Oh, there is also a Vampire Weekend mix from CRO. Another one, I didn't expect.
03.02.2026 10:22
👍 0
🔁 0
💬 0
📌 0
Oh, there is also a Vampire Weekend mix from CRO. Another one, I didn't expect.
CRO remixing a @blocparty.com song is something I did not expect to happen at all. As someone who enjoys CRO and is an ultimate @blocparty.com fan, this is an unexpected mix.
Unexpected music amalgamation through cloud engineers this morning:
1. A colleague made a joke that some other is our Chief Rendering Officer (CRO)
2. This made me listen to CRO all morning
3. I discovered open.spotify.com/intl-de/trac...
Still, as a core maintainer of a very large package distribution ( @conda-forge.org ) , I also see that updating the patches that we have on top of upstream releases gets a bit simpler with AI. Most of the AI benefits have been more personal automation tooling there.
Instead, if you get your change upstreamed, you get the benefit of
1. Maintenance/update to new dependencies
2. Being part of the unit test, i.e. that if changes in other parts of the system would impact your change, it will get adjusted.
This still leaves us with the issue of long-term maintenance. Internal forks of an open-source project have been quite costly, as you will need to continuously rebase the patches on top of the public version to get the benefit of new releases.
😅 Hoped for some good overview, but this would probably also be my approach.
Interesting! How did you determine that you want this exact machine? I'm also pondering getting a longaarch64 and a RISC-V one from China (once released) for some other OSS projects, but I find it hard to actually decide what would be a reasonable purchase.
Question scoped to software in physics: Do you currently use PPC64LE architecture hardware for anything in your scientific workflow and is it important to you to have binary builds of software for that machine? Are there high impact examples that you can think of? Or is everything x64 or aarch64?
As part of this, I also made my first (very tiny) contribution to CPython 🚀
Two weeks ago, I posted that Python 3.14 was the fastest available/usable Python in @conda-forge.org.To keep new releases as quickly usable as possible, I'm reviewing this week what we did in this release and what we can already do to make Python 3.15 even faster: uwekorn.com/2025/12/15/p...
It is only a minor contribution, but I'm honoured to have landed my first contribution to CPython: github.com/python/cpyth.... The benefit of testing to build CPython 3.15.0 very early on.
I was intruiged by this post and made similar (but not comparable) plots for the Python 3.14 availability on conda-forge: uwekorn.com/2025/12/01/p...
Conda in the Packaging Spectrum: From pip to Docker to Nix
If conda is a distribution, where does it fit alongside pip, Docker, and Nix? Part 2 of the 3-part series! 🚀
#conda #packaging #python
conda.org/blog/conda-p...
Week 41 this year was the week of the Python 3.14 release and also the last one I wrote a report on my @conda-forge.org activities: uwekorn.com/2025/10/20/w...
While I was at @pydataparis.bsky.social , I did not spend much time on @conda-forge.org, but in that week, at least some important milestones happened for the Python 3.14 migration, e.g. unblocking everything that depends on pandas. More details can be found in my report: uwekorn.com/2025/10/16/w...
Another week of reporting on my @conda-forge.org work: uwekorn.com/2025/10/14/w... This time, I started a mini-sprint on the Python 3.14 migration.
Another week in my effort to give some insight into my @conda-forge.org work, my post frequency definitely lacks behind my merge frequency, but here is my report on week 2025-38: uwekorn.com/2025/10/12/w...
Python 3.14 has been released two days ago. At the same time, we also uploaded a build of it to @conda-forge.org and about 40% of Python version dependent (mostly those with compiled code) were immediately available. See conda-forge.org/blog/2025/10...
At PyData Paris yesterday, I gave a non-technical talk on how you, as an engineer, can deal more practically and proactively with compliance topics. Slides at speakerdeck.com/xhochy/navig...
Ever wondered where all my GitHub contributions come from? I have document a week's worth of @conda-forge.org contributions: uwekorn.com/2025/09/18/w...
I plan to do this for 2-3 more weeks to get an overview where I spent my time. I hope there is also content in there that helps others.
Today is my last day of five weeks of parental leave which means that it's going to be a hard day for my chargers. All the laptops and tablet will need to be fully charged and updated to be ready for tomorrow's work. 🔥
Nice to be back at Paris again. My PyData journey started there 9y ago with my first ever talk about Arrow. This time, I'll tell you how to stay sane when you have to deal with compliance topics as a developer.
The study on an EU-STF underscores what the Sovereign Tech Agency has championed since 2022: that lean, independent, and targeted funding can meaningfully strengthen the open source ecosystem and thereby boost innovation, competition and resilience.
🎉 Eagerly trying to get 3.14 ready on conda-forge at the same time: github.com/conda-forge/...
@conda-forge.org and @conda.org are essential components of our stack. Thus, it is natural for us to work with the community there to make it ready for upcoming cyber(security) regulations. Read about our efforts at tech.quantco.com/blog/conda-r...
Reach out, if you would like to collaborate!
Setting up mail records is not simple but www.learndmarc.com was nice to understand how things work and whether we have done it correctly.
A nice feature of GitHub Copilot Coding Agent: I can start simple changes from my smartphone via chat, e.g. write an internal enhancement proposal. While it would be higher quality if made manually, it enables some more things "on the go".
Finally, I caught up with what other CTOs seem to be doing the whole day 🤣
