Philip did some more work on our evolutionary engine last night and called me this morning to let me know that the final integration test apparently works.

We've got some performance issues (i.e., it bogs down after a short number of generations), but those issues can be addressed. Basically, our core engine works.

So we can start evolving primitive electronic brains. Cool, huh?

Meanwhile, as if I weren't geeky enough, I'm going to start attending the meetings of this group.

Our neural net evolver is flexible enough to create nets not only for game AI, but multiple domains, from biometrics, to stock forecasting, to robot control systems.