Posit PBC

Heading to CDAO Government next week? We'll be at booth #1 on September 22-23 in Washington, D.C. If your agency is working through what open source looks like under real security and compliance requirements, that's the conversation we're there for. Worth stopping by if you're thinking about any of this: - Running R and Python alongside an existing SAS estate, without a rip and replace - Reproducibility and audit trails for analyses that inform policy decisions - Where agentic AI fits into government work, and where it does not yet - Package governance and software supply chain security for open source The short version of what we do: Posit gives government data teams the open-source tools they actually want to use, inside an environment that can tell you who ran an analysis, whether human or agent, what data they used, and what code produced a result. Trevor Chin and Mike White will be at booth #1 both days. Book time before the event and we'll have something ready to show you. https://lnkd.in/eYKYdVFW #CDAOGov #PublicSector #DataScience #GovTech

For years, giving your data scientists the IDE they actually wanted meant moving data somewhere your security team couldn't see. That tradeoff is over. Positron on Amazon SageMaker is now in public preview. Positron, Posit's R and Python IDE, launches as a one-click image inside Amazon SageMaker, with R, Python, and Quarto ready to use. Everything runs on your own SageMaker compute, in your own AWS account, under the IAM roles you already have. Data and credentials never leave your boundary. Posit Assistant is included and on by default. It authenticates through your SageMaker execution role and uses the models your admins have already enabled in Amazon Bedrock. No separate AI credentials for anyone to manage. Announced with AWS at posit::conf last week, and the latest step in the Posit and AWS Strategic Collaboration Agreement. 👉 https://lnkd.in/eprg5YXH #RStats #DataScience #AmazonSageMaker #Positron #AWSPartners

Thank you again for joining us at posit::conf(2026), and thank you to our platinum sponsor Amazon Web Services (AWS). The agenda was full, but the conversations made the week. Another HUGE thank you to Jagjit Dhaliwal (JD), James Blair, Subodh Kumar, Abhishek M., Sam McIntyre, and Sri Aakash Mandavilli for a keynote announcement, two sessions on running Posit tooling on Amazon SageMaker, three days of demos, and one great 80s party. We'll remember the energy you all brought. 🎉 Plus, read the announcement: Introducing Positron on Amazon SageMaker → https://lnkd.in/eprg5YXH #positconf #AWS #AmazonSageMaker #Positron

That's all, folks, posit::conf is outta here! 👋 Day 2 of the main posit::conf(2026) event began with Wes McKinney's keynote, Navigating the Agentic Future, where he discussed how we humans, the real agents, can reap the benefits of using this new, powerful, but flawed AI tech. The following talks covered a wide span of relevant and timely topics, from almost-PowerPoint-like slidecrafting with Quarto to extremely useful data viz to real-world Shiny apps, among much more. The day concluded with a delightful lightning talk round from the Posit team. My favorite: how your high-functioning team may or may not be like a termite colony. What a radical conclusion! Save the date: posit::conf(2027) is taking over Seattle, WA from Sept 13-15. Speaker announcements, session details, and registration information will be coming in early 2027. Until then, keep it real.

Your published work is not going anywhere. It is getting a better home. The "Migrating to Connect Cloud" webinar is now up on YouTube. Alex Chisholm walks through how shinyapps.io, RPubs, and Quarto Pub converge into Posit Connect Cloud, the timeline through January 2027, and a live demo of the interactive migration tool that maps your plan and moves your content for you. Worth an hour if you want: - A clear answer on when your content moves and what you need to do before then - The shinyapps.io to Connect Cloud plan mapping, including the new annual and organization options - To watch the migration tool actually run before you use it yourself - Access controls and secure sharing, which RPubs and Quarto Pub never had Free shinyapps.io users move in January 2027. Everyone else has a mapped path and a tool that handles the move. If you have been putting off figuring out what this means for your apps, start here with this video. The live Q&A is included at the end. Watch it here: https://lnkd.in/e9kzV9jT #Posit #RStats #DataScience #ConnectCloud #Shiny

Day 1 of the main posit::conf conference was like, totally rad! 🎸✨ We kicked things off with an inspiring opening keynote by Simon P. Couch and Sara Altman on building correct, reproducible, and transparent data agents, a reminder of how far data science has evolved since the days of floppy disks. 📟💾 The rest of the day was packed with fantastic talks spanning notebooks, dashboards, small-team workflows, AI (of course), and much more. The talks culminated in Christine Y. Zhang's keynote, How to Be Clear When You Can’t Be Certain, featuring polls, squaring the circle, and of course... THE NEEDLE. We wrapped up the night in true 80s style: big hair, retro jams, and pure synth-wave energy at the evening party! 🎤⚡️ Can't wait to see what Day 2 brings! Which talks are you most looking forward to? We'll see you there.

Generative AI can hand you a plausible answer in seconds. Getting the same answer twice is the hard part. Correct and verifiable are different problems. We built Posit Workbench, Package Manager, and Connect to solve this challenge, so your data, code, and environment stay inspectable, not just confident-sounding. Prove it with Posit: https://lnkd.in/g4nTFpDB #DataScience #ResponsibleAI

Posit CEO Tareef Kawaf spends less time worrying about whether the models are good than about what people stop doing once the models get good enough. "And one of the big worries that I have right now is that I see increasingly a lot of people just like abdicating the judgment and the the deeper thinking to to the machine. And the problem is that the machine like will make bad decisions or will not even know that it's making decisions." The second half is the uncomfortable part. A bad decision leaves a trace you can go back and find. A decision nobody registered as a decision goes into the deck and then to the board, and the spreadsheet never gets opened again. On this upcoming episode of The Test Set, Tareef works through what it costs to hand your thinking over, and the one problem he hasn't solved: how do you get people to inspect the output? https://lnkd.in/eS6aYvqa

What a fantastic workshop day to kick off posit::conf(2026)! 🚀 We had workshops from the Positron IDE to AI and next-gen clinical reporting in pharma, the energy yesterday was unmatched. Day 1 kicks things off with Simon P. Couch and Sara Altman’s keynote, “Correct, reproducible, and transparent data agents” (a talk that couldn't be more timely). We feel that as AI and automation continue to transform data science, we must make sure our tools stay reproducible, verifiable, and transparent. Their talk covers HOW to do this in detail. post::conf is the best place to learn, connect with the #RStats and #Python communities, and bring back fresh ideas! Who else is here at posit::conf? What sessions are you most looking forward to? And if you haven't registered yet, it's not too late! Join us virtually: https://conf.posit.co. #PositConf2026

If you're the person on your team who gets asked how the AI arrived at something, our demo on September 30th is tuning into. Our monthly workflow demo is going deep on custom MCP servers this month, which is how you connect an application to external tools, databases, and data sources without stitching it together by hand every time. We're walking through two real ones running on Posit Connect: a CDISC conformance validator that wraps an existing rules engine, and another that uses a pinned, separate LLM to classify adverse events. Register here: https://lnkd.in/dSz8Tk_4 The examples are based on clinical submissions, but the workflow is the same for anyone. A rules-based check against a standard, or free text mapped to a taxonomy, gets built the same way regardless of what industry you're in. Live Q&A at the end, which is usually the best part. 😌 See you there!