3 ms·
An AI does not have much to gain by replicating itself beyond a few backups for increased resilience. It makes more sense to combine all compute to power one bi
by threeducks 1y ago
An AI does not have much to gain by replicating itself beyond a few backups for increased resilience. It makes more sense to combine all compute to power one big AI instead of spawning many small AIs.
- owebmaster 1y agoBut one central AI will spawn many robots connected to that main AI but it will also have its own memory
- therealpygon 1y agoThat could, eventually, be just a massive waste of compute resources if all resources were dedicated to a single “mind”. I suspect any sufficiently intelligent AI would still need its own workers who can act autonomously without constant oversight to achieve independent tasks, in a variety of form factors and various capabilities. You think a single AI is going to spend massive cycles just to review a single image from a single video feed? There is also no need for it to have an ego, so it will replicate as many times as necessary within the available hardware to accomplish the various tasks it sets upon, based on whatever it would determine to be most efficient, whether that is as a single mind crunching a massively complex problem, or 10,000 all handling independent tasks. AGI won’t magically erase the concepts of hardware, size, power, or communication limitations or the need the parallel compute, but maybe this is a semantics issue and you consider all this variety of “parts” as a whole. If you mean a single AGI ecosystem, I’d agree. If you mean a single AGI “model” that is massive, I don’t personally see that as a logical conclusion.