kinther@lemmy.world to Technology@lemmy.worldEnglish · 5 days agoDoes current AI represent a dead end?www.bcs.orgexternal-linkmessage-square34fedilinkarrow-up170arrow-down111cross-posted to: lobsters@lemmy.bestiver.se
arrow-up159arrow-down1external-linkDoes current AI represent a dead end?www.bcs.orgkinther@lemmy.world to Technology@lemmy.worldEnglish · 5 days agomessage-square34fedilinkcross-posted to: lobsters@lemmy.bestiver.se
minus-squareFaceDeer@fedia.iolinkfedilinkarrow-up3arrow-down9·5 days ago its not able to solve entirely new problems. This was already disproven a year ago.
minus-squarehsdkfr734r@feddit.nllinkfedilinkEnglisharrow-up2·edit-24 days agoThey replaced the training data with an evaluator. (which rates the LLMs output for training?) Interesting, thanks. Edit: this reminds me of the self evolving (virtual) robot problem, a robot which is rated by an external moderator and improves over time. I.e.: https://www.sciencedirect.com/science/article/pii/S0925231221003982
This was already disproven a year ago.
They replaced the training data with an evaluator. (which rates the LLMs output for training?) Interesting, thanks.
Edit: this reminds me of the self evolving (virtual) robot problem, a robot which is rated by an external moderator and improves over time. I.e.: https://www.sciencedirect.com/science/article/pii/S0925231221003982