HN Explorer

11 comments

[ 2.6 ms ] story [ 42.1 ms ] thread
softwaredoug
It’s called Generic Neural “Elastic Search” - but does it use “Elasticsearch?”

Sorry if it’s a dumb question, poking around I don’t see where it interacts with Elasticsearch (maybe it doesn’t)

tlarkworthy
My guess: I think they mean the searches are elastic (i.e. F fuzzy). This is the same meaning that Elasticsearch the product emphasises through its product Name. There is crossover in functionality
biomodel
Seems like elastic here is a reference to the scaling and not to ElasticSearch. It searches by comparing vectors encoded by a neural net
caiocaiocaio
Names sounds like an NES emulator. Maybe one using GTK+.
blondin
man... people are fussy nowadays... i had the same exact thought of an NES emulator. what's wrong with this generation of hacker news people?!
vkaku
Yeah, please give it a different name, OP. It looks like you've put a good amount of work into the project itself;

However, as a normal tech user looking at various things, I would be confused by looking at the name too. Please give it a unique name - find one that isn't already on GH, PyPi or what not.

TL;DR: Brand identity = Unique name

pantulis
The project looks impressive, but it doesn't seem clear to me at a first glance what does it do or the advantage it has over other solutions.
JabavuAdams
What is "cloud-native semantic search"? What does it do?
z3t4
I'm not sure either, but: cloud=someone else's computer, cloud-native=living on someone else's computer, semantic=meaning of things, semantic-search=Find the meaning of something. Maybe it can explain what it self is.
jarym
The fact that any technology runs Dockerised or within K8 can’t be its principal selling point (above what the tech itself does) imo.

This ‘looks’ cool but I had to read a lot and still had to make some guesses about what it does.

How about a comparison of some real world use cases? For example, could I use this to store and query some structural information that might come from an hOcr or ALTO formatted ocr output? Can OCR even be plugged in? Can I intelligently index the diff of multiple versions of a file and query that?

I have no clear idea ‘cos the team behind it are so focussed in their literature on the technical detail.