Anaconda, Inc.

AI coding tools often struggle with conda because they can’t see your actual environment, installed packages, or channel configuration. Anaconda MCP changes that by giving MCP-compatible tools direct access to the real environment state, so actions like listing environments, creating new ones, and resolving package dependencies occur through conda rather than guesswork. This full demo shows how it works with Claude Desktop, including a guided setup and hands-on scenarios → https://bit.ly/3SLYoOm

A century before modern AI, Andrey Markov manually counted 20,000 letters from a Pushkin poem to show that predictable patterns can emerge from sequences where each item depends on the one before it. That work became the first Markov chain. Claude Shannon later applied the same idea to English sentences, and the core concept still lives on in today’s language models: predicting the next symbol from the context that came before it. Anaconda’s Senior DevRel Engineer Daina Bouquin is back with another #PythonTips, pulling back the curtain on modern data science and AI development.

The latest #NumericallySpeakingLIVE is now available on-demand! Anaconda’s Daina Bouquin and Nicole Schwartz explore Python profiling from the ground up, using intentionally inefficient code to show how profiling tools can help identify performance bottlenecks and reveal where code spends its time. They also share details about a Python profiler project in progress and invite community feedback. Catch the full episode on Anaconda's Video Hub: https://bit.ly/4w56UGE

Now available on demand: a fireside chat from Anaconda and Omdia featuring Mark Beccue, David DeSanto, and Ville Tuulos on building trusted AI workflows at enterprise scale. Learn why AI pilots often stall before production, what teams underestimate when scaling AI, and how governance and observability can help organizations deploy AI with confidence. A must-watch for technology, security, and business leaders navigating the path from experimentation to impact. Watch now: https://bit.ly/3SLXDVw

As organizations move AI agents into production, governance becomes just as important as capability. 🔐 In this on-demand webinar, Micheal Lanham, author of AI Agents in Action, 2nd Edition discusses how guardrails help create dependable agentic systems, covering validation techniques, agent-to-agent interactions, and governance practices that support oversight without limiting innovation: https://bit.ly/4fALLiG

Summer's here, and so is the latest issue of Numerically Speaking! Here's what you can find in our June edition: ⚡ Our CEO built a local AI agent in under 3 minutes — no cloud required 🤖 NVIDIA Nemotron 3 Ultra (550B parameters) is now available on Anaconda Platform 🐍 Anaconda Desktop Beta: Agent Studio + MCP bring AI to your Python workflow 📉 OpenAI weighs drastic price cuts as competition with Anthropic heats up ⏳ Last chance to download our Gartner® report on securing your AI supply chain Explore the full edition and don't forget to subscribe: https://bit.ly/44Awast

Anaconda’s recent acquisition of Outerbounds (acq. by Anaconda) reflects a broader shift in AI development: as AI systems become increasingly probabilistic and adaptive, the challenge is no longer just building them — it’s understanding, governing, and operating them. Reliable AI requires a deterministic foundation for environments, dependencies, workflows, and governance. With Outerbounds, Anaconda addresses that gap. Learn more → https://bit.ly/4vGQFiG

AI development is becoming AI-native. As organizations move beyond experimentation, the challenge shifts from model access to operationalizing AI systems that are reproducible, observable, governed, and scalable. Explore how our thinking has evolved and how recent investments across workflows, developer experience, and security are helping teams build and operate AI-native systems with confidence → https://bit.ly/4xoaiOd

New research shows AI coding tools dramatically increase code output and pull requests, but software releases lag behind due to bottlenecks in testing, validation, review, security, and deployment. Anaconda VP of Engineering and AI, Greg Jennings recently spoke with IT Brew’s Brianna Monsanto about these human bottlenecks: “Delivered code is not delivered value, it’s something that’s going to help us deliver more value, but now we have to pay more attention to the other steps in the process.” The takeaway: AI is accelerating development, but organizations need to optimize the entire software lifecycle (not just code generation) to turn velocity into shipped software. Read the full article: https://bit.ly/4emVxUA

A common PackagesNotFoundError workaround is pulling packages from PyPI or other sources outside existing governance controls. 🔓 Anaconda’s new main-x channel, now in Open Beta, adds ~6,000 pure Python packages built from source with the same validation and testing standards as main. Available with a free Anaconda account, main-x is designed for development and evaluation during Beta. Learn more: https://bit.ly/4v4B1hb