I'm kind of confused by the whole idea because I don't understand the lack of side effects. Knocking out the brain's learning system to cure alcoholism seems disproportionate, and I would also expect naltrexone to interfere with the ability to experience happiness (which many people seem to like).

From the Wikipedia article:

For tablet form, a patient following the Sinclair Method takes a 50 mg tablet one hour before every drinking session.

So it seems that the method is to time the disruption of opiod reinforcement to target when the person is drinking. Side effects may therefore be limited to other activities that typically coincide with drinking.

Though this also raises the issue of meta-akrasia, wouldn't people learn to not like taking the tablets if they don't enjoy anything they do for the next hour?