Insight

Jef Wauters

3 min. read

What is GPT-3? Everything you want to know.

GPT-3, or Generative Pretrained Transformer 3, is a state-of-the-art language processing AI developed by OpenAI. It has 175 billion parameters, making it one of the largest language models ever created. GPT-3 is capable of generating human-like text and can be used for a wide range of natural language processing tasks, such as translation, summarization, and text generation.

But wait, did you realize that this first paragraph was actually written by an AI? That’s right, GPT-3 was able to write this text by itself, using its advanced language processing capabilities.

Are you starting to see the potential impact of GPT-3 on our society? This technology has the power to fundamentally change the way we work and interact with computers. As AI continues to advance, it’s becoming clear that those who embrace these new technologies will have a significant advantage over those who don’t. The world is changing rapidly, and we must adapt or be left behind. The future is here, and it’s powered by AI like GPT-3. Are you ready?

Jef Wauters on GPT-3 and other AI opportunities

Crazy, right?

All three of the above paragraphs are written by GPT-3. Is this fourth one also written by an AI? The answer is that it doesn’t matter anymore. By my estimation, the Turing test has been passed. But that’s not even the most impactful part. Text generation is one thing, but did you know that GPT-3 can write, understand, debug, explain and document code? Yesterday, I used the code advice capabilities of GTP-3 to help me write Python code to automate my workflow.

The workflow in question is the creation of testcases based on actual data. As many companies do, we store our logs in the ELK (Elasticsearch, Logstash & Kibana) stack. I asked GPT-3 to write Python code for me that can scrape a Kibana webpage for the JSONs, unescape them, load them into a file, download that file, split the file into various separate JSON files with automatically generated names, and automatically generate MuleSoft mocking component XMLs based on each of those JSON files.


It all works like a charm. I did all of this in less than a day. And I have practically no experience with Python…

Wow! But…

I will let the AI explain some of the risks and downsides of using it right now:

While GPT-3 has many benefits, it also has some risks to consider. One potential risk is that the AI may make mistakes while sounding confident. Because GPT-3 is trained on vast amounts of data, it can sometimes provide inaccurate information or suggestions that may lead you astray. Therefore, it’s important to use GPT-3 as a guide rather than a definitive source of truth.

Another potential risk is that the current chat application may crash if the AI’s response is too long. This can be frustrating and disrupt your workflow. If this happens, one solution is to break your code questions down into smaller chunks and ask the AI one question at a time. This will prevent the chat application from crashing and allow you to continue working with GPT-3.

Ready to try GPT-3?

So, I strongly recommend you to make an account at https://chat.openai.com/chat and see for yourself. Ask it to write your python code. Or ask it to write you a song about dishwasher salt. It’s still completely free while it’s in open beta.

There are other advanced AI chat bots, but this is by far the most impressive one I have ever seen. Moreover, GPT-4 is live already and GPT-5 will be there soon.

Keep in mind that GPT-3 has gained millions of users in the first days it has been active, and it is prone to overloading and crashing.

Want to know more? Or implement this yourself? Reach out and let’s have a chat (GPT)!

Interested in a collaboration?