Unfortunately, I think you can only talk about consciousness in the Neuroscience world if you have a Nobel prize. However, I think the argument still stands that it's a reasonable assumption that your entire conscious facility derives from the brain. Unless you want to make an argument for a soul, then it's fair to assume neurons and glial cells are the responsible parties.

Do we know for certain that quantum effects within neurons don't contribute to consciousness? No, but the most parsimonious answer would likely be that that's not the case. We have neither evidence nor reason to require such a mechanism to exist. What do neurons lack that cannot account for consciousness? If you can answer that, then I would allow the possibility of some other force in effect.

There are plenty of Neuroscientists who study parts of neurons, so it's not quite atomic, but overall you are right. If you take a patch of a neuron (separate it out from the rest of the cell) and record from it with an electrode, you can actually see random fluctuations in voltage that are unitary in size (that is, the voltage will jump in set units). What you are observing is the channels that conduct current, made up of proteins that span the membrane, opening and closing. These fluctuations are actually caused by random events in the dish. However, these fluctuations stop if you look at the same channels in an intact cell. What happens is the membrane of the cell helps to average out all of these random fluctuations. Thus, biology is relatively good at dealing with the randomness of Nature, and is unlikely to be affected by quantum fluctuations.

Hibernate is a pretty good product. It has nice features that allows you to abstract tables into OOP representation. Joins can be specified optionally on column values. And while the XML can be hard to parse, it now allows for annotations to be done in the Java code itself.

That said, when Hibernate fails, it can often be hard to figure out why. It's error messages can be hard to parse, and difficult to trace where it originates. Also, I've definitely come across places where hacks had to be implemented into object relationships in order to get cascades to work properly.

If Hibernate has any really big issues, is that it probably promises to do more than it actually can. But with that said, if you need Hibernate to do things outside of its general scope, you probably don't want to use SQL in the first place.

Instead of placing all your requirements on a single product, it might be worth figuring out how each dataset can be stored most optimally. If you're doing a lot of key-value storage, you can store large datasets with MemCached, and write to disk.

If you need to do relational queries on some of that data, you can create a SQL representation that points to the data you want. Also SleepyCat DB might be worth checking out.

Apple has jumped in on the fun in this regard, by making key-value bindings a central part of Cocoa. Again, there are no silver bullets -- use the best tool for the job as always..