Prompt Engineering: The game-changing skill you need to master in 2023!

As the technology landscape continues to evolve rapidly, it's more important than ever to keep up with the latest advancements and acquire new skills to stay competitive in the professional world.

With cutting-edge technologies such as ChatGPT, Dall E, Google's Bard, and many more, the possibilities for enhancing our personal and professional lives are endless. However, accessing the full potential of these technologies is not as simple as merely possessing them. It's like having a key to Alibaba's cave, but you must utter the right words to unlock all the knowledge and wealth.

These technologies can be likened to a parrot, capable of reciting the perfect poem for you, but only if you give the right instructions. While the parrot doesn't truly understand the meaning of the words it recites, it has listened to the entire lexicon of human language, enabling it to produce the right combination of words for the right context.

In this ever-evolving technological landscape, leveraging these new AI tools through "prompt engineering" is essential to remain competitive and unlock its full potential.

What is prompt engineering?

Prompt engineering is the process of crafting input instructions for generative AI engines, such as ChatGPT and Dall E, to generate more accurate and precise outputs. The primary purpose of prompt engineering is to help these AI engines produce desired outputs by providing them with specific and relevant input instructions.

Prompt engineering enhances the accuracy and relevance of the outputs generated by these AI engines.

Without proper prompts, the outputs generated can be very generic, often not aligned with the desired output. However, with a well-crafted prompt, the AI engine can produce outputs that closely match the intended output. This makes prompt engineering an essential skill for anyone looking to leverage the potential of generative AI engines.

Real-world use case - Let us write a cover letter for a job application

Let us assume you want to apply for the following job requirement for a product manager.

Generic prompt

The default option that most of us use is to go to ChatGPT and ask the following question: "Give me a short cover letter for the position of product manager."

Here is the interaction with ChatGPT.

As someone for whom English is a second language, ChatGPT provides an excellent starting point for the cover letter for my job application. Although the output is good, it is not entirely usable, and I will have to edit it and incorporate my experience and qualifications before sending it. It is also quite generic and lacks excitement.

A more specific prompt

Let us see the effects of providing a detailed and specific prompt to ChatGPT. I included specific details such as my qualifications, experience, and the job description in my prompt. I also explicitly instructed the AI engine not to fabricate any information.

Here is the interaction.

The cover letter generated by ChatGPT is impressive. It contains highly specific and detailed information relevant to the company and the position requirements. It is nearly usable as-is, with only minimal modifications needed.

Although the cover letter is pretty usable, it still sounds generic. Let us see if we can improve it.

Fine-tuning the prompt

As a follow-up, I asked ChatGPT to make my cover letter sound more professional and highlight the benefits of hiring me.

Do you see the difference in outputs based on the prompts?

That, ladies and gentlemen, is an example of prompt engineering.

10 tips for creating a perfect prompt

Crafting suitable prompts is a skill that can be improved with practice, much like any other skill.

To help you develop this skill, here are 10 tips for creating a perfect prompt:

1. Master English

The most critical skill to master is English since it is the primary language used in most AI models and datasets, and the only way to interact with these models is by using English.

2. Determine the purpose of your prompt

Consider what you want to achieve with your prompt and what response you hope to generate.

3. Identify the target audience

Think about whom you are addressing with your prompt and the tone, style, and level of language that would be most appropriate.

4. Be specific and avoid ambiguity

Provide as much specific information as possible to help the AI engine understand the context of the request. Ensure that your prompt is clear and unambiguous to avoid confusion or irrelevant responses.

5. Use natural language

Write your prompt in natural language to improve the AI engine's understanding of the context and intent behind the request.

6. Consider the length of the prompt

Keep your prompt concise and to the point without compromising on important details.

7. Be mindful of the limitations of the AI engine

Understand the limitations of ChatGPT and the training data it uses to avoid unreasonable expectations.

8. Use appropriate keywords and phrases

Use keywords and phrases that are relevant to the request and will help the AI engine provide an accurate response.

9. Proofread your prompt

Ensure that your prompt is free of spelling and grammar errors that could affect the AI engine's understanding of the request. This is where your knowledge of English comes in. You can also use tools like Grammarly etc to correct your English.

10. Give follow-up instructions to refine your prompt

ChatGPT remembers the conversation (to a limit), so you can give it follow-up instructions to refine the output.

Resources for crafting suitable prompts

Here are some tools and online resources to help you craft a suitable prompt.

1. A community-driven collection of prompts for various use cases

https://huggingface.co/datasets/fka/awesome-chatgpt-prompts

2. An app that uses a BART model trained on the above data. In this app, you give a persona, and it will suggest a prompt for you https://huggingface.co/spaces/merve/ChatGPT-prompt-generator

3. Some tips to create prompts for AI art generators like Dall E, Mid Journey, Night Cafe, and so on https://pub.towardsai.net/design-your-ai-art-generator-prompt-using-chatgpt-7a3dfddf6f76

4. There is also a marketplace where you can buy and sell prompts for various use cases https://promptbase.com/

Conclusion

In conclusion, prompt engineering has become a crucial skill for developers and managers who want to leverage the full potential of all the new AI-powered products being launched nowadays.

Crafting a suitable prompt requires understanding the AI engines, their training data, and the user's needs.

Effective prompts are specific and unambiguous and use natural English to avoid producing generic responses.

As AI-powered products become more pervasive in the professional world, the ability to write good prompts is becoming a must-have skill to stay competitive.

By mastering prompt engineering, professionals can gain a competitive edge in their industries and contribute to the ongoing evolution of AI technology.

About Gunjan Karun

Gunjan is the founder of Quadralyst, a product development company specializing in rapid MVP and product development for startups.

He has been building apps based products for startups and enterprises for the last two decades utilizing new and emerging technologies such as Natural Language Processing, GPS, Blockchain, Artificial Intelligence, IoT, Cloud computing, and various Mobile technology to meet the needs of the clients.

Please contact me if you would like to discuss your product ideas.