As it is formulated, it is not possible to answer that question. There are several memory systems in the human brain: you have unconscious memories such as those relating to movement or speech, and conscious memories, such as those for names and places. These memory systems are extremely liable and vary quite a lot between individuals, depending on their genetic and epigenetic background, and global health (some disease states can boost or impair memory at a certain point). You see the difficulty in ascribing a maximum "potential" to a "global" memory. Memory is indeed thought to be retained (most likely) in particular synaptic networks rather than in individual neurons, but we're still far from understanding how these biological entities translate electrochemical signals into information.

As it is formulated, it is not possible to answer that question. There are several memory systems in the human brain: you have unconscious memories such as those relating to movement or speech, and conscious memories, such as those for names and places. These memory systems are extremely liable and vary quite a lot between individuals, depending on their genetic and epigenetic background, and global health (some disease states can boost or impair memory at a certain point). You see the difficulty in ascribing a maximum "potential" to a "global" memory. Memory is indeed thought to be retained (most likely) in particular synaptic networks rather than in individual neurons, but we're still far from understanding how these biological entities translate electrochemical signals into information.

Another factor to keep in mind is that not all of the memories stored at a given time are later recoverable, i.e., "forgetting". So, even though a lot of information about recent experiences can be recalled, only a few details of more distant experiences are retained.

Something to consider is that memory is an imaginary construct used to account for a phenomenon (and sometimes quite different phenomena). Memory is often described as a change in behavior with experience, but memory can exist without any behavior. What we call procedural memory can seem quite different to episodic memory, which can also be different to emotional memory. So I think a question one can ask oneself is, 'is there a better way to account for how vertebrates adapt to changes in their environment that is more compatible with what we know about neurobiology?'

In addition to what's been said above, take notice of the special abilities of certain mnemonics, capable to retaining (and recalling) vast amounts of data (e.g. rajan mahadevan; http://en.wikipedia.org/wiki/Rajan_Mahadevan) or individuals with savant syndrome (e.g. Kim Peek; http://en.wikipedia.org/wiki/Kim_Peek). These are indeed abnormal cases, but they can teach us of the capacity of the human brain.

Here's an interesting article in Scientific American that deals with the question you have asked.

http://www.scientificamerican.com/article.cfm?id=what-is-the-memory-capacity

May be it will give you some idea about the capacity of our brain to store memories.

Hi, this question is right up my alley. I am being diagnosed with savant syndrome because of my LTM and I think you might be interested. I can redraw extremely complex shapes months or years later and I believe randomness allows me to do this. I am coding a test suite to prove it scientifically. No one has ever done this before, but you might also be interested in checking out another savant named Steven Wiltshire. He redraws whole cities over a few days after flying overhead in a helicopter. I'm not sure what the time limit is on his photographic memory, but mine appears to have no limit as of yet.

I think when discussing human memory capacity in comparison to a computer, it is often incorrectly assumed that it would be beneficial for a person to store more. In fact, forgetting is often adaptive, and our memories have evolved in such a way that useless information is forgotten (e.g. where you parked your car last week), whole even well-learned memories which are not used tend to become less accessible.

Overall, because of the way new memories are taken in, there is no obvious limit. An expert in a field doesn't experience their memory becoming full or a need to delete to make space, and indeed, it tends to become easier to learn about something the more you know about it (e.g. an expert of neuroscience could read an article about this and retain it much better than a novice). So the analogy of human memory using up fixed storage space is not a useful one.