Life is metabolism, so a computer program or a silicon-based intelligence cannot be considered alive by current definition. Either we rework the definition or it will definitely not be alive. Also, sentience and intelligence are two very different things, I think we are very far from having ethical problems with switching off computers (if ever).
Computers and their programs consume energy in the form of electricity to do work. How is that not metabolism? I recall seeing a reference to a biological species that also consumes electricity.
In addition, they use that energy to (often) lower internal entropy and replicate themselves. They evolve with the assistance of a symbiote (humans), and are subject to competitive pressures for evolutionary fitness.
Mating is almost unheard of, but composition of parts is rampant.
That's exactly the difference, computers consume energy, while metabolism is the process of creation (and of course, consumption) of energy. An unplugged computer won't start to autonomously break down and convert chemicals into electricity to keep itself running, hence it's not alive.
You know what I meant, but in case it's needed, that's why I qualified my sentence with "autonomously". And no, a set of software/actuators to regularly plug the laptop into mains is not the same kind of autonomy, since it would just be the execution of an algorithm that a sentient being programmed. Without that, the laptop won't be subject to any necessity or pressure to harvest energy.