3 ms·
Fair. My jury is still out as to whether the current models are proto-AI. Obviously an incredible innovation. I'm just not certain they have the potential to g
by LightBug1 1y ago
Fair.
My jury is still out as to whether the current models are proto-AI. Obviously an incredible innovation. I'm just not certain they have the potential to go the whole way.
/layman disclaimer
- falcor84 1y agoAs you say, whether we call it "AI", or "doohickey", it is an incredible innovation. And I don't think that anyone is claiming at the moment that the systems as-is will themselves "go the whole way" - it is a technological advancement, that like all others should inspire practitioners to develop better future systems, that adapt some aspects of it. Perhaps at some point we will see a self-propelling technological singularity with the AI developing its own successor autonomously, but that's clearly not the current situation.
- LightBug1 1y agoDoohickey is so much more relatable ... I may call LLM's that from now on. Thank you.
- bluefirebrand 1y ago> And I don't think that anyone is claiming at the moment that the systems as-is will themselves "go the whole way" Dunno but I see plenty of people making exactly this claim every day, even on this site
- kmoser 1y agoThat will never happen. We may approach that state asymptotically but since AI output is stochastic, and humans' goals change over time, humans will always be part of the loop.
- falcor84 1y agoWhatever the formula for the probability of recursive self-improvement of AI may be, I am unfortunately certain that the fickleness of human goals does not factor into it.