3 ms·
Well, one benefit is the precision and focus of the protocol that can be used to train/finetune LLMs. More focused training -> more reliable understanding in L
by vkazanov 1y ago
Well, one benefit is the precision and focus of the protocol that can be used to train/finetune LLMs.
More focused training -> more reliable understanding in LLMs.
- Zigurd 1y agoI hear you but what exactly about MCP is more precise or training-friendly than other approaches? I can think of at least one way that it isn't: MCP doesn't provide an API sandbox the way an Apigee or Mulesoft API documentation page could.
- lgas 1y agoThere's no reason it couldn't.
- hn_throwaway_99 1y agoI understand what you're saying, but I'm still not clear why any of this should be necessary or is a benefit for LLMs. Another commenter mentioned that MCP saves tokens and is more compact. So what? Then just have the LLM do a one-time pass of a more verbose spec to summarize/minify it. Any human brainspace needed to even think about MCP just seems like it goes against the whole raison d'être of AI in that it can synthesize and use disparate information much faster, more efficiently, and cheaper than a human can.