "hiding data" to building a transparent trust relationship with your model provider. Read it here: https://lnkd.in/gScejmun

Imagine this: you have survey data from hundreds of respondents, and you need to generate individual PDF reports for all of them,  in recurring waves. Sounds daunting, right? That was the challenge facing the Decision Sciences & Innovation team at KS&R. By rebuilding their survey reporting pipeline with Quarto and Posit Team, KS&R turned a manual, time-intensive process into an efficient, reproducible workflow, now generating 100 reports in just 5–7 minutes. 🚀 The quote featured below comes from Ben Cortese, and the reproducible example shared alongside this work was led by Keaton Wilson, PhD, who played a key role in designing and implementing this workflow. Ready to do the same in 2026? 👉 Check out this reproducible example from Keaton Wilson at KS&R here: https://lnkd.in/e6WBY_B3 👉 Read the full customer story: https://lnkd.in/effZf-HX

Announcing gt 1.2.0: Enhanced Table Construction for Clinical Reporting We are pleased to announce the release of gt v1.2.0, a package designed to streamline the creation of publication-quality tables for R users! The updates in v1.2.0 are the result of a close collaboration between Posit and GSK, focusing on the complex reporting requirements often found in clinical and pharmaceutical industries (but that can benefit the larger user base, as well). • You can now create hierarchical row labels by passing multiple columns to the stub, which is designed for representing parent-child relationships in clinical trials or financial reports, with automatic value consolidation for a cleaner visual hierarchy. • The new summary_columns() function enables horizontal calculations (like row totals or averages) across multiple columns, complementing the existing vertical summary_rows() functionality. • With fmt_number_si(), you can automatically format values using SI prefixes (k, M, G, etc.).  • Significant improvements have been made to non-HTML outputs. Whether you are exporting to Word, LaTeX, or RTF, the new version ensures better text alignment, proper handling of Unicode characters, and consistent rendering of complex structures, such as column spanners. The gt package continues to evolve as a comprehensive tool for data communication, and we look forward to you trying it out. Learn more about gt 1.2.0 and its full list of features: https://lnkd.in/gaybJsTd

Databot: The AI Assistant for Accelerated Exploratory Data Analysis (EDA) 🤖 In case you missed it, we have released the research preview of Databot, a new AI assistant designed to dramatically accelerate the exploratory data analysis workflow. Databot is built specifically for data scientists fluent in Python or R and works as an extension within the Positron IDE. We designed it to be a powerful complement to the experienced data scientist, not a replacement. The process of open-ended data exploration can often take hours. Databot addresses this challenge by writing and executing its own code on the fly, allowing you to derive reliable, serendipitous, and transparent insights in minutes. • Reduce the time required for comprehensive EDA, transforming hours of manual work into a focused few minutes. • The user experience is intentionally code-forward, requiring careful review of generated code to ensure trust and reliability in the analysis. • Built to support data scientists working across the R and Python ecosystems. • Databot is a specialized, purpose-built agent for EDA, focused on rapid iteration and insight generation. Learn more about Databot, its research preview status, and how to get started in Positron: https://lnkd.in/e-5thnH8

✨ 2025 has been an incredible year of innovation, partnership, and impact in the world of data science—and it’s all thanks to you. ✨ We are so grateful to be the toolkit for your most important work. As you wrap up your final projects and "bundle up" for the season, we want to thank you for being the heart of the Posit community. From all of us at Posit, Happy Holidays! We hope you have a cozy break and a brilliant start to 2026. 🎁 ☃️   #HappyHolidays #DataScience #Posit #OpenSource

🚜 DairyNZ is putting cutting-edge data into the hands of 11,000+ farmers to drive profitable and sustainable businesses. Using Posit and Snowflake, they built innovative, farmer-facing tools like: ✅ Econ Tracker: Quarterly budgets and forecasts (down from 12-24 months!). ✅ Pasture Potential Tool: Interactive dashboards to measure farm productivity. ✅ Connected Farm Prototypes: Predictive tools combining wearable sensor data ("Apple watches for cows"!) and weather to mitigate heat stress. This robust data pipeline is powered by the scalable infrastructure of Posit Workbench/Connect and the trusted data in Snowflake. It's how they're delivering near real-time value and establishing the foundation for AI-driven animal science. Read the full story now! 👉 https://lnkd.in/gVSD26px #AgriTech #Posit #Snowflake #AI #DairyNZ

🗞️ Last Posit AI Newsletter of 2025! Curated by Posit’s Sara Altman and Simon P. Couch, this biweekly newsletter keeps you up to speed with what’s happening in the AI world, both at Posit and beyond. It’s opinionated, concise, and designed to keep you in the loop without overwhelming you. ❄️ This week's edition is a year in review: - A look back at 2025 in AI and data science - All the terms we've defined so far - Plus: chores 0.3.0, and more! 📬 Read the newsletter: https://lnkd.in/ejTW-Qqu

TruDiagnostic hit a critical bottleneck due to disconnected R and Python workflows and a heavy reliance on engineering support. These infrastructure challenges threatened to derail their mission of advancing biological aging research. By unifying their team on Posit Workbench with AWS SageMaker, they achieved instant transformation: ➡️ 1 Year of product development time saved. ➡️ 1,600 models trained in 1 hour (down from months). ➡️ 60% reduction in cloud infrastructure spend. Discover how TruDiagnostic empowered its researchers and transformed a bottleneck into a competitive advantage. 👉 https://lnkd.in/gnbbBE5r #TruDiagnostic #AWS #Posit #DataScience

Local LLMs for coding: Are we there yet? 🛑 The short answer: Not quite. Many developers want to use local models with tools like Positron Assistant for privacy and cost reasons. But a new deep dive by Simon P. Couch and Sara Altman shows that the trade-offs might be steeper than you think. Local models still struggle with complex reasoning compared to hosted giants. Running capable models locally requires serious (and expensive) GPU power, and you encounter higher rates of hallucinations and "lazy" coding. The verdict: If code quality and speed are your priorities, hosted models are still the way to go (for now). Check out the data and methodology behind their findings:https://lnkd.in/g8CxcQmq

Posit Workbench, part of the Posit Team platform, provides significant value to financial service organizations by delivering a centralized, auditable, and flexible development environment. It achieves this by providing three key areas of support: ✔️ Flexible IDEs and centralized project management ✔️ Guaranteed reproducibility  ✔️ Resource management and security Learn how Posit Workbench can help you overcome dependency management, access consistency, and operational risk. Read the blog: https://lnkd.in/eJPWdveZ