Apparently, when Superman Lives was in production, the producer (Jon Peters) would bring down young kids to look at the concept art on the walls - presumably to judge toy opportunities.
Said producer also reportedly forced a member of the art crew to let him wrestle him to the ground, insisted on a giant mechanical spider (which would later turn up in the Wild Wild West movie), and similarly wanted a crystal skull in the movie after seeing one on a cover of Time.
(NB: source from memory reading The Greatest Science Fiction Films Never Made, 2nd edition, David Hughes)
Because it's evidence of (ironically) a political climate that is poisonous to free speech. Opting out of the last episode may not itself be directly censoring the show, true.
But the reasoning of the decision - making an editorial choice to intentionally omit a particular view, and one that is actually most salient to the show as a whole, as a consequence of business and political lobbying* - surely raises very worrying issues; at which point is it seen to be acceptable to perform an almost stalinist revision of media? If it is ok to omit entire episodes, does it then become acceptable to excise portions of individual shows, or of particular individuals?
*NB: I don't mean direct lobbying to remove it. I mean lobbying that has created this atmosphere of voluntarily censoring anything that might be deemed 'controversial'.
That's true...except there's no evidence atall in the article to support any assertion that wheat prices will rise to such an unsustainable level as to cause 'riots'. In fact, the headline is deeply disappointing hyperbole that wouldn't look of place in the gutter press - precisely the sort of misrepresentation a science news story should NEVER make.
@cadrina: I watched that trailer in the highest possible def format, and I still have no idea what the hell is going on beyond chrome-thing-move-very-fast-ooohblurryblur-BOOMthingexplores.
@Kent: AI could be argued as being more of a state-space search than a series of math checks, though. Yes you have to determine, in some usually arithmetic manner, the optimality of a decision or plan - but the difficulty doesn't come from that arithmetic task.
Rather it comes from the combination of modelling/representing a 'world', determining the likely impact of actions (whether deterministic or not, i.e. MDPs), and being able to evaluate a myriad of alternatives very quickly. Thinking speed, believe it or not, is still an issue; IIRC even modern chess playing AIs require a set of precompiled 'best move' libraries alongside their reasoning.
Plus, IMO, in a game perspective you don't need the AI to 'make mistakes' as such; you more need to have a rational AI such that it can be roughly predicted, and thus potentially fooled into making mistakes.