Prompt Engineering Vs Context


Prompt Engineering Vs Context

Are prompt engineers actually needed?

There's a group of people on the internet who like to call themselves "Prompt Engineers." They see themselves as people who write good prompts for LLMs or other machine learning models. I have a feeling this job will likely go extinct, as it was never really valuable in the first place. With that said, just typing one sentence into an LLM isn't likely to yield the best results. Giving the AI good context is super important. That's why things like the “Canvas” features most major LLMs have, or “projects”/”workspaces” where the AI can get access to more information for context so it can understand the world around it, are extremely important.

So you don't need to hire a "Prompt Engineer". What you need to do is feed these LLMs and other AI tools good information as context. Perhaps give them access to your Standard Operating Procedures (SOPs). Give them background on the project. Give them the same orientation that you give a new hire over the first 90 days of their career, and then see what they can do. Context is key. Prompt Engineers, not so much.