[Question] Will superintelligent AI be immortal?

There are two points of view. Either such AI will be im­mor­tal and it will find ways to over­come the pos­si­ble end of the Uni­verse and will make an in­finite amount of com­pu­ta­tions. For ex­am­ple, Ti­pler’s Omega is im­mor­tal.

Or the su­per­in­tel­li­gent AI will die at the end, few billion years from now, and thus it will make only a finite amount of com­pu­ta­tions (this idea is be­hind Bostrom’s “as­tro­nom­i­cal waste”).

The differ­ence has im­por­tant con­se­quences for the fi­nal goals for AI and for our util­i­tar­ian calcu­la­tions. In the first case (pos­si­bil­ity of AI’s im­mor­tal­ity) the main in­stru­men­tal goal of AI is to find ways to sur­vive the end of the uni­verse.

In the sec­ond case, the goal of AI is to cre­ate as much util­ity as pos­si­ble be­fore it dies.

No comments.