Hacker News: onion2k

New comment by onion2k in "How Complex is my Code?"

onion2k — Sun, 12 Apr 2026 18:22:58 +0000

This isn't new. I've seen it for decades, including in situations where no one is at risk. I don't think it's often a fear thing.

New comment by onion2k in "Anthropic downgraded cache TTL on March 6th"

onion2k — Sun, 12 Apr 2026 09:50:58 +0000

I use Codex at home and Opus at work. They're both brilliant.

New comment by onion2k in "How Complex is my Code?"

onion2k — Sun, 12 Apr 2026 09:46:15 +0000

Well, maybe. It's projection, because I certainly don't make simple processes myself a lot of the time, but I do try to optimize them afterwards. I have a few decades of seeing people implement processes than I've had to use, and then had to simplify as I moved into more senior roles. I've had people push back quite forcefully when I've pointed out they do things like writing reports that no one reads or gathering data that teams ignore. People often fight for added complexity because their perception is that it's important, and that means they must be important because they're the one in control of it.

There is an element of projection because there is in most things people talk about; I'm speaking about this through my filters and biases after all. But it's grounded in a fair chunk of experience.

New comment by onion2k in "How Complex is my Code?"

onion2k — Sun, 12 Apr 2026 08:30:36 +0000

I'm an Engineering Manager, and I think I have a similar role just applied to people processes rather than code. One nuance though - a lot of the time I suspect it's deliberate complexity designed to obfuscate how little people actually do.

New comment by onion2k in "Show HN: Editing 2000 photos made me build a macOS bulk photo editor"

onion2k — Sat, 11 Apr 2026 21:04:50 +0000

I built my own photo viewer for OSX entirely because Finder doesn't have an 'actual size' option. OSX is pretty terrible for image management.

New comment by onion2k in "Starfling: A one-tap endless orbital slingshot game in a single HTML file"

onion2k — Sat, 11 Apr 2026 20:48:24 +0000

It started out using browser.click events and then switched to using browser.evaluate script injection. That's entirely valid for my use case.

New comment by onion2k in "Starfling: A one-tap endless orbital slingshot game in a single HTML file"

onion2k — Sat, 11 Apr 2026 18:28:17 +0000

It didn't exactly play it using the LLM, but it used Playwright to execute code in the browser to work out how it works and then wrote a script to inject into the page to play it. It was basically perfect AI getting skip * 2 on every shot even after a hundred planets. I didn't expect it to do quite so well with only 2 prompts.

New comment by onion2k in "Starfling: A one-tap endless orbital slingshot game in a single HTML file"

onion2k — Sat, 11 Apr 2026 16:19:55 +0000

I gave Codex 5.4 Playwright MCP access to the site and a prompt of "Use Playwright CLI Skill to open https://playstarfling.com/ and load the game. Work out how to play it, and devise a strategy to win." After a about half a dozen attempts it had figured the game out. Then I prompted it to "Score as much as you can." It wrote itself an auto-play script that just keeps going.

I stopped it running at 10866. That's currently the high score. I appreciate that this is pointless and proves nothing, but I've been experimenting with automating testing games (I work at a gaming company at the moment) so it felt like an opportunity to try an experiment.

New comment by onion2k in "JSON formatter Chrome plugin now closed and injecting adware"

onion2k — Sat, 11 Apr 2026 02:07:18 +0000

Surely it's reasonable to assume that a company doing some dubious 'marketing intelligence' scraping of people's data from a Chrome plugin is going to both inflate the numbers they put in offers and try to scam their way out of paying if you actually accept. I wouldn't consider them real offers. They're marketing. The real world payments, if you get them, would be lower.

New comment by onion2k in "Git commands I run before reading any code"

onion2k — Wed, 08 Apr 2026 14:38:45 +0000

That sort of project exists in an ocean of abandoned and dead projects though. For every app that's finished and getting one update every few years there are thousands of projects that are utterly broken and undeployable, or abandoned on Github in an unfinished state, or sitting on someone's HDD never be to touched again. Assuming a low change frequency is a proxy for 'dead' is almost always correct, to the extent that it's a reasonable proxy for dead.

I know people win the lottery every week, but I also believe that buying a lottery ticket is essentially the same as losing. It's the same principle.

New comment by onion2k in "Git commands I run before reading any code"

onion2k — Wed, 08 Apr 2026 13:40:45 +0000

Technically you're correct that change frequency doesn't necessarily mean dead, but the number of projects that are receiving very few updates because they're 'done' is a fraction of a fraction of a percent compared to the number that are just plain dead. I'm certain you can use change frequency as a proxy and never be wrong.

New comment by onion2k in "MegaTrain: Full Precision Training of 100B+ Parameter LLMs on a Single GPU"

onion2k — Wed, 08 Apr 2026 13:37:11 +0000

It’s too slow for the scale of pretraining.

There isn't really such a thing as 'too slow' as an objective fact though. It depends on how much patience and money for electricity you have. In AI image gen circles I see people complaining if a model takes more than 5s to generate an image, and other people on very limited hardware who happily wait half an hour per image. It's hard to make a judgement call about what 'too slow' means. It's quite subjective.

New comment by onion2k in "After 20 years I turned off Google Adsense for my websites (2025)"

onion2k — Tue, 07 Apr 2026 05:39:29 +0000

Creator-led channels are competing with AI-generated video channels that pump out many videos every day. The ad spend hasn't increased but now it's shared with people who have automated their channel's content production and who are likely getting the majority of what's available.

New comment by onion2k in "Employers use your personal data to figure out the lowest salary you'll accept"

onion2k — Mon, 06 Apr 2026 16:40:46 +0000

Everyone just says the top of the range.

New comment by onion2k in "Usenet Archives"

onion2k — Mon, 06 Apr 2026 08:41:14 +0000

Usenet was great in the late 90s and early 2000s. I posted a lot, and met some great people. I got a job doing tech review of books about WAP and WML from my posts in a group about the forerunner to mobile internet, and another job with a company making intranet software from some posts about ASP and vbscript. I've no idea where I'd go for that sort of forum today.

New comment by onion2k in "Employers use your personal data to figure out the lowest salary you'll accept"

onion2k — Mon, 06 Apr 2026 06:16:24 +0000

It works the same way here in the UK. Some companies ask your previous salary, and sometimes check your references, and sometimes your previous employer will disclose what your salary was. If it turns out you lied nothing bad happens, but you've just given your new employer a reason to dismiss you.

The main problem with the UK system is that it means that if you were underpaid before you're likely to continue to be underpaid in your next role (if you accept a low salary again). For that reason when I'm hiring I've stopped asking for someone's previous salary, and just ask them what they want instead. If it's in the right ballpark everyone's happy. If they lowball themselves I ask why and usually get "That's x% more than I'm on now.", which leads to a conversation about how they're underpaid and should be asking for more. If they ask for too much then I just don't hire them because I can't afford them.

There's a new law coming in where companies have to disclose salary bands now, which at least means people will understand the bottom end. That's going to make the salary negotiation part of hiring a lot easier.

New comment by onion2k in "The 1987 game “The Last Ninja” was 40 kilobytes"

onion2k — Mon, 06 Apr 2026 05:56:55 +0000

But for some reason, Firefox refuses to play back those kinds of files.

And that reason is because x264 is a free and open source implementation of the H.264 codec, and you still need to pay a license to use the patented technology regardless of how you do that. Using a free implementation of the code doesn't get you a free license for the codec.

New comment by onion2k in "Gemma 4 on iPhone"

onion2k — Mon, 06 Apr 2026 05:41:34 +0000

Software that doesn't work has been available for decades. It's not a good signal for vibe-coding.

New comment by onion2k in "12k AI-generated blog posts added in a single commit"

onion2k — Sun, 05 Apr 2026 09:42:06 +0000

What I meant was that it'd a good answer but maybe not perfect. For example, if you ask for a coffee recommendation you might get something that's in your top 5, but not number 1. That's better than getting a page of links where the top 3 have paid to be there, the next 5 are SEO-farms, and then maybe there's a site about coffee that will answer your question.

New comment by onion2k in "12k AI-generated blog posts added in a single commit"

onion2k — Sat, 04 Apr 2026 19:22:09 +0000

The fact is what people really want from a search engine is a single perfect result that answers their query exactly. An LLM does the 'single result' bit, but it's dubious whether or not it's a perfect answer. Most of the time that's probably not very important so long as the answer satisfies the search enough that the user is happy.

Google is trying to turn Search into that product e.g. the single answer to a given search. They could do that now with Gemini, but the ads in the results are what makes them money, and the backlash to embedding adverts into the output of Gemini would drive millions of people to OpenAI overnight. They have to do it slowly. Give it 5 years though, and search engine results pages will be a thing of the past.