If we take encrypt to mean hash (which is probably what the implementation would most likely mean), this still has the same problem that server-side hashing does. Hashing is fast, GPUs can burn through creating rainbow tables at incredible speed (see <a href="http://stacksmashing.net/2010/11/15/cracking-in-the-cloud-amazons-new-ec2-gpu-instances/" rel="nofollow">http://stacksmashing.net/2010/11/15/cracking-in-the-cloud-am...</a> ). Now if it's talking about implementing bcrypt (<a href="http://codahale.com/how-to-safely-store-a-password/" rel="nofollow">http://codahale.com/how-to-safely-store-a-password/</a>) in the browser that would be a real solution, but them the problem becomes adoption (at which point, you basically need to implement it on the server anyways). Plus, users would need to be signaled that this was a safe password box, vs. an unsafe one (which could easily be spoofed by malicious sites).