Explaining chatGPT to a 5-year old. Is it really coming for your job? Why should you care?

Vino Duraisamy
4 min readDec 7, 2022

--

Debunking the myths in 5 minutes!!

chatGPT flooded the internet this week! Students are excited to make it write their assignments. Marketing folks are wary of GPT taking away their jobs. Engineers are thrilled to make it write code snippets, documentation and what not!

Everyone is trying their hands at it, someone already developed a chrome extension to show chatGPT results in Google search results. Some of us even love chatGPT better than Google.

While the tech world is taken over by AI storm, let us look at what exactly is chatGPT and GPT3.

Should you be worried or be hopeful of a future with generative AI models like GPT? Why does it matter to you?

People are trying their hands at chatGPT and have mixed feelings about how helpful the chatbot is!

What is GPT3?

GPT: Generative Pre-trained Transformer. It is an auto-regressive language model that uses deep learning to generate human-like text. It is trained by OpenAI on Microsoft supercomputing engine.

Now let’s look at an analogy!!

A 6-year-old kid goes to school and got his books for the academic year!

There was no one to teach him anything, so he memorized those books himself. So, he sits at home memorizing all the books — English, Math, Science, Social sciences, etc. Memorizing might be an exaggeration but let’s keep it that way for the sake of simplicity.

Now, you ask him to write or recite a poem!

He can recite Shakespeare’s sonnets or whatever is in those books he studied. You ask anything outside of those books, he doesn’t know. Simple so far. Right?

Next, you ask him to prove a math theorem!

He is going to vomit a bunch of words from his Math book which may or may not make sense. No one taught this kiddo critical reasoning or logical thinking — that are needed for math. He simply memorized the books and can only parrot what is in them. With me, so far?

Can he recite the contents from all the books he read?

Although he memorized all the books, he cannot remember everything from the books word by word. So, he makes up his own sentences to answer your questions. Sometimes these sentences make sense and sometimes it is absolute word vomit.

So, can chatGPT replace Google?!!!

Ask chatGPT anything that happened in 2021 or later. See - It can’t browse the internet — which is exactly what Google does.

Can we re-train GPT to keep it up-to-date? Then it can replace Google!

Even if we re-trained it everyday, there will be 1-day lag. Microsoft invested more than $1 Billion in GPT3 and OpenAI. They couldn’t re-train the model due to the high cost associated with training.

Then, what good is chatGPT?

GPT is a Generative AI model. It can…
- be good at generating/creating texts (i.e., making things up lol)
- be absolutely terrible at logical reasoning or critical thinking tasks.

Limitations of GPT and language models, in general:

limitations of chatGPT, as told my the bot itself.
limitations of language models and Galactica — a specific language model trained by Facebook (now, Meta).

Conclusion!

No — chatGPT or GPT will not replace Google.

chatGPT was purpose built to be an intelligent chat assistant. Large Language Models(LLM) like GPT3 are trained on a large corpus of data to make it understand large swathes of natural language texts. LLMs are capable of doing general NLP tasks like sentiment analysis, sarcasm detection, question answering, text summarization, text creation and so on!

So, chatGPT is coming for your jobs if you are in Tech Support or Content writing. :( For others, it is going to make your work only easier!!

Thanks for Reading!

If you like my work and want to support me…

  1. The BEST way to support me is by following me on Medium.
  2. For data engineering best practices, and Python tips for beginners, follow me on LinkedIn.
  3. Feel free to give claps so I know how helpful this post was for you.

--

--

Vino Duraisamy
Vino Duraisamy

Written by Vino Duraisamy

Developer Advocate @Snowflake❄️. Previously Data & Applied Machine Learning Engineer @Apple, Nike, NetApp | Spark, Snowflake, Hive, Python, SQL, AWS, Airflow

Responses (3)