Once Upon AI Time

Since AI is actually a hot topic everywhere nowadays, I would like to use the opportunity and spread the word about the course ChatGPT Prompt Engineering for Developers by OpenAI and DeepLearning.AI which I have just completed.

The course is available under the link:

https://learn.deeplearning.ai/courses/chatgpt-prompt-eng

It is a very practical course, based on prompting and coding in Python, using the guts of ChatGPT 3.5.

But even if it is crafted for Developers, I think it may be really beneficial for broader audience.

What does it contain?

• Prompting good practices - how to be clear and specific, and how to give a model time to think and provide more accurate results.
• Model limitations - do you want to know about hallucinations and prevent your model from confabulating?
• Iterative prompt development - because if your prompt generates irrational results, it is not the model that fails, it is your prompt.
• Capabilities like Summarizing, Inferring, Transforming, Expanding - methods which can help you speed up and polish the results of plenty of activities you perform on your computer on a daily basis.
• Temperature configuration and setting up an AI bot - would you like to control the behaviour of AI and heat up or cool down, hmm?

What sets this course apart from similar ones?

This course is provided and led by OpenAI employees (while actually current revolution started with OpenAI moves). And it is free.

Is being a Developer a must in this case?

Actually you do not have to know anything about coding. You can be a basic user of ChatGPT without even an engineering background to learn some from this course about efficient prompting. But to really make use of the content, knowing basic data types and operating on variables is recommended. The deeper you understand coding and algorithm logic, the more you will understand and learn.

Why would I need a course if I use ChatGPT for prompting and my results are just fine?

Well, prompting is quite resistant to instruction inaccuracies and in many cases you will succeed. But first reason to take a course is that you may not even know what you do not know, and how you can use AI in various use cases you have not yet thought about. And the second reason is, and it is maybe even more important, that things do get complicated if you fail to achieve reasonable results in prompting with your casual approach and reasoning. Or if your results are just crappy, inaccurate and you are asking yourself why. If you do not really understand and feel the mechanics behind prompting and operating on LLMs, likely you will be unable to modify your approach to be successful and you will not meet your objectives.

And if you know Python or any other relevant programming language, you can easily go beyond the scope of this course and apply newly discovered ideas in your project or a business.

* Still no AI was used to prepare or validate this article ;-)