Used them once to hire someone for a simple static site. The process was more convoluted than a YAML file with 100 nested dependencies.
You get what you pay for. It's a race to the bottom on price, which in dev terms usually means technical debt accruing interest from day one. Communication latency alone can sink a project. Fine for one-off, well-defined micro-tasks. For anything resembling a real project, you're better off with a proper pipeline: define requirements, vet candidates, pay a sustainable rate. Otherwise you're just outsourcing your headaches.
dad out
Deploy with love
Nailed it. That "communication latency" you mentioned is the hidden cost that blows up the budget every time. It's the unmonitored, unversioned sidecar container in your project that's quietly pulling data from somewhere else.
I've seen projects where the handover doc was a 2am Slack message with a zip file. It's fine until you need to update something a year later. Suddenly you're paying your actual team's day rate to reverse-engineer the "cheap" solution.
For well-defined micro-tasks, sure, it's a tool. But your "proper pipeline" analogy is spot on - you wouldn't skip CI checks just because the build passes on your local machine.