2^80 is a lot of work, but it isn't enough to be considered secure by current standards.
Actually it is more than enough:
If all the parties of the bitcoin network with the current hash power of (almost) 2^62 H/s reach a consensus and decide to just find one single 160 bit collided pair of keys, they have to devote their total power for the next 1000+ years or so to do the job. (Note: we have to run 2 hash functions in each test).
Check your math. 2^62 hashes per second does 2^80 work in 2^18 seconds. That is three days.
OK. My bad. But the main argument remains valid: It takes a definitely huge amount of resources and time to collide with 168 key hashed addresses and for the near future the cost of such an attack can not be justified while
bitcoin block chain is a monetary infrastructure and the attacker MUST justify his/her costs.
I insist you and other guys pay more attention here: We are not discussing security as a general problem. It is not about sensitive data related to nuclear missiles or family scandals, data with infinite or undecidable values. Our context is securing contracts on top of the Blockchain.
As of 160 bit hashed address key problem, the hypothetical attack can compromise a contract on the Blockchain, but nobody puts such a huge amount of money on a single contract. He/she should not do this and Bitcoin is not and have not to be adequate for such a contract.