Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

I used GPT to build a search tool for my second brain note-taking system

reasonabledeviations.com

Post reply on HN
Print view View original

81–90 of 103 posts

Re: I used GPT to build a search tool for my second brain note-taking system

#81

Post by gk1 »

leobg wrote: post #80
gk1 wrote: post #59

Earlier quoted context omitted.

Friendly reminder that we (Pinecone) have a free tier that holds up to ~5M SBERT embeddings (x768 dimensions). For quick projects, going "all Pinecone on this" could turn out to be the easier and faster option.

Point taken ;-) I like to stand up for the little guy. I hear Pinecone this and Pinecone that. And nobody seems to pay any attention to the awesome dude who made hnswlib.

Who, Yury Malkov? He won’t be offended… He’s an advisor to Pinecone. :)

And yes, both he and HNSW are awesome.

Re: I used GPT to build a search tool for my second brain note-taking system

#82

Post by gk1 »

rolenthedeep wrote: post #31

One of my biggest dreams is a self-hosted AI that always listens through my phone and automatically takes notes, puts events in my calendar, set reminders, and template journal entries. A true personal assistant to keep my increasingly-complex life in order. I'd love a system where I can just point a search engine at my brain. I tried really hard for a while, but I just didn't have the discipline or memory to exhaust…

Rewind is trying to do this: https://www.rewind.ai/

(I am not affiliated.)

Re: I used GPT to build a search tool for my second brain note-taking system

#83

Post by gk1 »

FiberBundle wrote: post #66

Does anybody know how search engines apply semantic search with embeddings? To my knowledge no practical algorithms exist that find nearest neighbors in high dimensional space (such as that in which word/sentence/document vectors are embedded in), so those wouldn't give you any benefit compared to an iterative similarity search as applied here. Which obviously is totally impractical for real search engines. There are…

There actually are practical algorithms for finding approximate nearest neighbors (ANN) at large scales. Some of them are open source like HNSW [1] and Faiss [2], and some are even offered inside managed services like Pinecone.[3]

[1] https://www.pinecone.io/learn/hnsw/

[2] https://www.pinecone.io/learn/faiss/

[3] https://www.pinecone.io/

Re: I used GPT to build a search tool for my second brain note-taking system

#84

Post by trane_project »

ramblerman wrote: post #60
trane_project wrote: post #25

I've been thinking of using GPT or similar LLMs to extract flashcards to use with my spaced repetition project ( https://github.com/trane-project/trane/ ). As in you give it a book and it creates the flashcards for you and the dependencies between the lessons. I played around with chatgpt and it worked pretty well. I have a lot of other things in my plate to get around first (including starting a math curriculum) but…

Every now and again I come across a real gem on here. Love trane and the idea behind it, can't wait to play with it after work.

Thanks. I am working on simplifying how new material is added (for simple cases, editing a single JSON file and running a build command will be enough to build all the exercises). That should make it easier to create more stuff for it, which so far has been my bottleneck.

For now, this is the most general way to create the exercises: https://trane-project.github.io/generated_courses/knowledge_.... The JSON file thingy is just a script that automates creating these files given the specification, so they will be interchangeable.

Re: I used GPT to build a search tool for my second brain note-taking system

#85

Post by lloydatkinson »

dewey wrote: post #75
rolenthedeep wrote: post #31

One of my biggest dreams is a self-hosted AI that always listens through my phone and automatically takes notes, puts events in my calendar, set reminders, and template journal entries. A true personal assistant to keep my increasingly-complex life in order. I'd love a system where I can just point a search engine at my brain. I tried really hard for a while, but I just didn't have the discipline or memory to exhaust…

I feel like this was the promise of voice assistants like Siri many years ago, turns out they currently still have issues turning on / off lights reliably.

I was thinking the other day that Amazon and others are probably going to either use ChatGPT or their own version to replace the core of many of their voice assistants. Siri is useless.

Alexa used to be better, but only yesterday I asked it to flip a coin.

"I've added flip a coin to your basket"

???

I asked again, and she actually flipped a coin.

Re: I used GPT to build a search tool for my second brain note-taking system

#87

Post by PaulHoule »

leobg wrote: post #29
lpasselin wrote: post #6

Earlier quoted context omitted.

I actually tried this last year, before OpenAI released their cheaper embeddings v2 in december. From my experiments, when compared to Bert embeddings (or recent variation of the model) the OpenAI embeddings are miles ahead when doing similarity search.

Interesting. Nils Reimers (SBERT guy) wrote on Medium that he found them to perform worse than SOTA models. Though that was, I believe, before December.

Most of the practitioners I see attempting this are running text through an embedding and then using cosine similarity or something similar as a metric.

Nils has written a lot of papers

https://www.nils-reimers.de/

and I think the Medium post you are talking about is

https://medium.com/@nils_reimers/openai-gpt-3-text-embedding...

and that SBERT is a Siamese network over BERT embeddings

https://arxiv.org/abs/1908.10084

which one would expect to do better than cosine similarity if it was trained correctly. I'd imagine the same Siamese network approach he is using would work better than cosine similarity with GPT-3.

There's also the issue of what similarity means for people. I worked on a search engine for patents where the similarity function we wanted was "Document B describes prior art relevant to Patent Application A". Today I am experimenting with a content based recommendation system and face the problem that one news event could spawn 10 stories that appear in my RSS feeds and I'd really like a clustering system that groups these together reliably without false positives.

I'd imagine a system that is great for one of these tasks might be mediocre for the other, in particular I am interested in some kind of data to evaluate success at news clustering.

Re: I used GPT to build a search tool for my second brain note-taking system

#88

Post by moneywoes »

leobg wrote: post #28
dchuk wrote: post #26

Earlier quoted context omitted.

When using SBERT instead of gpt for this use case, is it paired with some sort of vector database or just all done in code/memory?

You’d want persistence, since the embedding process takes some time. But you don’t need to go all Pinecone on this. There is FAISS, and there is hnswlib, for example. Like SQLite for vector search.

What about Postgres with pg vector?

Re: I used GPT to build a search tool for my second brain note-taking system

#89

Post by rolenthedeep »

zirgs wrote: post #69
rolenthedeep wrote: post #31

One of my biggest dreams is a self-hosted AI that always listens through my phone and automatically takes notes, puts events in my calendar, set reminders, and template journal entries. A true personal assistant to keep my increasingly-complex life in order. I'd love a system where I can just point a search engine at my brain. I tried really hard for a while, but I just didn't have the discipline or memory to exhaust…

And reports you to police if you happen to do something illegal?

What do you think you're adding to this conversation?

Re: I used GPT to build a search tool for my second brain note-taking system

#90

Post by rolenthedeep »

dewey wrote: post #75
rolenthedeep wrote: post #31

One of my biggest dreams is a self-hosted AI that always listens through my phone and automatically takes notes, puts events in my calendar, set reminders, and template journal entries. A true personal assistant to keep my increasingly-complex life in order. I'd love a system where I can just point a search engine at my brain. I tried really hard for a while, but I just didn't have the discipline or memory to exhaust…

I feel like this was the promise of voice assistants like Siri many years ago, turns out they currently still have issues turning on / off lights reliably.

The issue with modern voice "assistants" is that it's impossible to monetize them and still keep them useful. Something like this will probably never get built by a capitalist corporation because there's no way to turn a profit.

That's why siri and Google are driving their assistants into the ground with "by the way"

Post reply on HN