Hacker News: ssalka

New comment by ssalka in "Why are rivers so mathematical?"

ssalka — Mon, 31 Aug 2026 18:22:26 +0000

Rivers form fractal shapes just like how Iterated Function Systems[1] and L-systems[2] form: by simple branching, carried out at varying levels of scale. A gentle reminder that fractals are nature's favorite building block.

[1] https://en.wikipedia.org/wiki/Iterated_function_system

[2] https://en.wikipedia.org/wiki/L-system

New comment by ssalka in "Muse Code and Muse Spark 1.2"

ssalka — Thu, 06 Aug 2026 15:39:48 +0000

I took a look at the black hole interactive demo/widget and dragged the Mass slider up. It crashed the widget.

New comment by ssalka in "Anti-fraud tools can't keep pace with robocall scammers"

ssalka — Fri, 31 Jul 2026 17:24:15 +0000

Cue the crypto bros touting their decentralized spam-detection blockchain

New comment by ssalka in "Claude Opus 5"

ssalka — Fri, 24 Jul 2026 17:24:57 +0000

I think some variance is to be expected since LLMs are typically non-deterministic, however that's a huge difference that I think warrants further explanation.

New comment by ssalka in "Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge"

ssalka — Tue, 21 Jul 2026 17:01:26 +0000

I think the Qwen team is probably aware that the NSFW community is very quick to adopt any new image gen model (see: Civitai). So, it seems like a good SEO approach to try and surface their model in search results that said community is likely already checking.

New comment by ssalka in "Show HN: GolemUI – Declarative Form Engine"

ssalka — Wed, 01 Jul 2026 17:03:33 +0000

There are lots of skills available that enhance Claude's design capability, eg:

https://www.tasteskill.dev/

https://www.usehallmark.com/

https://layers.jamiemill.com/

https://impeccable.style/

New comment by ssalka in "Only 2 of 128 YC-backed dev tools companies block unchecked merges"

ssalka — Tue, 26 May 2026 15:34:15 +0000

This seems crazy to me. At bare minimum, checking that simple install, lint/format, test, and build commands work without error should always be required in any code repository. That is, if you care to maintain the repo whatsoever.

New comment by ssalka in "New Shai-Hulud malware wave compromises 600 NPM packages"

ssalka — Mon, 25 May 2026 15:11:05 +0000

Affected packages & versions here:

https://airtable.com/appQsN6RIDefm93kd/shrMwE9m9GqU7pjU5/tbl...

Edit: this is not a follow-on attack from the one last week, OP's linked article was published on May 19.

New comment by ssalka in "Launch HN: Superset (YC P26) – IDE for the agents era"

ssalka — Fri, 22 May 2026 16:01:10 +0000

How do you guys plan to sustain the business, given that your product here is open source & already has many competitors doing similar things?

New comment by ssalka in "An OpenAI model has disproved a central conjecture in discrete geometry"

ssalka — Thu, 21 May 2026 02:18:07 +0000

You can make literally any position sound awful by saying that orphans will be killed as a result. Let's try to think posts through.

New comment by ssalka in "Meta blocks human rights accounts from reaching audiences in Saudi Arabia, UAE"

ssalka — Wed, 20 May 2026 17:03:04 +0000

[flagged]

New comment by ssalka in "Codex is now in the ChatGPT mobile app"

ssalka — Fri, 15 May 2026 15:31:58 +0000

Sometimes I get random inspiration for an idea while out on a walk or otherwise away from the computer. It's really nice to be able to throw a couple instructions out there, let your agent run with it, and see what it came up with later. Sometimes I do this 3-5 times before returning to my computer. IMO it's really nice to be able to start from X% done rather than 0% when I finally do sit down to review/iterate on the code.

New comment by ssalka in "Xs of Y – roguelike that names itself every run. Written in 4kLoC"

ssalka — Thu, 14 May 2026 21:07:48 +0000

honestly any indicator would help <3 love the game! I keep coming back to it haha

New comment by ssalka in "Xs of Y – roguelike that names itself every run. Written in 4kLoC"

ssalka — Thu, 14 May 2026 03:09:45 +0000

Not sure how you feel about including sound, but I feel like there needs to be a warning sound that plays when you're low HP, like in Pokemon. I keep dying because I don't realize I'm at low health

New comment by ssalka in "US Government releases first batch of UAP documents and videos"

ssalka — Fri, 08 May 2026 18:51:15 +0000

I feel like this is a symptom of AI psychosis

New comment by ssalka in "Ask HN: We just had an actual UUID v4 collision..."

ssalka — Fri, 08 May 2026 18:38:36 +0000

At one of my previous jobs, there was a function `createEntityWithRandomUUID` which would basically do the same thing as a light wrapper around database inserts. If a conflict occurred, it would generate a new ID and try again, up to 5 times I think. No logging to indicate whether any conflict actually ever happened.

New comment by ssalka in "Show HN: Ableton Live MCP"

ssalka — Sun, 03 May 2026 21:03:01 +0000

Things I would use AI for in music production:

1. Generating track layouts (add tracks + empty audio/midi clips throughout)

2. Generating MIDI sequences

3. Generating Serum patches

4. Extracting stems from existing audio

5. Automating common workflows (eg sidechaining)

6. Semantic search of sample library

That being said, I don't think I want a full agentic workflow for vibe-producing. Point solutions seems like a better fit for me, personally.

New comment by ssalka in "Mozilla Thunderbolt"

ssalka — Thu, 16 Apr 2026 16:07:01 +0000

I immediately thought "oh, the email client? It's AI now?" Then I realized this is Thunderbolt, not Thunderbird. Kind of an odd choice by Mozilla to have two products with such similar names.

New comment by ssalka in "Nucleus Nouns"

ssalka — Tue, 14 Apr 2026 19:45:27 +0000

This makes me think of domain models in domain-driven design. Very useful to think about what these models are and how it makes sense to set them up & relate them in your area of work.

New comment by ssalka in "LM Studio 0.4"

ssalka — Wed, 28 Jan 2026 20:05:57 +0000

Personally, I would not run LM Studio anywhere outside of my local network as it still doesn't support adding an SSL cert. I guess you can just layer a proxy server on top of it, but if it's meant to be easy to set up, it seems like a quick win that I don't see any reason not to build support for.

https://github.com/lmstudio-ai/lmstudio-bug-tracker/issues/1...