Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
gsuuon
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
9 ms
·
61.
▲
by
gsuuon
2y ago
That sounds like a really interesting project, is there anything to look at online?
62.
▲
by
gsuuon
2y ago
This sounds like an interesting series, I'd watch it. If you do go for livestreaming, keep in mind a big part of that is interacting with chat. If that's not something you're into then maybe recorded content with light editin
63.
▲
by
gsuuon
2y ago
Congrats on the HVM2 launch! Been following for a while, excited to see where this project goes. For others who are lost on the interaction net stuff, there was a neat show hn that gave a more hands-on interactive intro: https://
64.
▲
by
gsuuon
2y ago
I'd actually really love the ability to follow just the 'Show HN's of a particular user, for those who consistently post interesting projects. It would at least lessen the FOMO causing me to check HN a little too often.
65.
▲
by
gsuuon
2y ago
This is awesome! The comment filtering feature is a nice touch. Do I need to sign up to try local inferencing?
66.
▲
by
gsuuon
2y ago
I've used AI to test documentation, that is I give the AI some documentation I wrote as context and ask it to do something. If the AI gets it wrong, it can be an indication the docs aren't clear. I've thought about this since
67.
▲
by
gsuuon
2y ago
AI could also be used as a tool to help you develop the language - you could try to teach your language to a LLM, or finetune one to help generate the language, or simply plug in a grammar to the ones that can restrict by grammar and have a
68.
▲
by
gsuuon
2y ago
Are these multimodals able to discern the input voice tone? Really curious if they're able to detect sarcasm or emotional content (or even something like mispronunciation?)
69.
▲
by
gsuuon
2y ago
Is this limited to the ChatGPT UI? Hopefully this preferential treatment doesn't make it into the API.
70.
▲
by
gsuuon
2y ago
This is super cool! Both the writeup and the app. It'd be great if the search results linked to the HN story so we can check out the comments.
71.
▲
by
gsuuon
2y ago
Nice writeup! Very curious about the performance per VRAM with this compared to just quantization. Any plans to implement a cross platform version?
72.
▲
by
gsuuon
3y ago
Yup - it's very easy for distracted, tired work to become negative work.
73.
▲
by
gsuuon
3y ago
And for release builds, running the tests across a cohort of gaming profiles so you could get detailed results for various personas instead of just a guess and check in production. It'd be great for indie and hobby projects as well - g
74.
▲
by
gsuuon
3y ago
This would be cool to see as a game dev CI tool, an end-to-end playthrough test which would validate not only UI but the writing and game flow. Imagine getting a report that says something like "time to chop first tree: +20%"
75.
▲
by
gsuuon
3y ago
I spent way too long looking for an xplat lf/ranger that I can just use with cd without having to write the selected directory to a temp file (which then gets complicated if you want something portable across shells) so I made a direct
76.
▲
by
gsuuon
3y ago
I'm really excited for LLMs in gaming, like an Ender's Game / Homeworld type game where you can shout orders and the units scramble around or Stellaris where you have actual discussions with the other factions. Local and reli
77.
▲
by
gsuuon
3y ago
Excited for this, Vite running on a single bundler means you can finally trust what you're seeing with the dev server.
78.
▲
by
gsuuon
3y ago
> I have no authority and will not stand behind these claims if challenged. I want this on a shirt
79.
▲
by
gsuuon
3y ago
I write a hierarchical structured task list (very similar to op) with a neovim plugin I wrote ( https://github.com/gsuuon/note.nvim ). Sometimes I'll write these items as commit messages, make the changes and check
80.
▲
by
gsuuon
3y ago
I guess since the control vectors are applied at every layer, it becomes impossible to override (e.g. if you're consuming user-generated prompts)
81.
▲
by
gsuuon
3y ago
It's because of there being too many cooks in the kitchen for front-end. It's something people can see, and it's easy for anyone to have an opinion.
82.
▲
by
gsuuon
3y ago
This happened with one of my neovim plugins (also named llm!). In my case I was already thinking of renaming the project so I just went through with that. The other project's maintainer was nice enough to apologize and I ended up with
83.
▲
by
gsuuon
3y ago
This is cool! I like the hit protection / scars mechanic. Were the illustrations done with AI?
84.
▲
by
gsuuon
3y ago
This would make a great host extension - just add new extensions to the list and it automatically pull/build/installs the extension.
85.
▲
by
gsuuon
3y ago
I also had a similar thought, decided to implement it as a neovim plugin[1] to see what it feels like. Turned out okay! I use it daily now, but I think the important part is it's just a "mode" in an otherwise normal editor so
86.
▲
by
gsuuon
3y ago
That'd help, but a problem is they could still go closed-source and you wouldn't know - the store itself has no concept of open or closed source so it's not like you could check an "uninstall if it goes closed source&quo
87.
▲
by
gsuuon
3y ago
Not (just) a VCS, but this is the idea behind the Unison language: https://www.unison-lang.org/docs/the-big-idea/
88.
▲
by
gsuuon
3y ago
Doesn't this just offload the work of marketing the main product to marketing the free tool?
89.
▲
by
gsuuon
3y ago
Would really like something like this to generate infinite lofi to code to. I'm struggling to get this to not output vocals though - I've tried both instrumental and no vocals in the prompt.
90.
▲
by
gsuuon
3y ago
They're good for at least exploring the vocabulary of a new domain you're not familiar with. You can really dig into a topic and ask systematic things while jotting down new terms to look up and verify. It can help get you past th
More ›