Secret Service shouldn't protect vice president, because there's a higher target!
This gets repeated so many times, I wonder if we even should use crypto for anything that costs less than a billion dollars. Imagine that you can steal a billion by figuring out how to generate SHA-1 collision. You generate one, then steal the money. Then what? Then you still have a way to generate collisions.
Yes, the possibility of hash collisions is real and it's generally assumed we'll have SHA-1 collisions "real soon now". Just because there are more profitable things to attack doesn't mean that we shouldn't try protecting other systems. For example, consider MD5 collisions: anyone can generate one for free. If someone invented a way to produce collisions for SHA-1, after profiting from the highest profitable target, they will probably try to profit from the next profitable target, etc. and eventually we'll learn that there are collision. What's next? The next step is to protect our systems from collisions. So, we're back to square one.
The simplest way to protect from SHA-1 collisions, is of course switch to the better hash function. For example, BLAKE2 - it's faster and more secure. We'll win some time until it's broken (prediction: it won't happen anytime soon).
Another way to protect against collisions is to build your system in such a way that collisions don't matter. For example, they can use HMAC with a secret key.
This gets repeated so many times, I wonder if we even should use crypto for anything that costs less than a billion dollars. Imagine that you can steal a billion by figuring out how to generate SHA-1 collision. You generate one, then steal the money. Then what? Then you still have a way to generate collisions.
Yes, the possibility of hash collisions is real and it's generally assumed we'll have SHA-1 collisions "real soon now". Just because there are more profitable things to attack doesn't mean that we shouldn't try protecting other systems. For example, consider MD5 collisions: anyone can generate one for free. If someone invented a way to produce collisions for SHA-1, after profiting from the highest profitable target, they will probably try to profit from the next profitable target, etc. and eventually we'll learn that there are collision. What's next? The next step is to protect our systems from collisions. So, we're back to square one.
The simplest way to protect from SHA-1 collisions, is of course switch to the better hash function. For example, BLAKE2 - it's faster and more secure. We'll win some time until it's broken (prediction: it won't happen anytime soon).
Another way to protect against collisions is to build your system in such a way that collisions don't matter. For example, they can use HMAC with a secret key.