Skip to main content

AI School: How to Use Chat GPT

Chat GPT changed the conversation about artificial intelligence - the technology that is predicted to revolutionize how businesses and individuals interact with computers.

Despite its impressive potential, the service is far from user-friendly in all aspects.

In a series of articles, Techsavvyminds tests and guides you, the reader, through the basics of the most talked-about AI services. First up is Chat GPT from the American company Open AI.

Over half a year has passed since Chat GPT transformed the conversation about artificial intelligence. For companies, it has been said that AI can streamline everyday tasks by taking over repetitive tasks, assisting with presentation materials, and even handling email conversations.

Although the hype has been hard to miss, it hasn't been obvious to everyone to explore the possibilities of this new technology. Others have tried and realized that the shortcomings are still too significant to make a real difference in everyday life.

The only way to know for sure is to try - and probably continue to try as progress is made.

Chat GPT is a generative AI that has learned through machine learning. It creates new content on demand, based on learned data and the relationships present in the information. This is different from simpler algorithms that merely copy and repeat pre-programmed information. Chat GPT is a so-called Large Language Model (LLM). 


1. Getting Started

If you haven't already created an account and ventured into the complex world of Chat GPT, you must start by visiting chat.openai.com and creating a free account. That part is relatively simple.

A first encounter with Chat GPT is not unlike the first time you found yourself in front of a search engine. And just like then, it's up to you to take the first step. It can be a question or a simple request. Only your imagination sets the limits.

Giving Chat GPT a task is called providing a prompt. In short, it means asking the AI to do something, just as you would ask a person to find information - you use your everyday language. Chat GPT understands most languages but is primarily trained in English content and therefore performs better in that language.

As a first test, you can, for example, ask the AI to suggest a solution to a simple everyday problem you're currently facing.

When you ask a similar question in a search engine, you also get an answer based on pre-existing information on the internet. On the other hand, Chat GPT generates a response based on its accumulated knowledge. It doesn't present existing material to you as it is. Therefore, the answers can vary depending on how you phrase the question and the details and nuances you add.


2. Advanced Prompting

One way to get more refined results from Chat GPT is to demand more in your prompt. A common approach is to provide a "system prompt" before your actual task. A system prompt sets the framework for Chat GPT, such as explaining how it should respond.

For example, "you are a poet" is a system prompt that sets the context for Chat GPT. Then you can input the task, such as "Write a birthday card for my 10-year-old son when I give him the bicycle."



With advanced prompting, Chat GPT can assist in creating personalized and meaningful messages, allowing you to express your sentiments and make special moments even more memorable.


3. Limitations

Chat GPT has been trained on data up until September 2021 and, therefore, cannot provide answers on current events.

The exact details of Chat GPT's training are not known, but it is far from all-knowing. This particularly applies to niche events and individuals.

Regarding language, Chat GPT writes more or less flawlessly. It is advantageous to ask it for writing tasks. However, since it doesn't copy content from its training data but summarizes information based on its accumulated knowledge, factual details can be mixed up if the text you ask it to write is about a narrow subject where it has had limited sources to train on. This phenomenon has come to be known as AI hallucination.

It can also make factual errors. For example, if you ask Chat GPT for a recipe like "give me a recipe for Swedish meatballs" you may notice that it often comes up with taste combinations, but the quantities of each ingredient may be incorrect.




4. Keep Testing and Exploring 

Translating written text is something Chat GPT handles well. It doesn't translate word-for-word but rather sentence-by-sentence to better convey the content. The same can be said for programming, where it has shown the ability to solve coding problems at a relatively high level. Summarizing longer texts (copying a long text and asking it for a summary), proofreading, and providing feedback on the content are also worth trying to simplify your daily life.

By exploring the capabilities and limitations of Chat GPT, we can begin to grasp its potential impact on our lives. While it has revolutionized the conversation around artificial intelligence, there is still much to uncover and refine. As we continue to test and push the boundaries of this technology, we must approach it with curiosity, adaptability, and a critical eye. So, embrace the AI revolution and join us on this exciting journey into the world of Chat GPT and beyond.

Techsavvyminds, empowers you to navigate the realm of AI with confidence and creativity.

Comments

Popular posts from this blog

5 Proven Strategies to Pass the Microsoft Power BI Data Analyst - PL-300 Exam

Earning a certification in Power BI as a data analyst is a great way to validate your skills, enhance your career prospects, improve your skills, enhance your credibility, and demonstrate your commitment to professional development. To excel in this exam, candidates must have a strong grasp of Power Query and proficiency in writing Data Analysis Expressions (DAX). They should also possess knowledge in assessing data quality and be familiar with data security measures such as row-level security and data sensitivity.  The following skills are evaluated:  Prepare the data (25–30%) Model the data (25–30%) Visualize and analyze the data (25–30%) Deploy and maintain assets (15–20%) The Microsoft PL-300 exam is designed for candidates who want to validate their skills as Data Analysts. Here are some tips to help you prepare for and pass the PL-300 exam: 1. Review the exam objectives:  The first step in preparing for any exam is to review the exam objectives. These objectives pro...

How to append queries in Power BI

To append queries in Power Query, you can use the "Append" transformation, which allows you to combine two or more tables by adding the rows from one table to the bottom of another table. Here is how you can do this in Power Query: 1. Open the Power Query Editor and select the tables that you want to append. 2. Click the "Home" tab in the ribbon, and then click the "Append" button in the "Combine" group. 3. In the "Append" dialog box, select the table that you want to append to the bottom of the other table, and then click "OK". Power Query will create a new query that combines the two tables by appending the rows from one table to the bottom of the other. You can then apply additional transformations as needed, and load the resulting table back into your workbook or report. Alternatively, you can also use the "Merge" transformation to combine two tables by matching rows from one table with rows from the other table ...