Monorepo or Polyrepo for Microservices, how do you balance speed and all that complexity?

Hey everyone, I'm stuck on this classic problem: should I go with a monorepo or a polyrepo for a new microservices project? A monorepo could mean easier dependency handling and refactoring across services, which sounds great for getting stuff done faster. But then, I worry about really long build times, what if some tool breaks everything, and just how much mental energy it takes to deal with one huge repo. Polyrepos, though, give each team clear ownership, let you deploy things on their own, and keep codebases smaller and easier to handle. That definitely fits the microservices idea. But I can totally see a future with dependency hell, versioning headaches, and a messy development experience. What have you all seen? Did you find a good middle ground, maybe a hybrid way that uses the best parts of both? What kind of trade-offs did you hit with your DevOps, SRE worries, and how fast your team could move?