maCDzP

I used the embeddings from Voyage for a project in Swedish this summer.

The neat thing about Voyage was besides the speed of the service.

I think I had 250 million tokens and Voyage was the fastest. It took a couple of days on and off. I believe my napkin calculation showed that OpenAI would have taken months.

What a lovely community, right up my alley. Thank you!

Any tips if someone want to join in the Nordics?

SQL clicked once I read the original paper and it explained it terms of sets.

I want to crawl and plug in scihib to this and see what happens.

Have you seen this? https://github.com/icloud-photos-downloader/icloud_photos_do...

I am going to see if this can be used to forecast wins for sports betting.

I spent some time learning those queries for a project and when I grokked it they where very handy.

Let’s see where this goes in 1-2 years. There is a ton of money in the CAD industry.

Interesting, I got the opposite result. The full embeddings gave two or three clusters. How did you work with the hyper parameters of HDBSCAN?

So anyone up for a discussion on this paper at: https://www.alphaxiv.org/abs/2403.18103

Did you reduce the dimension before applying HDBSCAN?

Amazing video! For those of you that want to experience it - check out https://sailtraininginternational.org

I was born after Max Headroom aired, but for those of you that saw it while it aired, how was it?

I read in another thread that people are doing dev in a container. It’s supposed to be awesome.

How does this compare do developing inside a container?

This is what makes me come to HN every single day.

Harvest Coccoliths that have been around for millions of years and are believed of being responsible for the majority of carbon sequestration. Plus, they are pretty!

If this sort of thing interest you I can recommend the authors podcast “Cautionary tales”.

Y could probably build something with the shortcuts app that goes a long way - surprisingly long.

I haven’t read the paper yet but I want to share my immediate thought.

So the harder it is to “convince” the LLM on a wrong answer, that’s a proxy for low epistemic uncertainty?

I know you shouldn’t view the LLM as a “mind” - but I can’t help myself!

Also bookmarking to think about it.

My mind drifted towards Fourier transform. Using the transform as a way of describing a system with less entropy?

Or am I butchering all of mathematics by making this comparison?

I believe there is a 99% invisible podcast episode about “Great bitter lake”. It’s worth a listen.

It would be cool to write some tests to see how often it works out. I have noticed that LLM:s often creates command line options that doesn’t exist.

Security aside, I bet it would work more often than when I input something in the terminal.

I nice feature would be to just loop back the error and get ChatGPT to correct the error. You do this by running the command with bash -n (syntax check) and when it doesn’t return an error it runs the script.

Three months from now the next cloud outage at Google will be from “helpme delete that one weird file”.

May the lord have mercy on our machines.

Cool project using Processing, I’ve always wanted to play with that.

I want to this with Factorio. Build a huge computer within Factorio made out of belts. Make it seg fault and break out of the game.

What, Simone Weil the philosopher and André Weil are siblings?

That’s really cool.

I can recommend Philosophy This podcast on Simone Weil for those interested.

That is a beautiful website. I got lost in it and it created a sense of wonder. Nice.

I learned a ton from doing “Over the wire”.

After that I read “The Unix programming environment” and after that it clicked.

Story time - a thing I learned from TUPL.

I did an interview at the and they had a bunch of programming questions. One was to write a the shortest script to count lines in n a script ext file. I asked - can I choose language, ANY language?

From the book I remembered the command wc -l in Bash. The interviewer looked at the his colleague gave a small smirk. I felt right at home.

Thanks, that was a lot of insight, I appreciate it. Congrats for finding and executing a profitable strategy.

Could you expand on your trading bot? To bad for NASA denying your mugshot. I bet it would have generated many stars on GitHub.