The moral effort argument arises because of a vagueness in the criteria for evaluation of the outcome. Why would we want people to be more moral?

So they would deserve more moral praise? Or so they would do less of the things that hurt other people? If our goal is the former, the moral enhancer might be pointless since it would not cause the world to deserve more praise. If it is the latter, the moral enhancer might make a big improvement, and praise is beside the point. The unreliability argument seems to involve a fuzzy definition of "moral enhancer".

To me, that term means something that helps people act based on correct moral deliberation. If a drug's effect often systematically coincides with the outcome of moral deliberation, that's a morality substitute rather than a moral enhancer.

For instance, an anti-violence pill might reduce, for some people, the frequency of their immoral acts. We use a locked cell to stop some people from carrying out immoral acts.

These are not moral enhancers, they are morality substitutes. Good moral deliberation requires practice and experience. Learning to act on the results of your moral deliberation also requires practice and experience.

In the same way, playing a sport well requires practice and experience. Both involve honing judgment. Thus, the analogy with which the article started is more practical than the author seems to have supposed.

The challenge of designing a moral enhancer, a drug that would make you better at moral deliberation while skipping the practice and experience, has much in common with the challenge of designing a drug that would make you a champion at a sport while skipping the practice and experience of the sport. When technology can do the latter, maybe it can do the former as well.