Good LLM use cases

People and organizations are spending a lot of resources right now trying to apply LLM’s to solve problems. A lot of this effort looks more like “a solution searching for a problem.”

At the end of the day a LLM is just a very powerful ____(best next word guesser)__).

In my personal experience of tinkering with LLM’s, I want to lay out what I see as the most valuable use cases.

• Translation:

  • Translating pseudo code into syntactically correct code (probably the most economically valuable use case)

  • Translating between languages

  • Translating jargon into lay terms

  • Text summarization (in other words translating something lengthy into something short while retaining meaning or key points)

• Search and discovery:

  • Semantic search (taking a sometimes vague, incomplete, imprecise query and returning results the user wants)

  • Brainstorming/recommendation engine (surfacing related concepts)

• Data use cases:

  • Structuring unstructured data — this primarily involves categorizing and parsing text data

  • Generating synthetic data

I think the most useful applications will take advantage of one of the above core use cases.