Motivated by the intriguing functionality of gene regulatory networks we study chemical reactions (biological) circuits. We observe that those circuits are vastly different when compared to existing computing structures like logic circuits. In particular, the two strikingly different ingredients in biological circuits are feedback in memoryless computation and the stochastic behavior of devices in deterministic systems. Are these two biologically inspired concepts useful in improving the design of existing computing structures? I will provide a positive answer to this question and argue that progress in our understanding of biology depends on the development of new abstractions for reasoning about computation.