Hello guys!

I'm Bálint from Hungary, medical doctor, software dev, co-founder and builder of a next-gen vector database engine, metriq0.

I'm sure you've recently read a post about something like "I've fired my developers to use AI agents instead, now I'm broke bcs the bills are so high.." - well, yeah. If you think about AI coding apps, for example Cursor, they also have their enormous expenses, partly vector database costs (~30%).

Okay, but what are vector DB-s used for? You can think of them like databases for unstructured data where you can search based on similarity. They are used as the long-term memory of chatbots or the core of recommendation algorythms. And, they tend to be VERY expensive, if you pump in a lot of data.

Existing vector DB companies are tweaking the same 20 year old vector search algorithms, but their innovation does not match the need.

So here comes the idea, lets rethink the mathematical foundations of vector search, to be able to run it on cheaper hardware - without any compromises on speed and precision. It must not be as hard after all.. - that is what we were thinking when committed our first lines 2 years ago. [Narrator: it was very hard]. 😆

Today we are close to publish our first benchmarks and inviting people to try our engine. There is still a long way to go, but we are very excited!

I'm here to connect, chat, meet with likeminded people from the AI industry! Feel free to write me anytime 😊

Oh and this is our webpage:

metriq0.com