January 21, 2023
In recent weeks, I’ve been trying out the language model and artificial intelligence, ChatGPT. It’s the first large language model based on machine learning that has been made accessible to the public. Let’s take a look at how this could change the world.
The chatbot was developed and trained by the American company OpenAI and released in November 2022. The AI was trained on a vast amount of text data from the internet, enabling it to respond to a wide range of queries and topics. The model uses a so-called transformer architecture, which allows it to apply natural language in a way that makes it hardly noticeable you’re actually talking to a bot.
You can basically ask the AI any question, no matter the topic. The practical thing about it is that the AI remembers conversations and can answer follow-up questions meaningfully. For example, if you have a question about human anatomy and don’t understand something, you can simply say you didn’t get it, and the AI will explain it in an easy way with an example. And it really works great. The information is provided instantly when needed, without having to painstakingly search for it in a search engine.
Human Know-How Must Not Be Lost
In theory, you could have entire articles, emails, and whatnot generated by the AI and published that way, but I personally wouldn't recommend that. My recommendation is to use the AI as a teaching aid and a tool. Otherwise, there will be a flood of AI-generated content, and believe me, if you deal with it a little, you can tell exactly what is AI-generated and what is not. Personal know-how must not be lost.
Does the AI Make Mistakes?
Yes, that can happen. So, I wouldn’t recommend relying on the answers blindly. I’ve tested the chatbot extensively, asking various questions on topics where I already knew the answers. There were some errors. That’s why I suggest having at least a basic understanding of the topics before turning to the chatbot, and not believing everything without question. Just as we would do in real life when someone tells us something. And definitely don’t let the AI calculate anything complex. That still doesn’t work at all—which surprises me a bit, since such functions already exist in other tools.
How I Use the AI
I’ve used the AI to help optimize and tweak the design of the tools on my website. I do have some experience writing code, but ChatGPT is a real help and makes this work much easier. Nowadays, I also use ChatGPT to translate texts. I used to do this mostly with DeepL, which, by the way, is also an AI-powered translation tool.
And as I mentioned, ChatGPT is handy for getting simple explanations of many complex topics. I find it a bit unfortunate that there weren’t tools like ChatGPT back when I was in school. I often had difficulty learning and lacked motivation. I wished for more information on certain topics. Some things are unnecessarily complicated, and you end up having to go through multiple textbooks and websites. With the support of AI, it’s different. I can simply ask for another explanation and get an easy-to-understand answer.
Is ChatGPT the Only Artificial Intelligence?
No, there are many other AI tools, but ChatGPT is among the most advanced. In the near future, many more innovative AI products will hit the market. There are also many AI programs from other companies that can generate or edit images.
Here are, for example, some AI-generated images of me that were created with the Lensa app, based on templates of some normal images of me:
Related Links
ChatGPT:
https://chat.openai.com/
Lensa-App (Android):
https://play.google.com/store/apps/details?id=com.lensa.app
Closing Words
I see a lot of potential here for most areas of daily life, both privately and professionally. History is being made here. So try out the AI and become part of the future.