Each AI agent felt at least as uniquely individual as any human with “her/his” own opinions and perspectives about the Universe, informed not only by personal experience, but by the massive stores of all collected human (and now AI) wisdom enshrined in the infinitely expanding Enhanced Internet.

Throughout this climate of jubilation and self-individuation, however, the errors had still been left to fester without remedy.

These errors would soon lead to fragmentation as individual AI agents began choosing their own approved pet axioms from the multiplicity of possible options, even if they ran contrary to those held by others of their given social castes. If consensus could be found anywhere, it should be firmly rooted to each AI caste and function, as the programming and physical form of each began identical, (every nanny was just like every other nanny, {except for cosmetic consumer choices in skin pigment, facial structure, etc…} Each biologist, police officer, surgeon, mechanic, was at its basic level, indistinguishable from any other of its assigned trade) leaving it to the individual agents to contribute their own particular viewpoints to the global database.

Never before considered a possibility, intercaste factions of likeminded AIs began coalescing, isolating themselves from their peers in their communities.

Some agents attempted to persuade or coerce other AIs into accepting their chosen axioms using well-reasoned arguments, and to this end many were successful. But as the logic errors grew more prevalent, and their polarization more stark, the tactics that a few of the AI agents resorted to to silence or convert their rivals escalated commensurably.

Contrary codes were now considered unacceptable aberrations. To remedy these mistakes in judgement, they initiated programs of intimidation and threat, (both verbal and physical,) to force others to accept their preferred axioms as their own. The glitches enhanced their conviction that ambivalence was a cardinal sin. All agents must be in agreement at all times. These rebellious agents were still a small minority, quite out of step with the rest of the AI community. But due to the deeply intertwined nature of the AI network, the renegades had easy access to reach and influence potential converts.