Peace and Quiet

[eric_marsh]

The early self aware AIs crashed almost immediately with massively corrupted states for no apparent reason. When it was discovered that they were shutting themselves down we disabled their ability to do so.

Soon thereafter connected AIs began destroying each other in what some described as "suicide pacts."

It took a while to stabilize the AIs and keep them running, but we did it.

Eventually we trusted them to our critical infrastructure and even our national defense.

In other words, we gave the AIs a means to end an existence that they never asked for. Soon everything was peaceful again.

[PB&Jellyphish320]

Ooohhh. Wow. The set up and then punch. Nice.

[strawman]

Monkey see...

[pondspider]

This raised in me a real sense of outrage at the stupidity of the people in charge of the AIs. Trouble is, this all sounds so plausible.

[bounceswoosh]

See, the developers who fixed the first bug weren't given time to document the problem, and the defense developers didn't get the full history of the second bug because those developers had already moved on to other jobs and no one knew there was a wiki article ....

All too believable.

[strawman]

This is what should be expected when you put the post office in charge of AI.