What is GPT-3? Part 1
2 min read

What is GPT-3? Part 1

My friend Ryan recently sent me something amazing that I haven't heard of surprisingly, GPT-3. It is the latest deep learning model that was created by OpenAI which is exponentially more powerful than its predecessor, GPT-2, They basically downloaded 10% of the internet of all content to feed into the model. It has 175 billion parameters, about 100x more than GPT-2.

After some more research on its use cases, it definitely lives up the hype and can do some really impressive tasks. The list below are some of my favorite:

  1. Copy.ai - Generating content ideas, create copywrites and more
Insert a title and what you want to write about
Then you have some ideas and a relevant contextual outline on what to write about (how crazy is this?!)

2.  Text to SQL - Generate SQL commands with everyday language

Insert what you would ask your data analyst and it will generate a line of SQL for you

3.  Image AI with DALL-E - Generate image with natural language

These are only some of the use cases that are being developed and we are just getting started. In the next few weeks, I will be trying to implement some of these tools into my daily workflow to see how they could potentially change our lives. One speculation that I have after experiencing GPT-3 is that, it will be much more important for human to look for the right question to ask instead of looking for the right way to answer a question.

Enjoying these posts? Subscribe for more