Kant thinks that a free will is a will under moral laws and that freedom and the moral law are distinct thoughts that reciprocally imply each other. Fichte thinks they are the same thought.

If we decide rightly what to do, or use a correct procedure for making such decisions, that has to be because the decisions or the procedure rest on good reasons, and these reasons consist in the apprehension of truths about what we ought to do. Because these truths must constitute reasons for our decisions, and because in the rational order, reasons must always precede the decisions based on them, the truth conditions of claims about what we ought to cannot be reduced to, or constructed out of, decisions about what to do, or procedures for making such decisions.

Both Kant and Fichte thought of traditions of revealed religion as ways of symbolically (that is, with aesthetic emotional power) thinking about our moral condition. Both thought that religion would become more and not less powerful, emotionally and morally, if the claims of scriptures and religious teachings were taken symbolically rather than literally (whatever 'literally' might mean in the case of claims that are either nonsensical or outdated or historically unsupportable if taken as metaphysical or historical assertions).

What is central to morality is rational self-constraint (acting from duty), in cease where there is no other incentive to do your duty except that the moral law commands it.

It is both theoretically mistaken and morally wrong to regard others as objects of investigation rather than partners in free rational communication.

It is a cause of shame to any member of the human race to be a member of the same species some of whose members could vote for any candidate for president that has been offered by the Republican party. Such people seem to be motivated only by short-sighted greed, ignorance, fear and hatred.

Kant can provide, and has provided, a good model for philosophers to think about the relation of metaphysics to science and scientific methodology.