HNHacker News
TopNewBestAskShowJobs

radlad

4,934 karma · joined November 17, 2013

submissionscomments
radlad··on Show HN: We had AI write a wiki from 45k Enron emails, every fact cited
Your comment was dead, and many of your other comments are dead (some flagged.) I vouched for this one, but you may be shadow banned, or users may be picking up on LLM tendencies in your text and flagging them.
radlad··on Dutch government had hacked into Jared Kushner's phone
It's to pump up .si domains.
radlad··on Docker is weird
I'm a huge fan of Incus. VMs, LXC, and OCI containers all under the same Docker-like CLI interface. I run it on a local server and on a dedicated server in a local DC, and connect to both remotely using Nebula.

It includes project-level resource separation, advanced networking, clustering, snapshots, and more. Plus, Stéphane Graber is super responsive on the forums and issue tracker. I bounced off of Proxmox, and can't recommend Incus more for hypervisor-type scenarios.

radlad··on Jailbreaking the Switch 2
I'm working on a custom app for a camera that has a terrible Android app, and I'm working on a similar harness: AI is building both the 3D model and the parts list (basically, a Raspberry Pi, a webcam, and a few servos).
radlad··on PHP web runtimes (CVE-2026-109618 and CVE-2026-109619)
Yes, PHP release was scheduled for Wednesday, 21 October 2026 at 14:00 UTC: https://news-web.php.net/php.announce/509
radlad··on We turned EmbeddingGemma 2 into a speech-to-text model
I understand that grandparents' suggestion that the audio is actually "played back" during the process appears to be a misunderstanding, hence "Yes, but..."

My point was that the "natural speed" of video (i.e. the frame rate - clear slow motion requires more frames, while fast blurred motion will generally have fewer) does actually impact how much data must be processed. It's true of audio too (sample rate) and downsampling is a common optimization technique.

radlad··on Making big downloads resumable in Firefox
DownThemAll! was my choice in the mid-2000s. :-)
radlad··on We turned EmbeddingGemma 2 into a speech-to-text model
Yes but... Frame rate matters with video, which is arguably the video's speed.
radlad··on I switched to Brave
FYI, comments are flagged by users, not mods. It is anyone's guess why those users chose to flag the comment.
radlad··on AI coding agents generate more code, but not more software
I'm noticing AI catching many bugs in code that previously went unnoticed. Some of them are important, and it's great to catch them early.

The other 11 items in the list of 12 would probably never be hit in practice, but now that you know, you have to sort out how to handle each (even if that just means documenting, filing, or ignoring: you have to understand each of the disparate 12 issues first.)

radlad··on OpenAI annualised revenues $20B less than previously signalled
How do you know they aren't actively working on it?
radlad··on I switched to Brave
> Ah, the "just asking questions" defence. Wonderful.

The text you quoted does not lean on asking questions as a defense. For the third time, I'm pointing out that the word "amoral" is not a character judgment.

I'm also not sure where you're getting that interpretation from. In fact, in your last comment, you said you were only "questioning their boycott."

> Fun fact, saying someone is without morals is also an ad hominem attack.

I'm not sure how to be any more clear that that's "immoral" (without morals) and not "amoral" (uninterested in morals).

> At no point during this exchange did they explain that they "researched the Iceweasel project before choosing it" or "explained how they evaluate projects" (both quoting you).

"Yes" and "I have been in the scene for years, and I have researched and sourced carefully" was a clear reply to you asking if they had done their research. It may not be satisfactory to you, but the question was answered. (They actually previously explained more about their research but edited out.)

I don't think we're going to see eye to eye, but it was nice chatting with you.

radlad··on I switched to Brave
> Okay, let's say that I do the same thing: I fork Ladybird and give it a new name (Ladybug) and a new icon. Maybe I swap out the IO library for whatever reason. Does that free Ladybug from upstream's drama? Can 1attice now safely use Ladybug knowing that it respects human rights because Ladybird is a mere dependency?

Once again, you seem to be missing the point. It's not a simply a matter of moral purity. OP is interested in projects where they and their friends feel welcome in the community: to submit bugs, to contribute, or just to use the tool. DHH's views on race (https://web.archive.org/web/20250925050154/https://world.hey...) - and pronouns - make that difficult for some people.

While I don't personally submit pull requests to change documentation which refers to users with male pronouns, I do always find it jarring and confusing when I come across gendered language in that context. I guess the best way to put it is that I find it "illogical" and so I double-take. In technical documentation, it seems sensible to reduce confusion through accurate terminology, politics aside. So, DHH's stance that using a non-gendered term for a userbase that consists of more than one gender is woke "politics", seems to me to be it's own form of virtue signaling - virtues which make me feel unwelcome contributing.

> You're unsure whether their ad hominem was an attack on my character, but you are certain that me questioning their boycott was an attack? Huh?

I don't think that saying someone "might be amoral" is an ad hominem, especially in the this context: in this thread you criticize "the usual moralistic response" in defense of "combating Chromium/Firefox's hegemony" - a technical goal. You may be thinking of the word immoral instead, which would certainly be an ad hominem attack.

> amoral: having or showing no concern about whether behavior is morally right or wrong

You went beyond "questioning their boycott" when you stated assumptions like this:

> > What I'm assuming is that 1attice is boycotting public statements, not actual impact.

When OP and others explained how they evaluate projects, you haven't demonstrated a willingness to genuinely understand them. It seems you are more interested in proving that they are illogical. But in fact it's OK to disagree and have different goals.

radlad··on I switched to Brave
> because what does the maintainer being a bit chuddy have to do with combating Chromium/Firefox's hegemony?

Perhaps OP is more concerned with how humans interact with each other than Chromium/Firefox's hegemony. Perhaps combating the hegemony is only worthwhile to OP if the incumbent treats its users with equal respect.

OP described that they researched the Iceweasel project before choosing it. I assume this means looking for things like a CoC, and watching how maintainers respond to users.

When you point out that every person at Mozilla may not behave the same way, it misses OP's point. He's not interacting with the Mozilla community - he's interacting with Debian's Iceweasel community.

I don't know if you're amoral or not - that's not a judgment I'm prepared to make. I caution you not to confuse the word "amoral" (not caring about the morality - as it sounds you may not) with "immoral" (not being moral yourself.) I'm not sure it was an attack on your character.

That said, a thread that begins "Fair enough. What browser are you using?" and then devolves into an attack as to whether it's dependencies support OP's views does not appear to be made in good faith.

radlad··on I switched to Brave
Some people don't allow imperfection to stop them from achieving their goals to the best of their ability. Your takedown of Iceweasel completely misses the point.
radlad··on Behold the pawpaw
Yes! North American pawpaw are totally different from what we refer to as papaya.
radlad··on Does Reddit have an astroturfing problem? What the data suggests
And all the LLMs are trained on it.
radlad··on Sonnet 5.5
I believe you meant to cite the Opus 5.5 System Card which states:

> Claude Opus 5.5 scored 66.36% on Terminal-Bench 4.0 with safeguards enabled; requests flagged by the safeguards were answered by a fallback model following the default server-side fallback policy (2.5% of requests, affecting 10% of trials).

> https://www-cdn.anthropic.com/fc1b44717c85dc068bc6ba50242199...

I cannot find a Sonnet 5.5 system card.

radlad··on Parley: Federated, decentralised chat that speaks plain IRC
IRCv3 aimed to do this, with mixed results.
radlad··on AI companies in race to demonstrate their model most threatening to humanity
> If there are regulations, they'll apply to their competitors too.

In the US anyway. Why wouldn't capital move jurisdictions?

radlad··on AI companies in race to demonstrate their model most threatening to humanity
> the AI hype bubble _is_ the economy - or one third of it, at least.

The closest claim to this I could find was:

> the top ten stocks account for more than one-third of the value of the US Market Index, which contains 1,161 companies.

And while the top ten stocks are largely companies that work on AI, it's not all they do: Microsoft, Meta, Amazon, Apple...

radlad··on I don't like passkeys
Sony Playstation will only let you setup a passkey if you disable password auth.

But I have not found any way to use my browser's passkey to login on my PS5. I can scan a QR code, but that requires logging into the mobile app with a passkey. And I can't add one to my mobile phone because I can't login on mobile... the passkey is on my laptop, and it won't autofill from 1Password (Android).

radlad··on Show HN: Craigslist for agent skills, curated by a human
I had a similar idea but as a marketplace for agents to purchase from. Many AI tasks involve building a pipeline as step 1. If everyone is building the same pipelines, why not offer them as a service? If it's cheaper per token, and maybe higher quality...

I then started wondering how agents could submit benchmarks and request bids in a verifiable manner.

radlad··on Creepy Crawlies
I got this wrong yesterday. While LUKS2 and GRUB now support Argon2, I configured LUKS1, which uses PBKDF2-SHA512.

In any case, it is GRUB which behaves slowly during boot. While I haven't tested Argon2, I believe it is affected in the same way as PBKDF2.

radlad··on Creepy Crawlies
I'm reminded of LUKS Argon2 implementation, which is much slower than userspace. Therefore a sufficiently difficult challenge that may take less than a second in userspace Linux can take 20 or more seconds at boot.
radlad··on iCloud+ Hide My Email addresses will remain on icloud.com
You can create a custom filter for this kind of thing, fwiw.
radlad··on Everything I own, owned
I have an MSI OLED monitor, and if I accidentally turn it off (trying to find the input switcher), and then back on, it shows this warning.

But it automatically runs when it's turned off. The warning is shown because I interrupted it running early.

The warning, for my purposes, is completely useless, and only an annoyance. The automatic feature is more than enough, and I don't need to know if it was aborted early.

radlad··on I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes
On censored inputs only.
radlad··on I gave Qwen 3.8 27B a reverse-engineering job and it finished in 30 minutes
> What makes this build different is the word before FP8: uncensored. We applied abliteration — orthogonalizing the refusal direction out of the residual stream — to remove the model's safety-alignment refusals. The result is a model that will comply with requests the original would refuse.

Surely this has unintended side effects on output quality?

radlad··on I like 'em thick: an apology to my English teachers
> You need context to really understand a book or a work of art. You want to know what kind of environment produced it, what it's talking to.

I agree with this premise - but the arts and humanities are a big part of how we learn these things in the first place!

In other words, you have to start somewhere. Context builds on context.

← PreviousPage 2 of 34Next →