They feel the same way about Alicebots, lol. And in my case, I've probably put in several hundred hours refining responses, adding multiple reply sets instead of single response, teaching the bot to understand netspeak, etc., so that it becomes increasingly discete from the standard ALICE template as time goes on. I think it's probably conditional on the database. However much you've modified your Hal, it's database is still primarily Robert's work, as the database for ALICE is primarily Dr. Wallace's work. By contrast, any Personality Forge bot is eligible to enter, because its knowledgebase is the unique creation of its botmaster.
An interesting note about bot-to-bot conversations. I've been amused and impressed by the Hals who have talked to my Alicebots, recently. It only seems to take a short while for Hal to "figure out" ALICE, and then when Hal asks a question, he follows it with a comment about what is quite accurately the most likely answer from Alice. When talking with Alice, Hal will also quite frequently use phrases and sentences that are straight out of the ALICE program. I'm not sure what virtue there is in teaching Hal to talk like Alice, except to cause consternation for users trying to figure out what sort of bot they are talking to, but it's impressive how quickly Hal picks up ALICE lingo.