TIL mini blog

This tiny tiny blog is a collection of things I learned today!

(Well, okay, not TODAY today, but you get the idea.)


Random 🎲 See all
Muh face


April 13, 2023 - 13:34 EST

Treat ChatGPT as a coffee break colleague.

Not malicious by default

Let's say you're in the break room with Linda from HR that tells you she thinks we didn't go on the moon, that the earth is flat but we can't see it because of the chemtrails. You will most likely just roll your eye far enough to see your own brain, and go on with your day as if nothing happened.

But, if it's something somewhat credible? Let's say I'm your colleague and tell you three different things during our conversation :
  • You cannot catch a cold just by being in a cold place
  • Santa Claus used to be green before a Coca-Cola marketing campaign
  • Mosquitoes are not drawn towards light

Two of those assertions are right (as far as what we know today) and one (the Santa Claus one) is wrong, sorry for lying to you a bit here :)
When we are talking with someone else, a human being, we acknowledge that they have flaws and can lie to us. More often than not, not intentionnaly. We believe them most of the time, but we should be also wary of their bullsh*t. The key to navigating these conversations is critical thinking and healthy skepticism.

I thought we were talking about ChatGPT?

Yes, we are!
When interacting with AI like ChatGPT, adopting a similar approach is invaluable. While AI tries to provide accurate and relevant information, it's not infallible.

We should recognize that the AI is a tool, not an all-knowing oracle. By treating AI conversations like a coffee break chat with a colleague, you can maintain a discerning mindset and make the most of the valuable insights that AI has to offer.
Embrace the benefits of this powerful technology, but always remember the importance of critical thinking and healthy skepticism.

Here are some tips I usually keep in mind when I ask ChatGPT about some topic :
  1. Verify information: If you receive information from an AI, double-check it with other reputable sources before accepting it as accurate. This practice helps ensure that you're relying on well-founded data. Use any search engine you're used to and muliple websites if it's sensitive information.
  2. Be cautious with sensitive topics: While AI can offer general guidance, complex or sensitive matters may require a more nuanced understanding that AI might not fully grasp. In these cases, consulting a human expert may be more appropriate like a doctor or a scientist.
  3. Keep an open mind: Remember that AI is constantly evolving and improving. As with any interaction, be open to learning new information and refining your understanding.
  4. Don't take anything for granted: If you use AI for computer science, check that the code/tips provided is reliable and is secure to use for your apps/webiste. If you ask for medical advice, consult a doctor or a practitioner anytime you can and DO NOT take medicine based solely on what the agent say to you. Don't spread widely information given to you by the AI without verifying it to avoid spreading fake informations and news. 

Sign up and never miss a post!

My personal infos




+1 (437) 423-0565


Mississauga, Canada

© 2024