After playing around with ChatGPT, Dall-E, Bard and Bing Chat, I want to add another skill to focus on — "Clear and concise English communication." No matter how well you have thought about the problem in your head, if you are not able to articulate it with the right vocabulary, you might continue to get frustrated with the results.
Think about any recent incident in your office where you thought you were extremely clear in what you asked for, only to realise later that the other person created a completely different output.
But Won't AI Figure It Out?
You can argue:
- 1.LLMs can easily work around basic grammar mistakes
- 2.LLMs can do a basic translation job
- 3.LLMs can guess what you meant easily
These are all true (especially with general purpose language models) and you will easily get away with low-value use cases like text summarization and drafting a mail.
Where It Matters
If you are searching for specific types of outcomes (high-value add use cases), you will require sharp and clear communication skills along with basic prompt engineering. The models will evolve and in the near future this skill may not remain a differentiator. But as of today, if you want to create a productivity edge, this is a critical lever.