Xenakis, one of my favorite composers, was an architect originally. He went on to write - and utilize computers to write - a number of stochastic compositions. I always liked to think that, when asked what was going through his mind when he wrote a particular passage, he would say, "I thought, 'Good job, computer!'" Now, we laugh, but he was pretty Cagean in his use of stochasticism - like Cage he picked the "best" outcome from a series of chance operations. Xenakis "just" added his own mathematical basis alongside the randomness. In short, just had more inputs in the decision tree before selecting the "best" outcome.
The more inputs you have for a complex structure, the better chance that you'll like the outcome. Sure, you can leave it all to chance - if you roll dice enough, eventually you'll get a series that pleases you. But knowing, "I want the piece to start loud, be about X minutes long [perform operation to determine total length before first subroutine, but between 15 and 30 minutes], have eight major sections, and be quiet in the seventh of those" is pretty far along.
When I get on my high horse about "more inputs!", "more business requirements!", "more resources!", I'm really not trying to plan minutia. I'm trying to improve the likelihood that we'll have a pleasing result. To pull us back to "what are we trying to do here, other than implement this thing here because we can?". Because we often just don't have time to roll the dice as much as Cage did, given how complex the desired result can be.
So technology, architecture, and music DO have something in common: they can be structured, less structured, or both. The latter often gives unexpected and wonderful results, but knowing what to leave to chance informs how long you have to wait for the desired result.