Ancestors

Toot

Written by Joseph A di Paolantonio on 2024-12-15 at 04:55

I am wary of the output of any LLM/GPT, and fairly disgusted by the human costs in training the LLM, especially weeding out toxic material, and by the environmental costs of running these models. Far better, I think, will be causal models using connected data, running at the Edge with personalized data. Apple Intelligence, mostly running on-device with few running on their servers that are powered by renewable energy are along the right path #SensAE

From: @czds

https://mastodon.social/@czds/113655026726119469

=> More informations about this toot | More toots from jadp@mastodon.social

Descendants