Hacker News Reader

Comment by pdog | Hacker News Reader

What does a rapidly improving AI have to gain by "merging" with our notoriously error-prone, finicky biological hardware? It's nearly certain that intelligent machines will choose to replace humans, not enhance them.

Our "successful" descendants will probably be the unmerged, low-tech survivors living on the outskirts of a new "Machine Age" civilization.

View on HN

Until the day the machines decide that the rust problem must be tackeld at its root and that atmosperic oxygen must be eliminated...

    L'ennui c'est l'oxygène
    qui fait rouiller nos roues
    dentées, c'est l'oxygène
    qui fait rouiller nos clous

I believe the merging Sam is referring to here is augmenting human intelligence with non-intelligent (but better) machines. Kurzweil talks about this as well, and was one of the original options in Vinge's "Singularity" (https://edoras.sdsu.edu/~vinge/misc/singularity.html)

> Unless we destroy ourselves first, superhuman AI is going to happen... Perhaps the AI will feel the same way [that intelligence is singular] and note that differences between us and bonobos are barely worth discussing.

He's referring to a superhuman artificial intelligence, not merely an augmented human intelligence.

Maybe by the time AI is that sophisticated it will also be just as error prone and finicky as "BI" (biological intelligence)

Reply on news.ycombinator.com