3 ms·
> meant to distinguish AI systems from human intelligence > but then when somebody builds a system I mean this is really it. You still have to have a human to
by seiferteric 4y ago
> meant to distinguish AI systems from human intelligence
> but then when somebody builds a system
I mean this is really it. You still have to have a human to build these systems that specialist in one thing. Once you create a system that can automatically create those systems and it doesn't need humans anymore to solve novel problems, then there will be no practical difference in kind between human and AI intelligence.
- polygamous_bat 4y ago> Once you create a system that can automatically create those systems... Except we don't have that. We don't have one human that can create this system by themselves. We have a choice group of a handful of smart, motivated, and quite generously compensated humans working on these problems to create such system. As such, you are already surpassing the "general" intelligence level by quite a lot.
- seiferteric 4y agoNo, you misunderstand, the "systems" I am talking about are the ones built into our minds, like recognizing faces, or understanding speech. Humans can learn to speak and recognize each other automatically, but AI systems have to be built specifically to do each task.
- polygamous_bat 4y ago> Humans can learn to speak and recognize each other automatically, but AI systems have to be built specifically to do each task. I think that is a very generous take on what we do "automatically". After all, we have millions of years of evolution to build out all the neural circuitry that helps us speech or vision -- it's not like you can throw a soup of genes on the ground and out comes intelligence. What is machine learning doing, if not selecting, out of many possible parametrizations, the ones that are suited to understand vision or speech?