Vast
Vast has been getting a decent amount of buzz today, especially given the TechCrunch story .
It’s Naval Ravikant’s new deal. I like the idea of content aggregators and I like Vast, but my very first search using the product demonstrates some of the inherent challenges with their approach.
They make structure out of unstructured data by assessing what they find during crawls of web sites for things such as auto listings from a dealership website. They essentially create a normalized view of things for sale.
Try searching in the Cars section for BMW as the make and 2002 as the model. The intended result would show me BMW 2002’s (the model) from the 60’s and 70’s. Instead, I get BMW cars and motorcycles that were built in the year 2002.
These types of problems are not insurmountable and don’t indicate a completely flawed methodology, but exposes some basic challenges with any product that attempts to make uniform presentation out of data that’s inherently unique.