Elektrine
Log in Register
Paige Chat Timeline Gallery Friends Email Drive DNS Private DNS Domains VPN Kairo Nerve
Remote

Armin Ronacher

@mitsuhiko@hachyderm.io
mastodon 4.7.3
  • Open on hachyderm.io

Creator of Flask; Building at https://earendil.com — writing and talking about system arch, API design and Rust. Love my work? You can sponsor me on GitHub https://github.com/sponsors/mitsuhiko

Account here is largely inactive, moved to Bluesky.

6424 Followers
205 Following
50 Posts
Joined November 12, 2022
Blog:
https://lucumr.pocoo.org
Twitter:
https://twitter.com/@mitsuhiko
Bluesky:
https://bsky.app/profile/mitsuhiko.at
GitHub:
https://github.com/mitsuhiko
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
The post by @mitchellh@hachyderm.io gave me the final push to finish my GitHub "obituary". I originally started writing it after the Zig to Codeberg move. It's just a retelling of how I remember OpenSource pre GitHub and what it gave us. https://lucumr.pocoo.org/2026/4/28/before-github/
Before GitHub
Armin Ronacher

Before GitHub

Open Source before GitHub was reputation-driven and full of friction

100
14
88
4
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 4mo ago

More musings after some people got upset about the word clanker. https://lucumr.pocoo.org/2026/5/26/clankers/

lucumr.pocoo.org
23
5
14
2
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 4mo ago

Has been a while since I wrote about agentic engineering, so this time around some learnings of maintaining Pi as a junior maintainer to @badlogic@mastodon.gamedev.place https://lucumr.pocoo.org/2026/5/24/pi-oss/

lucumr.pocoo.org
21
2
6
1
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago

“If someone 50 years ago planted a row of oaks or a chestnut tree on your plot of land, you have something that no amount of money or effort can replicate. The only way is to wait.” https://lucumr.pocoo.org/2026/3/20/some-things-just-take-time/

Some Things Just Take Time
Armin Ronacher

Some Things Just Take Time

On friction, patience, and planting trees.

27
4
9
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

Motivated by how hard “equity” is to translate into German, I wrote about that missing word. https://lucumr.pocoo.org/2026/4/23/equity-for-europeans/

lucumr.pocoo.org
19
3
8
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

My kids and I built a little game with pi and codex over the weekend and it was a ton of fun. Some rough edges but with a bit of polishing I think getting agentic coding to be kids friendly is within reach. https://mitsuhiko.github.io/rope-man-game/

I know this website is not very receptive to coding agents still, but so much progress has been made in the last year. At this point they are such a pleasure to work with if you give them a bit of chance.

For working with kids we used vanilla pi, VoiceInk for text to speech and a small extension we built for drawing: https://github.com/mitsuhiko/pi-draw

mitsuhiko.github.io
18
4
1
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 4mo ago

Some weekend thoughts of recent mob behaviors. https://lucumr.pocoo.org/2026/6/6/communities-of-not/

lucumr.pocoo.org
10
2
7
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 2mo ago
I think this is a very bad move, and the people behind Codeberg should re-consider there stance. https://blog.codeberg.org/protecting-our-floss-commons-from-llms.html
blog.codeberg.org

Protecting our FLOSS commons from LLMs — Codeberg News

In Brief: Two motions regarding

4
1
2
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 2mo ago
Replying to
@bnjbvr@tutut.delire.party Because Github needs competition and now a potentially quite appealing alternative is gone for most new projects today. That wouldn't even be the issue, this rule will be arbitrary. It would be better for them to go the whole length and say: No LLM usage at all allowed. Then at least the line in the sand is clear.
4
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 4mo ago

Jarred rewrite Bun in Rust in a week with clankers. Say what you want about it, but that’s impressive. Also a reminder that this is happening while plenty of software engineers still dismiss the general utility of LLMs for programming. https://github.com/oven-sh/bun/pull/30412

GitHub

Rewrite Bun in Rust by Jarred-Sumner · Pull Request #30412 · oven-sh/bun

Blog post with details coming soon. It passes Bun's pre-existing test suite on all platforms (and fixes several memory leaks and flaky tests), the binary size shrinks by 3 MB - 8 MB, the benchm...

12
4
3
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

I analyzed my coding sessions and on the text interactions some words stand out. And well, they also show up on Google Trends as spiking. Oh and so much slop in my Twitter mentions and on GitHub. Thus here are some updated thoughts on all of this. https://lucumr.pocoo.org/2026/5/4/content-for-contents-sake/

lucumr.pocoo.org
12
0
6
1
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

Mario and I talked about Pi and agentic engineering with Gergely Orosz.** **https://www.youtube.com/watch?v=n5f51gtuGHE

12
0
1
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

I think @antirez ds4.c is important! I wrote down my thoughts on why I built pi-ds4 and why we need to focus our local model efforts stronger than we do currently. https://lucumr.pocoo.org/2026/5/8/local-models/

lucumr.pocoo.org
10
2
7
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

After the AI Engineer conference, which was full of useful criticism, I remembered that the most confident takes on AI often came from the least exposure. Rejection is easy, trial and error is expensive. I wrote about it. https://lucumr.pocoo.org/2026/4/11/the-center-has-a-bias/

The Center Has a Bias
Armin Ronacher

The Center Has a Bias

Why a measured position on AI tends to lean towards actually trying it.

10
26
7
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago

Trololol. I had the agent start using ua-parser-js, then it read the AGPL license after adding it and backed out immediately. This is hilarious.

7
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago

This evening ends with some darkness. "Just for fun". https://dark.ronacher.eu/2026/3/14/the-others/

dark.ronacher.eu
9
0
1
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@olasd @mitchellh @swheritage this is lovely!
4
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago

Since some asked: five months of Absurd in production. Durable execution on plain Postgres with tiny SDKs. The design held up fine, partitioning is still hard and some closing thoughts on the future of Open Source. https://lucumr.pocoo.org/2026/4/4/absurd-in-production/

lucumr.pocoo.org
5
4
2
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 2mo ago
Are there people still doubting the utility of these models?
1
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@audrius Don't be an asshole
3
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago

@schwa@mastodon.social @btel@mastodon.social the FSF is very much aware that if the case goes against them, then it weakens the GPL significantly. So they have to really be certain before they go in front of a court to know it’s going to rule in their favor. That’s why I think we won’t see many cases that want to actually look to careful. The implications could be vast.

4
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@pfmoore @ancoghlan there is a big difference. AI moves power from labor to capital. Outsourcing just equalizes between countries. Entire cultures are being lifted out if poverty in part because of remittances and outsourcing.
2
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@pfmoore Your discomfort is a position on acceptability based on your ethics. What it doesn’t give you is a view on how the technology behaves in reality. Those are separable. The post is about that second layer. I also mostly wanted to focus on the fact that the people that explore the space look like adopters, even if they are quite critical.
2
7
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@tintvrtkovic isn’t that what the Chinese models are? ;)
2
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@lfloeer@bonn.social I encourage you to read the post. I tried to explain there why we need to focus.
1
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@pythonbynight@hachyderm.io I hate to break it to you, but most of the things I write about or do are not novel :)
1
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@graves501 @mitchellh I sometimes go! I'm not the most meetup seeking person though :)
1
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@ancoghlan @pfmoore I think the issue with equating this to outsourcing is that one of the main arguments against outsourcing is at the same time the strongest argument for outsourcing: living standards. One side loses, the other wins.
1
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@realn2s @glyph You’re mixing two different claims. “This is unethical” doesn’t require use but you’re also making claims about behavior, impact, etc. Those aren’t moral questions and, as I argued, without direct contact, you’re outsourcing that.
1
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@glyph That to me is a confusing claim because I constantly write and talk about the cognitive challenges. Heck, my talk at AI Engineer even was about the topic.
1
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@aburka It's just a question of perspective. This might weaken copyrights which is something I generally advocate for :)
1
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@btel it is academic because of the argument I made in the post: people probably will be somewhat careful to fight that part out on court.
1
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago

@cuddle_puddle@mastodon.catgirl.cloud @ross@social.rossabaker.com Just for the record I remove the sharealike part from the blog license.

0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@realn2s @glyph you can absolutely reject something without using it and I said as much in the post. The point isn’t that rejection requires use but that without use, your view of how it behaves in practice is second-hand.
0
3
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago
Replying to
@jihyeok thanks. Need to figure out why this broke when I’m on my computer.
0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 4mo ago
Replying to
@goedelchen@mastodontech.de that was ages ago in AI years :)
0
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@btel to me that is largely a separate academic discussion. That applies to all LLM output and little to using it for reimplementing anything in particular. This made a from scratch implementation that had nothing to do with the original.
0
2
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@davidism OSS definitely will have to change. I see it daily in Pi. I have thoughts on that :)
0
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@lfloeer@bonn.social there are loads of projects that try to do a variation of it. It doesn’t pick a model and it doesn’t do the agent as well. Need the whole package!
0
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@aburka You can look at the code yourself and see that it's not anything like the existing implementation.
0
3
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago
Replying to
@djh That too is happening, there however I think the motivation has more to do with the economics around shipping and CO2 emissions.
0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@realn2s @glyph I'm curious why you think that way, because the article makes no such claim.
0
5
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@mdione yes, but that has little relevance to their ability to rewrite either GPL or proprietary software.
0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@aburka he (or rather his clanker) wrote all of the new code however.
0
5
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@abrahms that’s why it think we might see more trademark enforcement in the future
0
1
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@jmcs counterfeits are a question of trademark laws for the most part.
0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago
Replying to
@lfloeer how fast this is mostly depends on your Postgres deployment and what drives you use.
0
0
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 6mo ago
Replying to
@scottmichaud The difference here really is not the license but if there are CLAs in place. I wrote about this extensively in the past. Would point you to this article that I wrote about AGPL vs FSL: https://lucumr.pocoo.org/2024/9/23/fsl-agpl-open-source-businesses/
FSL: A Better Business/Open Source Balance Than AGPL
Armin Ronacher

FSL: A Better Business/Open Source Balance Than AGPL

My thoughts on license choices for balancing business interests and

0
4
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 7mo ago
Replying to
@aburka The person that maintained this library for 12 years is the person that wrote this code. If you trusted that person for 12 years to manage your dependency, why would it have changed all the sudden?
0
7
0
0
Open post
Armin Ronacher @mitsuhiko@hachyderm.io
· 5mo ago
Replying to
@lfloeer@bonn.social I don’t care about single binary :) I need efforts to focus so that one model in one harness in one inference engine works as good as possible. One can go wider from there.
0
4
0
0
Back
313k7r1n3
Elektrine

Tor hidden service

elekhj7afj4qnrr4yd3bkzslsyo5jgfxw3orgjkhlcxifueodybyiiad.onion

I2P eepsite

j6b6cyk6gjmepjih7jjadxgxvvf3lzzujljuu2v4biemzpg3naya.b32.i2p

Platform

  • Email
  • Chat
  • Timeline
  • VPN
  • DNS

Company

  • About
  • Contact
  • FAQ
  • Lite (no JS)

Legal

  • Terms of Service
  • Privacy Policy
  • Transparency Report
  • Report Abuse
  • Warrant Canary
  • VPN Policy

Support

  • support@elektrine.com
  • Report Security Issue
Mail client setup IMAP mail.elektrine.com:993 POP3 mail.elektrine.com:995 SMTP mail.elektrine.com:465
© 2026 Elektrine. All rights reserved. Server: 03:38:01 UTC