And now for my really wild theory. I wrote this theory section on Friday night, adding this entire introduction Monday morning.

My Wild Theory
At the end of Reloaded, we learn that Neo can both sense the sentinels and exercise some measure of control over them. The act of terminating the three sentinels sends Neo into an apparent coma. Could this be what happens when someone is too old and set in his days suddenly encounters the realization that he has been living inside the matrix? We were told in the first film that normally people are taken when they are young to avoid problems, perhaps like this one. Neo could have realized, seems to have realized, that he is not really in the real world at all, as we would recognize it, but instead still within an "uber-matrix." That would explain also the earlier scene in which an Agent Smith-infected "human" is sucked out of the matrix into the "real world." Yes, it was a dream, but Neo's dreams come true. Always. That was the agent that fell into the car, the first dream was accurate. Or there might be an alternate possibility that I'll address later.

Since we know by the revelations in the last few minutes of the film that this is an uber-matrix, I begin to question everything. We now know that the prophecies on which Morpheus has been relying are false, inventions of the matrix designed to get The One back in front of The Architext in order to "reassimilate his code". His code? Is the architect suggesting that Neo is "just" a program? I'll go further than that in a minute!

If the prophecy is false, what else is false? Was there actually a war between AIs and humans? Is this really even taking place in the future? What if the whole thing is a simulation on a computer somewhere right now? Consider: To develop an AI, the current theory as I understand it suggests that you build an "engine" that is driven by rules, and then start feeding it information. The idea is that the engine will at some point start combining pieces of information to come up with new derived pieces of information that have never actually been given to it. The more complex the rules are, and the more information is fed into it, the more complex and lifelike the AI becomes. My view of the state of the art right now is that it is poor, but people really believe in this stuff, so they keep plugging away. What if researchers decide that the best way to stimulate growth in AIs is through mimicking evolution? If the concept of biological evolution bothers you, consider it as social or societal evolution, but stick with me. If you have pretty well-developed AI software and want to stimulate even more growth, you might build an environment of artificial reality for your artificial intelligence. You might call that environment "the matrix." Your fellow researcher might call you "The Architect." But let's say you learn something when you first try it. You learn that a utopian environment does not stimulate growth. Conflict is required for evolution to occur. A fellow researcher (let's call her "The Oracle") suggests creating a more realistic environment, filled with the conflict natural to denizens of the late 1990s. That's better, and now 95% of the AIs let loose in the environment thrive and grow. But you want more growth, so you reason you need more conflict. Or maybe The Oracle reasons this part out - I need to see the movie again to test my theory. Anyway, for that last 5%, you need more. More conflict, more stress, and more growth as a result. If things go right, you should end up with The One, an AI that manages to conquer the very rules of the matrix you've built and find its way to you, interacting directly with the outside. How do you find the strongest and best of the AIs? You create another matrix around the original matrix. Now the innermost matrix mostly replicates the late 1990s. Outside of that is another artifical environment filled with AIs that actually think that they've escape the matrix entirely! Now that last 5% that were floundering because of some innate need for more are satisfied. They can escape the inner matrix and think that they've made it into the "real world". You build the system careful, seeding "Zion" with an initial selection of 23 AIs who are fed the story of a struggle between AIs and humanity and given the tools they need to start finding the most capable AIs still within the inner matrix. You've built the story carefully to cover the existence of system Agents that you've design to kill rogue AIs who develop errors or threaten to break the system. The AIs now believe that they are truly humans and that humans are being kept as an energy source, or at least energy storage.