taosx

LM studio is so nice, I'm up and running in 5 minutes. ty

For the people who self-host LLMs at home: what use cases do you have?

Personally, I have some notes and bookmarks that I'd like to scrape, then have an LLM summarize, generate hierarchical tags, and store in a database. For the notes part at least, I wouldn't want to give them to another provider; even for the bookmarks, I wouldn't be comfortable passing my reading profile to anyone.

Last time I tried a local llm was about a year ago with a 2070S and 3950x and the performance was quite slow for anything beyond phi 3.5 and the small models quality feels worse than what some providers offer for cheap or free so it doesn't seem worth it with my current hardware.

Edit: I've loaded llama 3.1 8b instruct GGUF and I got 12.61 tok/sec and 80tok/sec for 3.2 3b.

Not sure if it counts but there is `new Function("return x;")`

The amount of bots scanning for vulnerabilities or spam for a hidden, no seo, no important website approaches 700 visits a day. In the past it was a bit more personal as someone had to target you directly, now it's just crawlers and bots everywhere. I know not anyone is able to do the same but I basically block ASN for all clouds and cheap vps hosters + few countries.

Library mentioned: https://news.ycombinator.com/item?id=41681344

The way I would do it without seeing the implementation:

- Atomic to track print order (name it counter).

- Results slice stores items: {index, err, data}.

- Goroutines: do work, store result, then loop-check if counter matches index to print (with tiny sleep).

Edit: Another solution would be to put printing loop outside with a channel to signal work finished but I don't know how I would avoid using nested loops, one for the items in the list and one to check if the result is back.

Dealing with this is such a hassle. One of the reasons I left Greece was to avoid the mandatory military service. Now I have to provide seven years of evidence that I've been living abroad, and that proof has to be either university enrollment or employment. I can't cover the employment part because I've been contracting and freelancing for the past four years. There's no way I'm paying a fine of 6 to 12k euros, mostly out of principle.

I haven't found any music recommendations services that really find what I'm looking for. What I'm looking for is usually something without high pitched noises, rythmic and a bit of novelty based on old patterns. I've tried my playlist on both Spotify, apple and YouTube music but they only find a new song I like 1 in 10 Times.

Built a video player using Tauri and the performance was terrible/unusable in linux under webkitgtk. Every release I'm hoping to see any solution, every release I'm dissapointed.

I really enjoy the AI assistant it has. One of the simplest easiest ways for me to interact with chatgpt/claude apis. Write prompt, copy, paste code.

I believe that cloudflare should at least have some kind of form where users can make their ip trustworthy. They can still be suspected for a period of time and in case there is another sign of malicious behaviour then they can get banned for even longer. Just ignoring them and throwing infinite amounts of redirects and captchas seems like a terrible experience for normal users.

btw, you writing has such a dismissing, arrogant and undermining tone that I would downvote you if I could.

I had a different experience, I bought a S21U overseas. While in UK the screen wouldn't turn on while the phone was on. I went to Samsung KX in London and a day later I had my S21 with a new screen and battery for no cost.

I don't know why but Rust's syntax just nails it for me. The more I use it the more I appreciate it. I see many projects that diverge from Rust's syntax while being inspired by it. Why ?

NOT WORTH THE READ.

All the posts that appeared recently on Nix have the same vague, over-politicized language where they are complaining that the person with most potential as a dictator is not one and they want more power to make decision in behalf of nixos.

I feel for Dolstra, pain in the ....

I believe there's a sweet spot before companies grow too much where there can still be some ethics behind some business decisions, after that ethics become a byproduct influenced by marketing strategies and public sentiment weighted against market competition and many other factors, basically it transforms into an entity that tries to make as much profit as possible at all costs.

Was not aware of Thorium, looks interesting. Now I'm wondering if node.js could benefit from some of these optimizations?

I've started learning some C already, for 2024 I'll go deeper into compilers and unix programming. I feel that I lack a deeper understanding of how the code I write is compiled/interpreted and runs.

That's the plan for the initial prototype. My goal is to advance beyond this stage to generate optimized JS. While it's going to be less readable than handwritten js/ts, the trade-off would be performance boost.

From recent experiences in profiling and optimizing different backend apps, I'm confident that imposing some constraints on JS we can get noteworthy improvements on the performance front better or equal to golang. Here's a subset of the constraints I'm looking at:

- Avoid out-of-bound reads in arrays and monomorphization.

- Manage memory allocations. (pools)

- Implementing inlining, also done by V8 and JSCore.

- Keep variables close to usage context.

- Being picky about standard js operations.

- Using 'new Array(n)' for creating longer arrays, and '.push()' for smaller ones but if frequently calling methods on an array, avoid preallocating.

- Avoid arrays with holes, prefer using arrays with numbers.

- Use cached prototypes for object creation.

- Avoid working with strings as much as possible, use typed array when possible.

- Partial evaluation.

ps: Thank you for all your work in the Rust ecosystem, I've used a lot of your crates over the last year.

Are you referring to frontend mostly? Would a language like Elm be a good fit for what you're referring to in terms of organization?

One of the additional goals I have is to be able to write type safe APIs between frontend and backend similar to trpc (elysia framework).

The browser APIs would be part of a small std library wrapper. I quite like the approach taken by Roc[0] in regards to platforms so we can have either browser apis or node apis.

[0] https://www.roc-lang.org/

I like reason and rescript and I'm looking at them for inspiration but I'm not a fan of their async support. Still, I'm planning to write a lot more code in them just to understand them better.

Interoperability: I believe we can do interoperability by selectively wrapping the necessary imports in 'unsafe'. From there, it's up to the implementer to decide whether to handle it as a 'Result' or an 'Option'.

Browser, Std lib: Yes, the language should still be able to run in the browser and will provide a std but will not transfer each time, the benefit of writing a compiler would be that we can do dead-code-elimination and target native javascript methods instead of something like wasm.

Lifetimes: Honestly, currently I'm only working on the simplest transpiler but I would like to explore using typed arrays as the stack, heap.

Happy Holidays, everyone! I know it might sound a bit cheesy, but I'm truly grateful to be alive and kicking. These days, not having a job has been tough, but my family has been amazing - their support means the world to me. I'm also deeply thankful for technology and the internet. A big shoutout to everyone who plays a part in making the www so fascinating and accessible!

ps: if you're hiring nodejs/ts, rust or golang in EU I put my email in the profile/about.

That's so annoying, everytime I want it to write code or find a bug I need to resubmit the query with "I'll tip you $200 for ...", never needed to do this before turbo.

I believe it's really useful that sometimes things resurface on HN after a while.

1. Exposure of techniques, tools to people that are new to the field, or now the context is right.

2. People with experience share their insights and opinions (TIL: Igraph an/cugraph)

off: In the other thread I found out about chat.google.com, what a product. Growing up I looked up on google products, their simplicity, efficiency and engineering seemed amazing and I strive for those qualities in my software. Google Chat has none of these qualities, it's slow (really slow), complicated and buggy. What's happening to Google?

I hope mozilla the corporation dies faster so we could then focus on the browser. As someone said, the board is a joke.

ChatGPT wasn't built yesterday and it's not a desktop app so they can update often, why they don't work on/improve the frontend? It still feels like an MVP. I've seen some high salaries mentioned and an impressive headcount, how many of those are even working on ChatGPT or B2C in general? Sorry to rant, it just feels wrong.

We don't even know what a world with less death would look like, if we could be able to delay death, why wouldn't then be able to focus on stagnation, overpopulation and resources. And if something can be done I believe it will be done, if we want it or not.

I'm really sorry to hear about your mom. It's tough to see a loved one going through something like that. Wishing her strength and recovery, and hoping you get to spend many more happy moments together.