Today, I'm going to show you how you can use AI and leverage your work with GPT-3.

It is a cutting-edge technology and has the potential to revolutionize the way we interact with computers.

By following this short guide, I hope that you can get started and reduce the amount of complexity that I see in this field.

AI is eating the world

Despite being THE fastest-growing technology in the world, most people have no clue how it works and how to use it.

At the end of this guide, you will know:

• What AI, GPT-3, ChatGPT, LMs, and few other abbreviation mean
• What makes GPT-3 different
• How it works
• 7 application areas where using GPT-3 is a huge advantage 
• What can you build with GPT-3

Finally, understanding AI and ways how you can use it gives you endless opportunities to find a new high-paid job or build your own business.

If you are still thinking about whether should you learn about AI or not, then here is some advice from @michaelsayman (ex-Google, ex-Facebook).

The first part of the guide is about the technology itself and the way how it works. The second part includes practical examples and use cases.

Let's dive in.

Part 1: Everything about GPT-3

What is GPT-3?

GPT-3 (Generative Pre-trained Transformer 3) is a language generation model developed by OpenAI that uses machine learning to generate natural language text.

The first version of GPT was released in 2018, followed by GPT-2 in 2019, and GPT-3 in June 2020.

GPT-3 is the third generation of the GPT series of language models and is the largest and most powerful version to date, with 175 billion parameters. It is designed to generate human-like text across a wide range of topics and can be fine-tuned for specific tasks such as translation, summarization, and question-answering.

I know what you're thinking... AI can generate text several years ago and why it makes sense now...

What makes GPT-3 different?

In the world of AI, the size of trainable parameters matters a lot. The more parameters, the better outcomes. The more parameters model has, the more data developers need to train it. The more data used to train the model, the more advanced model it is. The size of trainable parameters of GPT-3 is 10x more than any previous model.

It is one of the most advanced language models currently available.

What are ChatGPT and LMs?

ChatGPT is a long-form question-answering AI that can answer complex questions. Basically, ChatGPT is a function of GPT-3 model.

So, it's a language model that understands you (sometimes even better than people) and can give you deep and detailed answers to different variety of questions (sometimes even better than Google).

But what is a language model?

A language model (LM) is a type of machine learning model that is trained to predict the likelihood of a sequence of words in a given language. The goal of a language model is to assign a probability to each possible sequence of words, with higher probabilities assigned to sequences that are more likely to occur in the language.

Language models are used in a variety of natural language processing (NLP) tasks, including machine translation, speech recognition, and text generation.

They are an important component of many NLP systems, as they help the system understand and generate human language.

How does GPT-3 work?

Time to prompt.

To generate text, GPT-3 takes in a prompt, which is a short piece of text that provides context for the text that the model is being asked to generate. The model then uses this prompt and its knowledge of the language to generate a sequence of words that is likely to be coherent and relevant to the prompt.

Prompt can be anything you want to ask GPT-3. The more specific prompt, the more chances to get a good answer.

Advantages of using GPT-3

1. High-quality text generation.
2. An efficient and cost-effective way to generate large amounts of text.
3. Versatility: GPT-3 is able to generate text in a variety of languages and styles, and it can be customized to meet the specific needs of different users.
4. Easy to use: GPT-3 is relatively easy to use. Users simply need to provide a clear and concise prompt to the model, and it will generate text in response.
5. Continuously improving: GPT-3 is an active area of research, and the model is constantly being updated and improved.

GPT-3 Alternatives

GPT-3 is not alone here.

There are several alternatives to GPT-3 that are also used in the field of artificial intelligence (AI) and natural language processing (NLP). Some examples include:

1. PaLM (Pathways Language Model): The model is created by Google. It's not publicly available. However, it's 3x larger and more powerful than GPT-3. It's a multimodal model across vision, sound, and language all at once. So, I'm waiting for its public launch.
2. LaMDA (Language Model for Dialogue Applications): LaMDA is a language model created by Google. LaMDA made headlines when a Google engineer was fired for calling it so realistic that he believed it to be sentient.
3. BERT (Bidirectional Encoder Representations from Transformers): BERT is a pre-trained NLP model developed by Google that can be fine-tuned for various NLP tasks such as language translation, question answering, and sentiment analysis.
4. ELMo (Embeddings from Language Models): ELMo is another pre-trained NLP model developed by researchers at the Allen Institute for Artificial Intelligence (AI2). It can be fine-tuned for various NLP tasks and has been used in a number of research projects.
5. ULMFiT (Universal Language Model Fine-tuning): ULMFiT is a pre-trained NLP model developed by researchers at fast.ai. It can be fine-tuned for various NLP tasks and has been used in a number of research projects.
6. RoBERTa (Robustly Optimized BERT Approach): RoBERTa is a variant of BERT that was developed by researchers at Facebook AI. It is designed to be more robust and efficient than BERT and has achieved strong results on a variety of NLP tasks.
7. Transformer-XL: Transformer-XL is a variant of the Transformer model that was developed by researchers at Google. It is designed to be more efficient and able to handle longer sequences of text, and has been used in a number of research projects.

How to get started with GPT-3?

It's time to do some AI things. Let's work with GPT-3.

There are several ways to get started with GPT-3 and they are based on your knowledge:

1. No-code option: use ChatGPT or an application that uses GPT-3. There are a number of tools and applications available that use GPT-3. Some of these tools are free to use, while others may require a subscription or payment. For example, here is a list of 400+ apps that use GPT-3.
2. GPT-3 API [no-code option too]: OpenAI offers an API (Application Programming Interface) for accessing GPT-3. To use the GPT-3 API, you will need to sign up for an API key and follow the instructions for making requests to the API. It is required some basic understanding of the code. However, if you don't know the code, you can use No-Code tools for making calls to GPT-3 API. For example, here is a tutorial on how to set GPT-3 with Bubble.
3. [For developers] Local installation of GPT-3: If you have the resources and expertise, you can install and run GPT-3 on your own computer or server. This option will require more setup and may not be feasible for everyone, but it can be a good option if you need to run GPT-3 locally or have specific requirements that are not supported by the API.

Tips for using GPT-3 effectively

Prompts are keys.

1. Understand the capabilities and limitations of GPT-3: GPT-3 is a powerful tool, but it is not a substitute for human intelligence. It is not able to think like a human (yet), and it is not able to understand the context or meaning of the text it generates.
2. Choose the right GPT-3 model: GPT-3 comes in a range of sizes, with different capabilities and price points. It is important to choose the right model for your specific use case. The smaller models may be sufficient for many tasks, while the larger models may be necessary for more complex or demanding tasks.
3. Provide clear and concise prompts: GPT-3 is most effective when it is given a clear and concise prompt that specifies the task it should perform. The more specific and focused the prompt, the better the results will be.
4. Use the API wisely: GPT-3 is accessed through an API, which allows users to send requests to the model and receive a generated text in response. It is important to use the API wisely and not overuse it, as it can be expensive and there may be limits on the number of requests that can be made.
5. Post-process the generated text: GPT-3 is capable of generating high-quality text, but it is not perfect. It is a good idea to post-process the generated text to ensure that it is accurate, coherent, and appropriate for your specific use case.

At this stage, you know GPT-3 basics and are ready to join to AI world. Let's look at the real-world use cases.

Part 2: GPT-3 use cases and Examples

Top-7 areas of GPT-3

Most of the GPT-3 applications work in these 7 areas:

1. Content generation: GPT-3 can generate a wide range of content, including articles, blog posts, and social media updates.
2. Language translation: GPT-3 can translate text from one language to another.
3. Summarization: GPT-3 can summarize long documents or articles, making it easier for users to understand and digest large amounts of information.
4. Question answering: GPT-3 can generate answers to a wide range of questions.
5. Chatbots and conversational agents: GPT-3 can build chatbots and conversational agents that can engage in natural language conversations with users.
6. Text classification: GPT-3 can classify text into different categories or labels based on its content. This can be useful for tasks such as spam detection or sentiment analysis.
7. Text completion: GPT-3 can complete partial text inputs, such as sentences or paragraphs, by generating appropriate and coherent text. This can be useful for tasks such as predictive typing or autocomplete.

And here is the forecasted cumulative global AI revenue for 2025 from Statista.

What can you build with GPT-3?

There are some of the existed GPT-3 applications. You can use these ideas for your inspiration:

1. Marketing applications: Copysmith, Jasper, Copy.ai. Content creation is the most popular case of GPT-3.
2. Creative applications: Fable Studio. One of the strong sides of GPT-3 is storytelling. So, Fable Studio used GPT-3 for generating dialogues for the children's book Wolves in the Walls.
3. Chatbot applications: Quickchat. Emerson AI is the company Quickchat's chatbot persona that is known for its general world knowledge.
4. Data analysis applications: Viable. Viable is a feedback aggregation tool that identifies themes, emotions, and sentiments in surveys, help desk tickets, live chat logs, and customer reviews.
5. Coding applications: Stenography. Stenography is a program that uses GPT-3 and Codex to automate the process of writing code documentation. 

Other use cases of GPT-3

It's not all. You can also use ChatGPT in your day-to-day. Here are some examples of how people used ChatGPT:

• Create video in 5 minutes (AI writes the script, AI creates video and sound, and edits it 🤯)
• Create summaries of meetings
• Generate content ideas
• Generate an SVG icon
• Create a list of US cities and convert it to a CSV file
• Create an AI dating coach :)
• Create 3D models
• Create patterns for clothes
• Write poems
• And much more

Interesting AI startups

As I want to provide more value for you, I gathered a list of interesting startups in the AI space:

• Brancher.ai - Connect AI models to build powerful apps - without code
• Sitekick - Create a beautiful landing page in minutes
• BCG-3: Generative AI for Slides.
• YouTube to Summary - Convert a YouTube video (in Spanish or English) into an illustrated summary
• Thread with recent investments in AI startups

What to do next?

If you are excited about AI and the opportunities it can create, then here are a few tutorials to learn AI by doing small practical projects.

• Generate avatars with Stable Diffusion for free (no-code)
• Create 80 web designs with Midjourney (no-code)
• Build an AI image generator with OpenAI and Node.js

Conclusion

The potential impact of GPT-3 on industries and society is significant and multifaceted.

On the one hand, GPT-3 has the potential to revolutionize various industries: customer service, finance, healthcare, and others.

On the other hand, the widespread adoption of GPT-3 could also have significant societal impacts. For example, it could lead to the automation of certain jobs, potentially leading to job displacement and economic disruption.

Meme of the week

P.S. If you found this newsletter useful, please let me know about it on Twitter. I appreciate any feedback or support.