yet another LLM

First, if you didn't get the joke, you shouldn't be trusted to develop, build or—worst of all—invest in AI.

But if you do get it, yet still use LLMs for tasks they aren't built for, like classification, prediction or generation—of language, even—based on structured data; please tell me your LLM implementation is just a proof of concept.
Please tell me you have plans to build a model specific to your input once you've validated your idea.

This is not against LLMs—LLMs are wonderful—this is against using your car's exhaust as a leaf blower.

This is kinda only half serious (kinda 👀) so please don't get mad at the person who sent you here. That said, if you were sent here, you're likely a stakeholder in one or more LLM-powered products that could perhaps use a specific model instead, so please pause and reflect on whether your application deserves one or not. Thank you.

Inspired by the wonderful nohello.net.