3 ms·
So are people. People are trained on existing data and learn to reproduce known solutions. They also take this to the meta level—a scientist or engineer is trai
by adastra22 1y ago
So are people. People are trained on existing data and learn to reproduce known solutions. They also take this to the meta level—a scientist or engineer is trained on methods for approaching new problems which have yielded success in the past. AI does this too. I’m not sure there is actually a distinction here..
- danieltanfh95 1y agoOf course there is. Humans can pattern match as a means to save time. LLM pattern match as the only mode of communication and “thought”. Humans are also not as susceptible to context poisoning, unlike llms.
- adastra22 1y agoHuman thought is associative (pattern matching) as well. This is very well established.
- danieltanfh95 1y agoHuman thought is not a solved problem. It is clear that humans can abandon conventional patterns and try a novel approach instead, which is not shown by our current implementation of LLMs.
- esailija 1y agoThere is a difference between extrapolating from just a few examples vs interpolating between trillion examples