This blog is a good read if you are still poking in to your mobile to read some interesting and relatable Tech stories. Here I share my Experience which resembles the way Generative AI is behaving these days.
Today I was curious to understand and learn the behaviour of Gen AI. So, I was learning a topic on a Kubernetes policy tool Gatekeeper that implements Open policy agent (OPA) a well know CNCF project. As GenAI has been trending in these days I thought of taking help of two well known tool ChatGPT and Gemini.
I spent good amount of time on ChatGPT and Gemini, In this 3 hours I started asking about basic Questions and it worked liked wonder. I thought I am with the best companion ever for learning.
Resemblance with Real experience:
Reminded me when I was a fresher and started giving my job interviews . I had all the theoretical knowledge and I was able to crak the first round and was high on confidence.
These flawless responses inspired me to explore further with more complex queries, hoping to have an insightful conversation. Instead, the tool responded with inaccurate information frequently, When I pointed out these inaccuracies, it responded with partially corrected statements, each reply was prefixed with "Apologies for the oversight"< ......>".
I realised AI brilliance can sometimes be deceptive..
this again mirrored moments from my interviews:
During my second round of interviews, some of the interviewers were kind enough and allowed me to correct myself when I made mistakes. I cleverly navigated these situations by confidently prefacing my answers with,
"Sorry, I just recalled".as I heard somewhere " you should be confident in the interviews even if you are wrong". Interestingly this approach proven worked many times.
After few conversation from these AI tools , my curiosity turned into amusement.
As I delved deeper, the responses began citing :"As of my last update in January 2022",<...>
Despite feeling frustrated and tired after almost 5 hours . I admired the confidence of these AI tools they were still eager to assist . ChatGPT's Prompt was Continuously beckoning with "How can I help you today?". Gemini Star blinks as though analyzing the forthcoming question's response.
Suddenly, few college days memories flashed in the back of my mind which was more relevance in this context of Gen AI tool.
Here begins the Story :
It was the 2nd semester in my engineering (in 2009). Himanshu and I breezed through the semester until the exam schedule dropped. Usually we work hard day before the semester and struggled to avoid back papers 😁. thanks to those summery (Alok) book📖.
Let me explain the hard work phase :
Somehow, our hostel mates recognized Himanshu and me as the brilliant students, relying on us for answers to all their questions like ChatGPT or Gemini. We didn't want them to feel hopeless. During semesters our hostel room was houseful, Some of them seemed to be Dull but they had hope on us.
To start with
- we close the door and open YouTube and play a Verse from Bhagavad Gita . The most play one was chapter-2 verse-47.
"Karmanye vadhikaraste Ma Phaleshu Kadachana |
Ma Karma Phala Hetur Bhur Ma Te Sango Stv Akarmani ||
Meaning :
You have right to perform your prescribed duty, but you are not entitled to the fruits of actions.
Through this sloka, we inspire/ motivate them to at least fill up the answer booklet rather leaving it blank, at least you can hope to pass.
Now everyone is energetic Ready to consume all the Knowlwdge like blackholes.
I start reading the summery book📚 (Alok) and Himanshu illustrates them like the Krishna 🕉️.
As always there was no question or doubt. A classy example of last hope.
Now its time to break for sleep to wake up on time to the exam hall.
In the exam hall, as I received the Thermodynamics question paper , I quickly noticed that I only had answers for three 1-mark questions and one 5-mark question.
I acted like a AI tool 😉, decided to write those four answers at the beginning, Because first impression lasts long. From there, I began writing whatever came to mind, sometimes straying out of context and occasionally presenting contradictory views. However, I never left the answer paper blank, understanding that my duty was to perform, regardless of the outcome being beyond my control.
This is how the renowned Gen AI tools are Behaving these days 📝
On result day, I wasn't particularly worried, as I understood the potential outcomes. However, like any normal human, I still felt a sense of hope. Coincidentally, I had passed with a C-grade, which was beyond my expectations. Now I realized the power and sweet spot. And the history repeats.
Moral of the story:
Gen AI Is immature now so use Responsively. It may mislead often
Validate each piece of info you get from tool then Accept or learn
Can give you baseline Idea, To know the fundamentals please refer official guides and evident articles to enforce your analysis and research.
Cherised memory with Tossali, Himanshu and Sidharth 🍻.
Do let me know if you like my stroy !