After digesting James Devlin’s
articulate (and still ongoing) series on building
a poker bot, an obvious near-term prediction can be made: prepare
for a future of bot on bot action. That is, as more knowledge of
building software that can play on-line poker masquerading as humans
spreads, inevitably the ratio of bots to humans will increase. Far from
causing the on-line poker industry to collapse, it will thrive under
such conditions, gaining more and more “players”, even as the proportion
of actual humans in a virtual seat shrinks. An average of
humans-per-table may never hit zero, but it could become vanishingly
small. Once it gets lower than 1.0, most on-line poker games will be
bots vs. other bots.
A good number of these bots, possibly large subsets of them, will be colluding. Since this requires communication between them in some way, what you will have here are distributed expert systems, programmed for conflict. In short, you’re halfway to Skynet. This, I think, would be a much better story for the Terminator franchise: where Skynet arises out of an intense escalation of poker-bot conflict, and the extermination of the human race is merely the side effect of Skynet’s (successful, though misguided) obliteration of all opposing poker bots.
Devlin’s articles also paint a picture of how disturbingly easy it is to hijack software on the Windows platform. There’s some Skynet potential there as well. For example, it is easy to see how a zombie network could easily become a poker playing powerhouse. Chances are one or more of them already is.
Image from a lego model by robbed.