AI
JetBrains shelves Kotlin Notebook as AI-native tools reshape developer workflows
Two major tech firms have abandoned notebook interfaces for non-Python languages, citing shifting developer practices and the rise of AI-powered coding tools.
The New Stack
JetBrains has announced it will discontinue Kotlin Notebook, its interactive coding plugin launched three years ago. Beginning with IntelliJ IDEA 2026.2, the tool will be unbundled from the IDE, handed to the open-source community under an Apache 2.0 license, and receive no further updates from the company.
The move mirrors Microsoft's February decision to deprecate Polyglot Notebooks, a VS Code extension enabling C# and other languages to run within the Jupyter format. Both companies frame their departures similarly: developer workflows are changing in the age of AI-assisted coding.
Why notebooks failed outside Python's domain#
The underlying issue may be less about AI displacement and more about cultural fit. Jupyter notebooks originated in the Python and data science ecosystem, where exploratory analysis, inline visualization, and collaborative documentation form a natural working style. When JetBrains and Microsoft extended the notebook interface to Kotlin and C# developers, they encountered communities with different development patterns and priorities.
"Kotlin Notebook didn't reach the level of adoption we expected," said Marco Behler, JetBrains' head of the JVM ecosystem, adding that "AI tools have changed how developers explore code, prototype, and iterate."
Application developers in these languages never embraced notebooks the way data scientists embrace Jupyter, leaving both efforts underpowered by weak adoption.
Jupyter and Colab thrive while others fade#
Jupyter itself remains robust. GitHub's Octoverse report showed Jupyter Notebook repositories grew 75% year-over-year, reaching 2.42 million projects. AI-tagged repositories using Jupyter nearly doubled, now numbering over 400,000 — demonstrating the format's centrality to machine learning and experimental work.
Google Colab has followed a different trajectory. Built explicitly as a Python-first environment for data scientists and ML researchers since 2017, it avoided the audience problem that derailed Kotlin and Polyglot notebooks. Rather than abandoning the notebook model, Google has been integrating AI features directly into Colab, rolling out an "AI-first" version in June 2025 that embeds agentic capabilities into the interface.
