Well the article is right in one sense: Postgres full text search is probably better than you think it would be and is very usable at non-massive scale
However, I still wouldn't recommend using it. The queries and indexes are all unintuitive, complex, not very readable and require learning a bunch of new syntax anyways (as you can see in the examples in the blog post). So just take the effort you would spend doing that and instead learn something with more features that will continue running no matter how large your scale.
While the blog mentions elasticsearch/solr (which are awesome and powerful), it doesn't mention sphinx (http://sphinxsearch.com/). If I were trying to make a simple search engine for my postgres/mysql powered site, I'd use sphinx. The indexer can be fed directly with a postgres/mysql query, and it includes simple APIs for all languages to interface with the searcher daemon. And of course it is super fast, has way more features and can scale larger than you'd ever need (sphinx powers craiglist's search).
However, I still wouldn't recommend using it. The queries and indexes are all unintuitive, complex, not very readable and require learning a bunch of new syntax anyways (as you can see in the examples in the blog post). So just take the effort you would spend doing that and instead learn something with more features that will continue running no matter how large your scale.
While the blog mentions elasticsearch/solr (which are awesome and powerful), it doesn't mention sphinx (http://sphinxsearch.com/). If I were trying to make a simple search engine for my postgres/mysql powered site, I'd use sphinx. The indexer can be fed directly with a postgres/mysql query, and it includes simple APIs for all languages to interface with the searcher daemon. And of course it is super fast, has way more features and can scale larger than you'd ever need (sphinx powers craiglist's search).