S02E02: The Love-Hate GenAI Relationship

Rutvi Dresswala
Oct 1
3 min read

ChatGPT was born in late 2022. We were introduced in early 2023. It was love at first chat.

I was fascinated—completely in awe of this new technology that felt straight out of a sci-fi movie. It wasn’t just me. OpenAI and ChatGPT were everywhere - in newspapers, on my YouTube feed, in conversations with friends, and at work, where we were busy exploring how and what to build with LLMs. Even my parents - who had taken forever to warm up to tech- started using LLMs for their daily tasks because they now had easy access through WhatsApp.

I like to think of myself as a techie, and what I saw (or rather predicted) was a fundamental shift in how technology would be built and how human work would evolve. And now, almost three years post-ChatGPT’s launch, my theory is that it has fundamentally changed how we build and work—but it hasn’t replaced human thinking. Rather, it’s enhanced it, giving us back precious time to focus on thinking and growth.

I’m not scattering gyaan here as a pro-AI advocate (though I am one 😉). I’m writing from experience - specifically, the experience of developing a healthy relationship with AI, which (swear to God) took me almost 1.5 years! So today, let’s dive into the phases of my (hopefully) healthy relationship with GenAI.

Like any romantic relationship, GenAI and I went through the rite of passage of multiple phases:

💫 Infatuation Phase

We all went through this. Playing with ChatGPT for funny little tasks - writing a few emails, correcting documents, planning trips, etc. Harmless, fun, and magical to see a new technology at work.

❤️ Honeymoon Phase

This is when the dependency started. I used it for consequential tasks: applying for my French residency, redoing an article for the INSEAD website, drafting PRDs, even rewriting code for efficiency. Hot and heavy, I began to trust the LLM more and more.

😬 Toxic Dependence Phase

You know when you trust a new partner too much and it blows up in your face but you’re still in denial? That was me with ChatGPT. I swear it should have graduated from INSEAD because it did most of the work. At one point, my attention span was gone, deep work became a struggle, and brain fog settled in. Ay ay ay—I was a bit of a mess.

💔 Breakup Post-Toxicity Phase

After a lot of contemplation and threats from my old-school, “things-should-be-done-a-certain-way” boyfriend, I decided to cancel my OpenAI Pro subscription and actively reduce my LLM use. It was tough - this was midway through my MBA, and everyone around me had their LLM buddies. But slowly, the fog lifted, and I felt like I was finally thinking again, not just CTRL-C/CTRL-V’ing my way through.

🫂 Patch-Up Phase

Eventually, I realized unless I planned to become a farmer in rural India, I couldn’t ignore GenAI or refuse to use it. I had to learn to use it better. This began a new phase: being more intentional with my prompts, more cautious with the responses.

🌱 Healthy Phase

This is where I am now - what I believe is the healthiest way to use LLMs:

It’s your assistant who never wants a promotion.
It’s great for redundant tasks, deep research assistance, editing content, etc.
It’s also a little stupid - so be clear and detail-oriented in what you want.
It will do a good job, but for consequential tasks, you still have to verify responses.
Most importantly: you are the thinker. The LLM is the support. Don’t let it be the other way around, or you’ll end up back in the toxicity phase 😛.

I’m not sure if this has been everyone’s experience with LLMs, but it’s been mine. It took a while to fully get on board with this beautiful tech and understand more clearly how it might change the world we know 😊

I’d love to hear if this resonates with you or what your experience has been. I’m always a ping away on LinkedIn.

Until next time, I hope you and your LLM buddies are doing well 😛

Fin.