To know that we don't know
The reason I no longer participate in such discussions is that they usually settle on just two positions:
1) Free will is irreconcilable with physics as we know it, therefore it must be a supernatural phenomenon, thus proving the existence of the soul, God, etc.
2) Free will is irreconcilable with physics as we know it, therefore it must be an illusion; that is, we don't really have free will and all our thoughts and actions are in fact mechanically generated by the deterministic laws of physics and possibly the random effects of quantum phenomena.
I'm not buying either of them. Humans have a rather long history of believing that anything they can't yet explain must therefore be supernatural. Lightning, disease, seasons, and the very existence of complex life forms were once thought to be explicable only by magic or by the intervention of a deity. Time passed, knowledge grew, and the real explanations were discovered.
That we do actually have free will seems to be so self-evident from experience that declaring it non-existent because it can't be fitted into the present theoretical model of physics looks like the kind of arrogance that no actual scientist would commit. Historically, observed phenomena which cannot be accommodated by existing theory are taken as a strong hint, at least, that the existing theory is incomplete.
Try to imagine people 1,000 years ago trying to figure out what makes the Sun shine. They could not possibly have arrived at the correct answer since it depends on a phenomenon (thermonuclear fusion) which was utterly beyond their knowledge at that time. It would be easy to imagine two factions developing -- one saying that sunlight couldn't be explained by then-known physics and must therefore be supernatural (God was causing it), and the other saying that sunlight couldn't be explained by then-known physics and must therefore be an illusion (the Sun is really not shining at all, we just think it is, due to some kind of flawed perception). Of course they would both have been wrong.
One certainly couldn't fault those people for not knowing about thermonuclear fusion. But one could fault them for refusing to recogize the possibility that the real answer might depend on some phenomenon not yet discovered.
Right now, all we can do is leave the problem of free will in that category. All we can say right now is that we have an observed phenomenon which doesn't fit the current model, and that we don't really know what that means. The next few decades will almost certainly bring a full and detailed understanding of human brain functions, simulation of those functions in computers, and eventually uploading of human minds into computer systems, allowing for a trillionfold increase in our intelligence, emotional sensitivity, and richness of sensory environments. That is, we will soon have vastly more actual data about free will (and every other aspect of how our minds work), as well as vastly increased ability to analyze such problems. We will figure it out. Until then, the honest thing to do is to admit that we don't know the answer.