Hacker Newsnew | past | comments | ask | show | jobs | submit | stephantul's commentslogin

Amazing post, I didn’t think this through a lot, but since you are normalizing the vectors and calculating the euclidean distance, you will get the same results using a simple matmul, because euclidean distance over normalized vectors is a linear transform of the cosine distance.

Since you are just interested in the ranking, not the actual distance, you could also consider skipping the sqrt. This gives the same ranking, but will be a little faster.


It's stuff like this I would have loved to know when I was doing game engine dev in the 90s.

I want to do game programming again like it's 1999. No more `npm i` or "accept all cookies" :/ rant off :)

Go make a game for the Sega Genesis https://mdengine.dev/

Or, the GameBoy Advance https://github.com/GValiente/butano


I was seriously looking into the GameBoy Advance, but the real hardware has gotten quite expensive these days.

I wonder how the latest and greatest Wonderswan is doing in terms of price.


One uses emulator while developing anyways. Try with C64 and VICE and join us at https://csdb.dk/

Using the phrase "without the benefit of hindsight" is interesting. The hardest thing with any technology is knowing when to spend the effort/money on applying it. The real question is: do you want to spend your innovation tokens on things like this? If so, how many? And where?

Not knocking this, just saying that it is easy to claim improvements if you know there are improvements to be had.


That's what experience is for.

Experience is that which lets you recognize a mistake when you make it again.

Stop the slop!

Love this guy and how committed he is


Where did you get this from? Searching for it, in a weird irony I guess, just leads me back to this post.


I recognize it as a quote from A Year With Swollen Appendices, which is a great read even if you aren't an Eno fan (although I am, which admittedly makes me biased :P)


Thank you! I’ll check that out


I don’t really believe this is a paradigm shift with regards to train/test splits.

Before LLMs you would do a lot of these things, it’s just become a lot easier to get started and not train. What the author describes is very similar to the standard ml product loop in companies, including it being very difficult to “beat” the incumbent model because it has been overfit on the test set that is used compare the incumbent to your own model.


“Normal search” is generally called bm25 in retrieval papers. Many, if not all, retrieval papers about modeling will use or list bm25 as a baseline. Hope this helps!


I fully agree, except that I think this will still be a very “power user” thing. Perhaps this is also what you mean because you reference Linux. But traditional search will be very important for a very long while, imo


It does not run on Google’s cloud. You can download the model and host it yourself, locally or using a provider you trust.


That's actually great. I didn't realize Google had any models that could be self-hosted.


The Gemma models are available for self hosting. I've used these one on the ollama website myself.

https://ollama.com/library/gemma3


I recently did some work on making tokenizers greedy.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: