This article is interesting but Edo seems to dump hard on MongoDb. Anyone with more tech chops on vector tech to shed light on his claims ? Namely Pinecone is the only true vector db? Seems a bit self aggrandizing https://www.zdnet.com/article/pinecones-ceo-is-on-a-quest-to-give-ai-something-like-knowledge/ My take is if vector db are so important to GPT then Google, AWS, etc (Mongo) will offer them and make a better competitive landscape. Edo seems to be trying to be like a new monopoly in a tech that honestly id be shocked if others couldn’t replicate
I’d be shocked if pinecone (a startup amongst very well known and positioned competitors) comes out on top, any team that uses a vector db should just use an extension like pgvector on any Postgres db rather than fully managed and far more expensive pinecone. In other words— pinecone is very easy to use and simple but expensive. But if your team is using a vec db in the first place you shouldnt really be worrying about ease of use and more so latency and cost at scale. Also there are approx 20 platforms tools/extensions to built a vector db. You also don’t even need a pure vec db to have the same sort of search using cosine sim or dot product I have been wrong before idk
Can't decide what is the bigger letdown, your ending comment or GoT's ending..
@BARD. does that mean everything before it was among the best comments you've ever read?
Some funny stuff here , did not see that coming :) plot twist Okay good stuff. It’s just wierd to me to read this Edo guys comments like he’s the only one who knows how to build one, when at the same time he says he learned how to do it at Amazon. this is how CEO talk when you take 150 million in funding and are called Edo Liberty
The vector db landscape is already very competitive. Many startups, Azure/cloud services offering their own solution
Pinecone is a wrapper on top of GCP actually