Don't Use ChatGPT Until You Watch This Video

You've probably already started using ChatGPT, right? It is not an exact science. Anyone can ask questions and it will give you answers. In school we learned that there are no bad questions. In the ChatGPT world, there are. It won't tell you that, but it will give you poor quality answers and you'll waste a lot of time going back and forth. I was there, very upset, so I decided to spend some time finding the best pointers to get the most out of ChatGPT. My first piece of advice for you is to take advantage of the personalized instructions that you will find when you click on your profile icon.

So when you access custom instructions, you have the ability to provide ChatGPT with a note that explains who you are, what you want, and how you want ChatGPT to respond. This way, you won't have to repeat your preferences in every conversation. So the first question is: what would you like ChatGPT to know about you to provide better answers? It may be helpful to provide some context about yourself. So, for example, if you're a teacher, tell them. If you are a student, accountant or lawyer, tell them. This way, you'll be able to get answers that are more relevant to your line of work.

If you want answers that are relevant to your region, please indicate where you live. Not the exact address, but just the area where you are. For example, you could say "I'm a software developer based in Chicago and I work with Python." Or you could say, "I'm a marketing professional in New York and I write advertising copy." Well, you get the idea. The next question is: how would you like ChatGPT to respond? We come up with a few ideas to get started, like: how formal or informal should ChatGPT be? For example, you could write, "The language and tone should be friendly and informal." You can decide whether ChatGPT can have opinions on topics or remain neutral, and whether responses should be long or short.

And

this

can really make a difference because you can get rid of a lot of frustrating trade-offs. So, for example, let's say you're good at Excel and use ChatGPT to help you when you get stuck, but you don't like seeing all those explanations about solutions. So what you could write is: "When it asks you for Excel formulas, simply provide the most efficient formula without any explanation." If you were a programmer and didn't want to see all those programming explanations, you could write "When I ask you for code, just give me the most efficient code with code snippets without further explanation." Well, that's really helpful.   In

this

case, I'm picky about Excel formulas.  I'm going to leave this and click Save.

Now, let's start a new chat and I'll ask you for a formula to make sure this works. Let's say I'm in Excel and I'm having difficulty updating this formula. Currently, it returns everything that is greater than 12,000, but I want to change this to be between 12,000 and 15,000. And I have no idea how to tell it to do so between these values. I'm going to go back to ChatGPT and write my formula, tell it to update this Excel formula so that it returns values ​​in column B that are between these two.   And when I send this it just provides the solution without any explanation.

Now, without that custom setting, this is what you would get. And sometimes you would end up with much longer explanations like this. They're great if you don't know how to use them, but if you do, they can be quite annoying. Another thing that could be quite useful is to tell ChatGPT to always tell us the confidence level of your response. This could be very useful for factual topics. Now, we can also expand on this and say, "When your answer includes facts, always provide a valid URL with the source of your answer. And if you speculate or predict something, let me know." Okay, so let's try this.

I'm going to save this, go ahead and start a new chat and ask him about the FIFA World Cup winners of the 90s. When I run this, I get: 1990: West Germany, Brazil, France.   The trust level is high and I get a list of valid URLs directly from FIFA.com. Okay,   so these settings can be really useful.  I'm sure you will find them useful. Now let's move on to the directions. The first is to write like you. So,   if you ask ChatGPT to write some text for you,   the results will probably sound a little generic,   right? Even if you emphasize the personalized instructions, tone, and style you want, it may not adequately reflect your writing style.

Now, the good news is that you can teach him to write in your own style by giving him some examples. So first, we'll explain to ChatGPT what we're going to do with this message. "I would like you to help me write articles for my productivity blog." Simply replace this with   any type of blog or article you need.  "First, I want you to understand my writing style based on the examples I give you. You will save my writing style in 'LG_STYLE'." Now, this makes it easier to refer back to later. "After that, you will ask me what the topic of my specific content is.

Then you will write the article using LG_STYLE." Well, let's try it. Okay, so you understand what we're trying to do and you're ready to look at some examples. I'm going to go get a copy of my website. Let's copy this paragraph from the About page and paste it as an example one of LG_STYLE. Now I will give you a second example of LG_STYLE. I'll just go and grab the other copy from here and paste it in. So I'll summarize my style as informative, personable, and aimed at making a connection.  And now you are ready to write our content.

So, I want to write an article on the importance of daily coffee for productivity.   Great, start writing it in my personal writing style, the one I saved earlier. Now, of course, you can continue working on this and improve it. You can also come back to this chat and ask him to write other related articles. And by the way, if you love your daily coffee, subscribe to this channel because we all love coffee around here. The next step is self-criticism.  So another great option is to ask ChatGPT to review your own text and provide feedback. It sounds funny now, but it actually works well.

So let's say I asked ChatGPT to provide me with a summary of why Python and Excel can work well together and I get this answer that I'm not very happy with. So I'm going to ask you to act as a critic. Be ruthless, analyze the text and tell me where it can be better. It will review the response and provide step-by-step feedback on potential issues. For example, with clarity and how adding specific examples could improve the summary, how we could highlight specific capabilities. He also went ahead and revised this original answer, providing examples of the strengths of Python and Excel.

From there, you can further improve this answer by asking him to specifically emphasize and include some of the previous advice he gave us. It's very good, right? The next step is self-motivation. So how about we use ChatGPT to auto-indicate and optimize your own message?   This is how you can do it. So, let's say I want to send an email to my team encouraging them to participate in our team building event. To get the perfect message, you could ask ChatGPT: "Write five perfect ChatGPT messages that really show the power of ChatGPT. Focus your messages on writing an email to my team encouraging them to participate in our team building event.

Before you write anything, ask me questions until you're sure you can create the optimal prompts." They will usually come back with a few questions about the topic, in this case, about the purpose, dates, locations, and other useful information.  Now, after I give my answers, for example, for any particular incentive or benefit, I type "nice dinner, good breakfast, no work", then I send this and five possible messages are created for me. These are indications that will help me get the best answer. One of them is: "Compose an email to your creative team encouraging them to join a special team-building event.

Mention the importance of taking a break from work." Another is: "Compose an email inviting your team to a fun, informal team-building event. Emphasize the importance of building strong relationships within the group and the opportunity for a nice dinner."   Okay, so I like number three better. I'm just going to tell you to use it. Now it just gives me a message, but I actually want it to execute it. So can I do it? Start writing the email. Now we can see that it has a nice theme, the tone is soft and casual and highlights bonds and relationships.

Well, then we can see it spread in different places, as if we can share stories, laughter and delicious food. We will have a big breakfast. There is no work, there are no deadlines.  So, it's a very personalized email. Now, compare this to a case where I don't optimize the message. So, I tell him to write an email to my team to participate in our team building event. It's no surprise that you receive a very generic email about team building and what team building events are designed to do in general. So, as you can see, automated requests can give you a much better result.

But you can also take these self-instructions one step further. So if you go back to the step where we got the different messages, we could ask you why one message would work well. You give me detailed information about why you think this message is effective, but I'm not so sure. So I'll ask, "Which message do you think is best?" It tells me that message three effectively combines the key elements I provided, so it might work well in my case, and it did. Often when you ask ChatGPT to type something, the response can be quite long. Therefore, one way to avoid this is to tell it from the beginning to stick to a specific word count.

Now, if this is something you want in every response, you can add it to your custom settings. Otherwise you can add it to your message. So, for example, we want to know the advantages of using XLOOKUP over VLOOKUP in Excel. I will add that the maximum length of the text must be 500 words. A few moments later. Okay, so let's check again. I'll copy this, go to a new page, type "word.new" to open a blank Word document in the browser, and paste the text. We can see down here that we have 469 words. Now, another alternative is to tell it to reduce the length of a text it already gave you.

So a good suggestion for this is: "Now, say the same thing more concisely and prefer to use only 60% of the words or whatever percentage you need. You could try to reduce it further, step by step, until you get the crunch." that you like. Your audience will thank you." Specify the output format. ChatGPT has many different output formats, not just plain text. So, for example, you can tell it to give you the answer in a table format. Let's say: "Create a table with the winners of the FIFA World Cups between 1990 and 2018. The headings should be year, winner and runner-up." So, we got a table.

We can simply highlight the content, use the shortcut Ctrl+C to copy it, move to an Excel sheet and paste the table with Ctrl+V.   Now, alternatively, you can tell it to output the table as CSV. Then, all you need to do is copy the code from here, paste it into a text editor, and then save it as a CSV file extension. Now if you need this, say for your website, you can output it as HTML.   Moments later... I'm going to try this. So I'm going to take this code and paste it into our site in an HTML editor.

Then when I go to the preview, I can see the table, but it doesn't look very good. I prefer to have borders and spaces. So let's improve the message and ask it to add borders and spaces. Later that night. Now when I replace the code, I have a table that is much easier to read. Good? Therefore, it can also output in different formats like JSON or XML. Or if you need it as a Pandas data frame, you can do that too. Test it. Here are some tips to get the most out of ChatGPT or any large AI language model.   In reality, it all comes down to the old principle: garbage in, garbage out.

Keep this in mind and you will definitely make better use of these new tools. I hope you found this useful. Subscribe if you haven't already and I'll see you in the next

video

.