Prompting - Not the New Googling

I had a reflection the other day. If you've used chatbots like ChatGPT, you probably know that generative AI models don't work like Google. What do I mean by that?

Well, I've noticed that when I and others start using AI chatbots, it's easy to reflexively type a Google search query into the chat bar. The field looks like a Google search field, and we have learned to use Google effectively. But if you use Google-like queries with, for example, ChatGPT, you are often disappointed, as more often than not, you get a vague or sometimes even incorrect answer.

As John Schulman, one of the creators of ChatGPT, pointed out, "Our biggest concern was around factuality because the model likes to fabricate things." Chatbots have a tendency to provide convincing but false information, to the extent that the creators refer to it as "hallucinating."

An example I can provide is from when I held an internal AI course at Aliby. I created an exercise to demonstrate this. The participants were given the assignment to find Europe's highest mountain. They received two different answers from ChatGPT and Bing Chat. A lively discussion ensued!

We found that the solution may be that instead of using AI chatbots like Google, we can take advantage of their strengths by using the right "prompt." This is something that will become increasingly important.

Those who master the art of "prompting" or "prompt engineering" will be the most productive in their workplaces. There are even job ads where they are looking for Prompt Engineers.

So, in this new world of generative AI models, you can be the most efficient one at your job and be one of those who finds the right approach and opens the doors to an exciting and productive future.

PS And for those wondering, Mont Blanc is apparently not Europe's highest mountain; it is Mount Elbrus in Kazakhstan. What did I know? You never stop learning. ?? DS

#aliby #ai #prompts #promptengineering #persensai