Discover, evaluate, and access relevant embeddings in your go-to framework. Skip all the infra for scraping, cleaning, indexing, and updating high-quality embeddings.
For teams building Layer 2 solutions for enterprise customers, bing/google based web browsing is a huge pain point. Being able to hit higher quality public stores of information is a great time saver.
Will be nice to know their approach to ranking and curation of these embeddings. Once you open the floodgates, signal-to-noise gets bad fast. But, if you can ensure clear differentiation between high-quality knowledge bases with SLA contracts and the huge number of 'free' user uploads then I can see this taking off.
Complete tangent. *embedding.store* - how is this url still available ?
This looks like a great idea to me. I mean, why go through the trouble and cost of cleaning-up, embedding, storing, indexing, and querying a vector database for snippets of text when someone else can do it for you at significantly lower cost?
This is amazing, data companies out there that still spend a huge amount of time + developer cycles, building and maintaining infra for scraping and cleansing even from high quality sources can benefit massively.
Highly relevant to POI/Geo data and many more industries as well!
11 comments
[ 3.1 ms ] story [ 42.2 ms ] threadWill be nice to know their approach to ranking and curation of these embeddings. Once you open the floodgates, signal-to-noise gets bad fast. But, if you can ensure clear differentiation between high-quality knowledge bases with SLA contracts and the huge number of 'free' user uploads then I can see this taking off.
Complete tangent. *embedding.store* - how is this url still available ?
Highly relevant to POI/Geo data and many more industries as well!