Well, that’s not really possible. The problem is that for any laser the phase wanders randomly over time compared to an absolutely perfect frequency reference. The more stable the frequency and the narrower the line width the slower the wander, but the phase will wander. For two sources to wander around together they have to be tied to the same random walk. That can mean splitting a single source, or it can mean seeding two sources with the same seed laser. It can also mean actively driving the phase of different sources to match that of some reference source. However, in all cases one way or another all outputs are tied back to one reference source.

Now, if you make two independent sources, their phase difference will stay constant for something like whatever the shorter coherence time of the two. However, while you can expect the phase difference to not drift much on time scales short compared to the coherence time, at any given moment you have no idea what that phase difference is and it will wander through all possible values on a time scale long compared to the shorter coherence time. So that lack of knowledge or control of the relative phase prevents you from doing all the things we do with coherent sources. Yes they are coherent for useful lengths of time, but with an unknown and varying phase difference.