They are also helpful in terms of space savings. I worked at a startup in Palo Alto in 2011 where we used a 100 GB in memory (Redis) Bloom Filter to de-duplcat

They are also helpful in terms of space savings. I worked at a startup in Palo Alto in 2011 where we used a 100 GB in memory (Redis) Bloom Filter to de-duplcate links coming in from a real-time social media stream (Facebook, Twitter, Forums, etc..). While Bloom Filter does introduce false positives (collisions) we were willing to take the bet on that rather than allowing duplicate social media posts in our pipeline. If we wanted to store this information in a regular hash table, we would require a lot more space than the 1 bit per entry of a Bloom Filter.