Slide 1

Slide 1 text

K E Y N OT E Press Start: Python's Next Generation プレス・スタート:Python の次世代 Carol Willing PyCon JP · 2026 ▶ Tap to Start

Slide 2

Slide 2 text

おはようございます O H A Y Ō G O Z A I MA S U Good morning! Thank you for having me.

Slide 3

Slide 3 text

LV. 60 Player 1: Carol Wi ing • • • • • • Python core team and 3-term steering council member Jupyter core team and 2017 ACM Software System Award PSF Fellow, 2-term PSF board member PyLadies, pyOpenSci advisor, napari core team Interested in cognitive science, learning, and education ll Love empowering humans by scaling technology's impact

Slide 4

Slide 4 text

Python's next generation

Slide 5

Slide 5 text

I love games.

Slide 6

Slide 6 text

As a child, I played Atari video games. When I was 12 years old, I wrote code to play Breakout on an Apple.

Slide 7

Slide 7 text

• import pygame • Set up constants for game board • Create Paddle, Ball, Brick, Game • game.run() • Play! When I was 60 years old, I used pygame and Claude to write Breakout.

Slide 8

Slide 8 text

No content

Slide 9

Slide 9 text

Try porting Breakout to Pyxel. Learn more at tomorrow's keynote with Takashi Kitao. 詳細は、明日行われる北尾隆氏の基調講演でご確認ください。

Slide 10

Slide 10 text

? Why do we play games ? なぜ私たちはゲームをするのでしょうか?

Slide 11

Slide 11 text

? Why do we play games ? なぜ私たちはゲームをするのでしょうか? solve problems fun friends relaxation exploration achievement entertainment choice

Slide 12

Slide 12 text

tt Games help us become be er coders.

Slide 13

Slide 13 text

ll Games, like using Python, requires ski s.

Slide 14

Slide 14 text

Games teach us to adapt. We practice until we succeed. We can use AI tools to adapt and succeed.

Slide 15

Slide 15 text

Today's Quest Map 今日のクエストマップ 2 3 ★ Python Power-Ups Visualization Systems and AI Community Fast, flexible Python Massive data, interactive graphics New solutions with code The players behind it a l 1

Slide 16

Slide 16 text

Level 1 · レベル1 Python Power-Ups Pythonパワーアップ Faster performance and exibility improves everyday code. fl 1 / 3 levels

Slide 17

Slide 17 text

TRY 3.15 Insta uv and run Python in the terminal. uv python install 3.15 uv run python 3.15 python Fo r f re e - t h re a d e d b u i l d : uv run python 3.15t python - - - - ll https://docs.astral.sh/uv/getting-started/installation/

Slide 18

Slide 18 text

TRY 3.15 What's New in 3.15 tt h ps://docs.python.org/3.15/

Slide 19

Slide 19 text

TRY 3.15 PEP 810 Explicit lazy imports for faster startup time

Slide 20

Slide 20 text

TRY 3.15 PEP 814 Add new built-in type frozendict immutable hashable

Slide 21

Slide 21 text

TRY 3.15 Statistical profiling profiling.sampling Stack is sampled periodically. Snapshots of the call stack are saved. Functions that consume signi cant CPU time will appear frequently in the collected samples. Di erent than the deterministic pro ler, profiling.tracing (cprofile) fi fi ff tt h ps://docs.python.org/3.15/library/profiling.sampling.html

Slide 22

Slide 22 text

TRY 3.15 More performance Free-threaded Python and stable ABI abi3t. Upgraded JIT compiler. PEP 803

Slide 23

Slide 23 text

TRY 3.15 O cial release for 3.15 is coming soon. Thanks Hugo van Kemenade for managing the 3.15 release. O cial release date is October 1, 2026. h ps://docs.python.org/3.15/ ffi ffi tt tt h ps://devguide.python.org/versions/

Slide 24

Slide 24 text

Python 3.15 is HPC ready for scaling.

Slide 25

Slide 25 text

Level 2 · レベル2 Visualizations 可視化 Massive data and interactive graphics accelerate research and understanding. 2 / 3 levels

Slide 26

Slide 26 text

napari is a fast, interactive viewer for multi-dimensional images. Supports numpy, zarr, and xarray. Used for microscopy and geospatial. https://napari.org/

Slide 27

Slide 27 text

PyVista Tetsuo Koyama is a PyVista and visualization expert. See his talk today.

Slide 28

Slide 28 text

fi VisPy is the current stable Python library for high-performance, interactive 2D and 3D scienti c visualization. It uses OpenGL to display large datasets and integrates with Python GUI and notebook environments.

Slide 29

Slide 29 text

datoviz WebGPU wasm https://datoviz.org/examples/webgpu/live.html?id=showcases_protein

Slide 30

Slide 30 text

Use pixi when working with GPUs. fl https://matthewfeickert-talks.github.io/reproducible-cuda-work ows-with-pixi-scipy-2026/

Slide 31

Slide 31 text

Level 3 · レベル3 Systems and AI システムとAI Use systems engineering to level up with AI tools. 3 / 3 levels

Slide 32

Slide 32 text

? What's my view on AI? AIに対する私の見解はどのようなものでしょうか? It's complicated. それは複雑です。

Slide 33

Slide 33 text

Complexity is caused by two things: dependencies and obscurity. John Ousterhout

Slide 34

Slide 34 text

The physical world is complex.

Slide 35

Slide 35 text

The physical world is complex.

Slide 36

Slide 36 text

The physical world is complex.

Slide 37

Slide 37 text

The physical world is complex.

Slide 38

Slide 38 text

The physical world is complex.

Slide 39

Slide 39 text

The physical world is complex.

Slide 40

Slide 40 text

The physical world is complex.

Slide 41

Slide 41 text

No content

Slide 42

Slide 42 text

PyCon US Lightning Talk Manabu Terada Pragmatic system using AI to organize and find photos. https://github.com/pyconjp/pyconjp-image-search

Slide 43

Slide 43 text

Marimo pair An agent working side by side with a marimo notebook. h ps://marimo.io/pair h ps://marimo.io/blog/marimo-pair Agentic Data Science with Marimo Pair by Eric Ma h ps://youtube.com/watch?v=DvSxrCcfWFs tt tt tt tt h ps://ericmjl.github.io/blog/

Slide 44

Slide 44 text

Multimodal AI, Automation, and Robots https://www.airoa.org/ https://github.com/airoa-org/airopi-share https://arxiv.org/abs/2509.25032 https://arxiv.org/pdf/2505.12583 https://arxiv.org/pdf/1808.10703 https://airec-waseda.jp/en/about_en/ https://www.sugano.mech.waseda.ac.jp/en/ https://www.bbc.com/news/articles/c4g9e34yzvgo https://github.com/petercorke/robotics-toolbox-python

Slide 45

Slide 45 text

★ Community コミュニティ The players behind it all. Pythonistas grow the language, the global community, and world impact.

Slide 46

Slide 46 text

Local, Regional, Global PyCon JP Association https://www.pycon.jp/ Python Asia Organization https://pythonasia.org/ Python Software Foundation https://python.org

Slide 47

Slide 47 text

No content

Slide 48

Slide 48 text

Level up your Python ski s. ll conference.pyladies.com/about

Slide 49

Slide 49 text

ll Ca to Action

Slide 50

Slide 50 text

2023 NEC C&C Prize Guido van Rossum Thank you Guido. I fulfi ll tt h ps://www.candc.or.jp/en/2023/ceremony.html

Slide 51

Slide 51 text

Introducing Satyrn AI https://satyrn.ai/

Slide 52

Slide 52 text

https://satyrn.ai/

Slide 53

Slide 53 text

Introducing Satyrn AI Jupyterのようでありながら、エージェント向けのもの Satyrnは「エージェントのためのJupyter」です。ローカルAIの活用を成功に導くサブエージェントやツールを構築す るためのプロジェクトであり、私たち自身の手で、私たちのために開発されています。 https://satyrn.ai/

Slide 54

Slide 54 text

Introducing Satyrn AI オープンソース 地域所有の ローカルファースト Koudai Aono (PEP 750 co-author) is doing work on the Satyrn Engine. https://satyrn.ai/blog/2026/zen-and-compass/

Slide 55

Slide 55 text

We're taking this journey and exploring together. We hope you wi join us and support our mission to make AI systems more accessible, trusted, and humane for everyone. AIシステムを、すべての人にとってより利用しやすく、信頼でき、人間味のあるものにする ll ll Carol Wi ing

Slide 56

Slide 56 text

Choose your game. Create your game. Share the game with friends.

Slide 57

Slide 57 text

No content

Slide 58

Slide 58 text

Ready, player? Press start. 準備はいいか、プレイヤー? スタートボタンを押せ。 ll Slides: speakerdeck.com/wi ingc

Slide 59

Slide 59 text

Links Willing Consulting willingconsulting.com GitHub @willingc Mastodon @willingc LinkedIn in/carolwilling PyCon JP · 2026