GPT3

Some GPT3 Practical Applications And Sample Generated Content

If you have not heard of GPT3 yet you should know that at the moment of writing this article GPT 3 may be one of the most advanced text/code generators in the world. 

First I would like to offer my tribute to The Guardian as through their article I heard about GPT 3. 

Next I’m part of the beta program and I can say I am impressed in some areas while not in others. 

One area where GPT 3 is truly impressive is its ability to summarize content. Don’t have time to read a book but you’re interested in its main ideas for example? Let’s see what GTP 3 can do about it.

For example I provided the following input with a temperature of 0.7 and maximum length 3000 tokens: 

Summarize the book The Five Love Languages.

And GTP3 replied:

The book The Five Love Languages by Gary D. Chapman explores the different ways people express and receive love. He argues that there are five love languages: words of affirmation, quality time, receiving gifts, acts of service, and physical touch. He provides tips and advice on how to identify and effectively communicate in each love language.

If you did not read the book you may not find its summary interesting. But I can assure you what you read above is the main idea of the book in as few words as possible. I find this feature really impressive.

Next I decided to test its abilities to generate interview questions. I’m a PHP programmer and I went to around 100 interviews so far so I know a thing or two about PHP interviews. I provided the following input to GTP3 with the same temperature and tokens:

Create a list of 13 questions for my interview with a PHP programmer:

And GPT replied:

1. What is your background in programming?

2. What experience do you have with PHP?

3. What do you think are the strengths of PHP?

4. How do you approach problem solving in programming?

5. What have you done with PHP that you are particularly proud of?

6. What do you think makes a great PHP programmer?

7. What do you think separates the best PHP programmers from the rest?

8. What challenges have you faced in your programming career?

9. How did you overcome them?

10. What resources (books, websites, etc.) do you use when working with PHP?

11. What do you think makes a great website or application?

12. Have you ever encountered a problem with a website or application that you couldn’t solve?

13. How did you go about solving it?

What is interesting enough is the fact that the last two questions are the behavioral type. If you wonder how I know about behavioral questions and want to learn more about these kinds of interviews or how to prepare for them I recommend the following Coursera course: Advanced Interviewing Techniques. And I can say GPT3 beats some technical people that asked me interview questions. In my country technical people are extremely focused on hard skills while basically ignoring the possibility or impossibility of me fitting in the team. They guess this is the job of the HR but the technical leaders know the team best, not the HR.

There are people who still laugh about AIs capabilities but many will lose their jobs. This is obviously not the job apocalipse as a lot of jobs in the creativity area will come into existence. 

Which are the people who will get beaten by AIs? 

The “smart” folks who don’t read books, take courses or train themselves because “they already know enough” or “they are always right”. These demigods will not find their way easy in this new world created by AIs. As more and more people realize emotional intelligence is the skill of the future and even present for some organizations. If you wanna compete with robots you need to develop emotional intelligence – I actually have an idea regarding this but I have not started coding it yet. No matter how smart or capable a robot will be, emotions may never be understood and transposed mathematically.

A technology like GTP3 will soon be advanced enough to replace average technical managers for example. What do average technical managers do: discuss specifications with the client and translate them to the team. An average technical manager is like a boss, unable to motivate the team for example. So the average ones are just information gathering people and translators.

Let’s see if GTP3 can replace the average ones. 

I started with the following input and reply.

Create a list of features for an eCommerce website:

-Product catalog

-Product search

-Shopping cart

-Checkout

-Payment processing

-Order management

-Customer account

-Customer service

-Shipping

-Returns

-Coupons and discounts

-Analytics and reporting

I continued, of course, in the same manner.

Create a list of questions to help my client decide what kind of Product catalog he wants:

1. What kind of products do you sell?

2. What is your budget for a product catalog?

3. How many products do you need to include in your catalog?

4. Do you need a custom catalog or will a generic one suffice?

5. What is your timeline for creating a catalog?

6. Are there any specific design elements you would like to include in your catalog?

7. Who will be responsible for creating and maintaining your catalog?

You can see where I’m going with this. Give the answers to the questions above to a lead developer (not the team leader which is already busy with the team), ask him to discuss details with the client and you don’t need a manager anymore.

Of course this whole thing seems far fetched for the moment. Just wait for technologies like this to get perfected and implemented in applications. The potential cannot be evaluated.

Where does GPT not shine in my opinion? 

Writing niche content. I actually tried to create some niche blogs using GPT3 a while ago and I got quite disappointed.

Let’s see again how things are. Let’s say I want a digital marketing blog and need an advanced article in the PPC category. I actually borrowed a headline from Search Engine Land for this.

So I provided the following input:

Write an article about How paid search marketers can address brand measurement challenges and grow in their careers.

And received:

As a paid search marketer, you know that brand measurement is essential to success. After all, how can you show the value of your work if you can’t measure your impact on the bottom line?

But brand measurement can be tricky. You need to account for both short-term and long-term effects, and you need to make sure you’re comparing apples to apples (e.g., comparing brand awareness in your target market to brand awareness in a control group).

Fortunately, there are a few ways you can address these challenges and grow in your career.

1. Get educated on brand measurement.

There’s a lot to learn when it comes to brand measurement, but it’s worth the effort. The more you know about the topic, the better equipped you’ll be to tackle challenges and show the value of your work.

There are plenty of resources available on brand measurement. Start with some of the basics, like this article from Search Engine Land, and then move on to more advanced concepts.

2. Use the right tools.

There are dozens of brand measurement tools out there, so it’s important to choose the right ones for your needs. Some paid search marketers prefer to use general tools like Google Analytics, while others prefer specialized tools like Moz’s Search Metrics.

The key is to experiment and find the tools that work best for you. There’s no one-size-fits-all solution, so don’t be afraid to try out multiple options.

3. Stay up to date on industry trends.

The world of brand measurement is always changing, so it’s important to stay up to date on the latest trends. This way, you can be sure you’re using the most effective methods and tools.

There are a few great ways to stay informed. Subscribe to industry publications, like Search Engine Land and Search Engine Journal. Attend conferences and webinars. And connect with other paid search marketers on social media and in online forums.

4. Be prepared to adjust your approach.

No matter how well you plan, there will always be some trial and error involved in brand measurement. You need to be prepared to adjust your approach as you learn what works and what doesn’t.

Don’t be afraid to experiment. Try new things, and be willing to change course if something isn’t working. The goal is to constantly improve your results, so never stop learning and adapting.

5. Grow your career by helping others.

As you become more successful in brand measurement, you can start to share your knowledge with others. This is a great way to grow your career and make a lasting impact on the industry.

There are a few ways you can help others. Write articles and blog posts about your experiences. Speak at conferences and webinars. And mentor other paid search marketers who are just getting started.

Paid search marketing is a demanding field, but it can be extremely rewarding. By taking the time to learn about brand measurement and staying up to date on the latest trends, you can set yourself up for success. So don’t be afraid to challenge yourself and grow in your career. 

Now you may find the article good enough as it provides some inspiration and it does not look like it has been written by a robot. There are a lot of blogs out there that provide content just like this. But those blogs have little traffic and they will never ever get big. The obvious problem is the fact that the content is way too general.

Of course the developers of GPT3 have already thought about this and they give you the option to train your model. This is something I have not tried yet because I don’t know exactly where to get the training data. I have an idea how I could create it but it would take a lot of time.

Of course with GTP3 and proper training models you could theoretically create niche article generating applications. There is quite some competition in this area on the web.

So these are some of the applications of GPT3, next time I may try to generate some code or even refactor it. 

Leave a Reply

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.