jzombie

Do you foresee this being faster than SIMD for things like cosine similarity? Apologies if I missed that context somewhere.

  Location: Florida on paper (U.S. citizen born & raised); Currently residing in Nicaragua
  Remote: Yes
  Willing to relocate: No
  Technologies: Rust, TypeScript, Python, Docker, some PyTorch / Tensorflow
  Résumé/CV: https://www.linkedin.com/in/jeremyharrisconsultant/
  Email: jeremy.harris@zenosmosis.com

I'm timezone agnostic and on-call.

I am using ChatGPT to help me build a language that has a domain-specific use case with non-LLM models. The interpreter is written in Rust and the language resembles a mix of Lisp and SQL.

I have something similar.

https://github.com/jzombie/globcat.sh

Nothing fancy, but gets the job done.

It's not a tradeoff that is worth it to me.

I don't want a Facebook extension to have access to arbitrary data that I copy and paste in my browser, which may be unrelated to the app I'm developing at the time.

I just responded to an adjacent query with the info.

https://news.ycombinator.com/threads?id=jzombie#42072665

It seems to do well for a lot of searches, though some are questionable, but I believe that I know why. I'm training some different autoencoders to give it some different perspectives.

The code lives here: https://github.com/jzombie/etf-matcher

The ad-hoc vector DB I've created lives here: https://github.com/jzombie/etf-matcher/blob/main/rust/src/da...

> Likes/dislikes are stored in local storage and compared against all stories using cosine similarity to find the most relevant stories.

You're referring to using the embeddings for cosine similarity?

I am doing something similar with stocks. Taking several decades worth of 10-Q statements for a majority of stocks and weighted ETF holdings and using an autoencoder to generate embeddings that I run cosine and euclidean algorithms on via Rust WASM.

Lynx wrapped in Docker:

https://github.com/jzombie/docker-lynx

I am working on something similar and applied for your role! I prefixed my email with "ETF Matcher" just to correlate it here.

Maybe build something to do stock (or crypto, or some other market) analysis of some sort?

FaxGPT

globcat.sh is a tiny little shell script that concatenates multiple files together, as specified by a glob pattern, and sends them to standard output.

I think it's safe to say that if the tensorflow-metal plugin is no longer supported, the project is dead on Apple.

I'd like to be wrong here, and while I'm making assumptions, I'd say that probably anything else related to neural networks on Apple is dead as well.

I agree. This advice should not be ignored; and if the OP is complaining about performance, the answer is lying here in plain sight.

I may have misinterpreted this comment by thinking you meant that as we squeeze more and more transistors into a small amount of space, the resulting waveforms would start to resemble analog more so than digital.

The enclosure of that device looks very similar to a Focusrite Scarlett Solo. Curious if it happens to be a repurposed one.

> Don't think any algorithm is going to help me with that though.

Quant trading is exactly that.

I'd get rid of the fake reviews. After reading that, I definitely don't trust the "Over 50+ clients."

I don't think it's a viable business model if you flat out have to lie about it and don't even know who your target audience is.

Excellent article. Judging by your graduation year I estimate that I am roughly 8 years older than you. I am hopeful that you still have plenty of time.

Self-proclaimed leaders and experts basically signal to me that they stopped bothering to upskill, leveled off, and use their new title as a form of justification or personal flattery.

What evidence is there that intelligence can exist without neurons?

Is your function to also remove and simplify information? I ask because that seems like a pretty simple explanation. I was trying to study ML, but think I should just read children's books, or maybe not even read at all.

So, what you're telling me is that every thing they say has already been said before, completely verbatim? Like, if I asked it to write a story about a dog named Jebediah surfing to planet Xbajahabvash, it would basically just find a link to someone else's story that wrote about the same dog surfing to the same planet? That sounds like an infinitely large amount of combinations. Perhaps the internet is just infinitely large, squared (or even circled).

So, if I write in, "yes, no" it will output, "maybe?" or will it just do a Google search, or will it just output either "yes" or "no?"

And more importantly, will it give the same response every single time?

So, you're saying an LLM is a just a database that does text retrieval?

> LLMs don't increase information complexity in any way.

False.

LLMs can synthesize and combine information from their training data and generate new insights, ideas, and expressions that weren't explicitly present in the training data.

This adds layers of interpretation, inference, and creativity that enhance the complexity of the information they produce.

> ...by the very nature of their mathematics

What are their mathematics? Please enlighten me.

I don't usually answer it.... usually, my voicemail answers it. I wonder if that qualifies as an answer as well.

If I'm not mistaken, I believe this issue is worse on Sonoma than Ventura, from what I have read so far.