Coding agents have gotten really good at one kind of work. You scope a feature, edit some files, run the tests, ship it. It all happens on disk. But that is not how data work feels. You load something, you look at it, you run a cell, you watch how it responds, and you decide the next move from whatever is sitting in memory. And until now, your agent couldn't see any of that. It only saw the files. Never the live state. This episode, that wall comes down. marimo pair drops a coding agent right inside a running notebook, with full access to every variable Python is holding in memory. The notebook becomes a shared canvas. You point, it runs the code. You tell it to zoom in on the Picasso paintings, and the chart just updates. No MCP tools to wire up, no schema to describe. Just Python, and an agent that can finally see what you see. Trevor Manz is back to walk us through it.

You ask an AI a question and it answers with total confidence. Most of the time, a confidently wrong answer is just an annoyance. But what if the question is medical, and there's a real patient on the other end? In that world, a hallucination isn't a bug, it's a patient-safety event. Sumit Gundawar is a London-based software engineer who builds the clinical platform for a UK longevity and aesthetic-medicine clinic, and his whole argument is that in high-stakes AI, the model is the easy part. Earning trust is the real engineering. We dig into grounding, refusal logic, human-in-the-loop design, and the messy frontier of longevity and biohacking, plus a live demo of an assistant that refuses to answer when it can't back up the claim. Let's get into it.

This episode is a fun crossover from our Python news and tips podcast, Python Bytes. We have had some big changes over there. Brian Okken has moved on and Calvin Hendryx-Parker has joined the show as the new co-host. To kick off this new era, we decided to do a longer and more personal episode called "All Our Tools". The idea is both of us talk about some of our most useful day-to-day developer and business owner tools that we think you all would find useful. It was so well received, that I'm bringing it to you all as a crossover episode. Enjoy and we hope you find something new and awesome to help you with your software and data science day to day.

OpenAI just acquired Astral, the company behind uv, Ruff, and ty. And if your first thought was "wait, is uv toast?", you are not alone. But here's the twist Charlie Marsh shared with me: he thinks they may ship more open source at OpenAI than they ever did at Astral. On this episode, we get into the acquisition, the mixed feelings, the future of your favorite Python tools, and what it's like to build right at the center of the AI universe.

If you've ever been to PyCon, you know one of the best parts of the expo hall is Startup Row, a stretch of booths where early-stage companies built on Python show off what they're creating. But only attendees get to walk that lane, so let's bring it to everyone. In this episode, we stroll down Startup Row together. We kick things off with the organizers, Jason and Shay, who share the program's origin story going back to Paul Graham and the PSF, plus some surprising stats, including two unicorns among the alumni. Then we meet five startups: Tetrix, bringing AI to institutional investing in private markets. Arcjet, security that lives inside your app as an SDK. Phemeral.dev, serverless hosting built for Python web apps. CapiscIO, an identity and authority layer for AI agents. And Pixeltable, a multimodal database from Marcel Kornacker, co-creator of Apache Parquet. See if you can spot the theme running through them all. Let's go for a walk.

You wake up, brew the coffee, open GitHub, and there it is. Another pull request on your open source project. Thirteen thousand lines added. No issue filed first. No discussion. Just "here, please review this for me." Over the past year, GitHub activity has spiked roughly twelve times in a few short months, and a huge chunk of that signal is landing on the same small group of maintainers who were already stretched thin. The curl bug bounty got buried under AI-generated noise. Jazzband, the home of Django classics like pip-tools and the Django debug toolbar, hit what its maintainer called an "apocalypse" and started sunsetting. Even CPython just shipped fresh guidelines on AI-assisted contributions this week. So what does all of this actually look like from the receiving end of the pull request? On this episode, Paolo Melchiorre joins us to tell that story from inside the maintainer's chair. Paolo is a director of the Django Software Foundation, an organizer of PyCon Italy, a Django Girls coach, and he has spent the past year carefully collecting examples of how AI is reshaping open source contributions. The good, the bad, and the extra fingers. We dig into his PyCon US talk on AI-assisted contributions and maintainer load, why AI is best understood as an amplifier rather than a new kind of contributor, the wildly different policies across 86 open source foundations, whether projects banning AI today are reacting to last year's models.

Your documentation has two audiences now - humans reading the rendered HTML, and AI agents trying to make sense of your library. Rich Iannone and Michael Chow from Posit are back on Talk Python with a brand new Python documentation tool called Great Docs that takes both seriously. Rich is the creator of Great Tables, and before that the R package GT, the man has a serious eye for design, and he's pointed that energy at the Python docs ecosystem. We'll talk about how Great Docs spins up a polished site in three commands, why every page ships as Markdown for your favorite LLM, how it leans on Quarto for executable code blocks and tabbed install sections, and where it lands against Sphinx, MkDocs, and Zensical. Plus, you'll meet Tablin. Here we go.

What if your database worked more like Git? Every change captured as an immutable event you can replay, instead of a single mutating row that quietly forgets its own history. That's event sourcing, and Chris May is back on Talk Python, fresh off our Datastar panel, to walk us through what it actually looks like in Python. We'll cover the core patterns, the libraries to reach for, when not to use it, and why event sourcing turns out to be a surprisingly good fit for AI-assisted coding.

When OpenAI trained GPT-3, they didn't roll their own orchestration layer. They used Ray, an open source Python framework born out of the same Berkeley research lab lineage that gave us Apache Spark. And here's the twist: Ray was originally built for reinforcement learning research, then quietly faded as RL hit a wall. Until ChatGPT showed up. Suddenly reinforcement learning was back, as the post-training step that turns a raw language model into something genuinely useful. Edward Oakes and Richard Liaw, two founding engineers behind Ray and Anyscale, join me on Talk Python to tell that story. We'll trace Ray from its RISE Lab origins at UC Berkeley to powering some of the largest training runs in the world. We'll talk about what Ray actually is, a distributed execution engine for AI workloads, and how a few lines of Python become work running across hundreds of GPUs. We'll cover Ray Data for multimodal pipelines, the dashboard, the VS Code remote debugger, KubRay for Kubernetes, and where Ray fits alongside Dask, multiprocessing, and asyncio. If you've ever stared at a single-machine Python script and thought, "there has to be a better way to scale this", this one's for you

The cloud is convenient until it isn't. You upload your photos, sync your contacts, click through the cookie banners. Then prices go up again or you read about a family that lost their entire Google account over a medical photo sent to a doctor. At some point, the question shifts from "why would I run this myself?" to "why aren't I?" My guest this week is Alex Kretzschmar, head of DevRel at Tailscale, longtime host of the Self-Hosted podcast, and co-founder of Linuxserver.io. We cover what self-hosting really means in 2026, the apps worth running yourself like Immich and Home Assistant, why Docker Compose ties it all together, and how Tailscale lets you reach any of it from anywhere, without opening a single port. If you've been thinking about pulling your digital life back behind your own walls, this is your roadmap.