Mastodon Feed
chipotle@mstdn.social ("Watts Martin") wrote:
New post: "Just slap an LLM on the front end—how hard can it be?" https://coyotetracks.org/blog/slap-an-llm-on/
LLMs are hit-and-miss for generative output, but what about as front ends to other systems? Would that be more reliable and simpler? Apple’s “Personalized Siri” travails suggest no on both counts.