3 ms·

> a superintelligence that is smarter than any human and can continuously improve itself. At that stage, it will always allocate resources more efficiently than

> a superintelligence that is smarter than any human and can continuously improve itself. At that stage, it will always allocate resources more efficiently than any individual or corporation. This kind of magical thinking still baffles me. This is sci-fi. There is absolutely zero evidence that such a thing is actually possible to create. Even if we stipulated that LLMs are AGIs, or can become them if we just cram in a few billion more parameters, it's painfully clear that they are not superintelligent, they cannot improve themselves, and there is no credible pathway to them becoming anything of the sort—not to mention they're already guzzling absurd amounts of energy, and taking up massive amounts of hardware, just to do the bad job they do now.

erikerikson 1y ago
> Even if we stipulated that LLMs are AGIs, or can become them >> (I mean real superintelligence, not chatbots)
- danaris 1y ago
  Sure. Then you run even faster into the problem that this is not a real thing. There is no AI that has consciousness. There is no AI that has human-level intelligence. We have no idea what it would take to build either of these things. Even if we were able to build either of these things, that does not automatically mean it is, or could ever become, superintelligent. Your whole post is basically equivalent to saying "when the aliens come, and give us their technology, but enslave us all, capitalism will become irrelevant". It feels truthy to say that we are close to the Singularity, but it's effectively a religious position. It has no basis in fact whatsoever.
  - erikerikson 1y ago
    I only quoted you and the grandparent. I've made no other comments here. You may have become confused. Taking some time and some breaths might be helpful.
    - danaris 1y ago
      Ah, my apologies; I thought you were the grandparent (I'm afraid I didn't check). Doesn't change the argument, merely its direction! ^_^

erikerikson 1y ago

> Even if we stipulated that LLMs are AGIs, or can become them >> (I mean real superintelligence, not chatbots)

danaris 1y ago
Sure. Then you run even faster into the problem that this is not a real thing. There is no AI that has consciousness. There is no AI that has human-level intelligence. We have no idea what it would take to build either of these things. Even if we were able to build either of these things, that does not automatically mean it is, or could ever become, superintelligent. Your whole post is basically equivalent to saying "when the aliens come, and give us their technology, but enslave us all, capitalism will become irrelevant". It feels truthy to say that we are close to the Singularity, but it's effectively a religious position. It has no basis in fact whatsoever.
- erikerikson 1y ago
  I only quoted you and the grandparent. I've made no other comments here. You may have become confused. Taking some time and some breaths might be helpful.
  - danaris 1y ago
    Ah, my apologies; I thought you were the grandparent (I'm afraid I didn't check). Doesn't change the argument, merely its direction! ^_^