As programmers we strive to create clean modular code with good separation of concern and little to no boiler-plate. But for the longest time however, multithreaded code has gotten the better of us by introducing challenging concepts such as semaphores, double checked locking, starvation, deadlocks, race-conditions and other unfortunate heisenbugs. With Carl Hewitts message-oriented actor model it all got much easier though. Languages like Erlang, Scala and F# offer a way to model both asynchronous, concurrent and parallel systems in clean cohesive modules which retains the maintainability of single threaded code, relieves the cognitive overload of traditional multithreaded programs while keeping their great performance characteristics. In this talk we will look at how F# agents and the actor model in general can help you create the simple concurrent applications you dreamed of. We will also see how they can help us solve other problems like optimizing for batch processing and even make distributed systems almost as easy as those that run on a single machine.