I never expected that my professional life as a chef would intersect meaningfully with artificial intelligence. In my world, precision is measured in grams, timing is instinctive, and success is something you can taste immediately. AI, on the other hand, felt abstract—something distant, technical, and largely irrelevant to my day-to-day routine.
That assumption did not hold for long.
***
Realising This Was Not My First Encounter with AI
In retrospect, my exposure to AI did not begin with the recent surge of tools like ChatGPT. I had already experimented with AI earlier than I initially realised.
In early 2022, I subscribed to Rytr.
At the time, discovering Rytr felt transformative. It genuinely felt as though an entirely new world had opened—one where writing, ideation, and content generation could be accelerated in ways I had never experienced before. There was a sense of novelty and excitement, almost like witnessing the early stages of something significant.
I remember that my housemate (who was also my workmate at the time) used to wonder why I always woke up early, like 4 AM, just to sit in front of my laptop. It was difficult for me to explain how excited I was to explore this new tool.
Little did I know that, by the end of that year, the world was about to be shaken by AI.
***
The Shift: From Novelty to Infrastructure
Fast forward to now, and the landscape is fundamentally different.
With tools like ChatGPT and Claude, AI is no longer perceived as experimental. It has become embedded in daily routines.
What changed is not just the technology, but its role.
AI is no longer something I occasionally explore. It is something I rely on:
- answering questions
- assisting with reports or planning
- helping resolve everyday uncertainties
There has also been a linguistic shift.
It is no longer “Google it.”
It is increasingly, “ChatGPT it.”
That subtle change reflects a deeper transition—from searching for information to interacting with intelligence.
***
Realising What AI Actually Is
For quite some time, I believed that using chat-based AI tools was the full extent of AI. I assumed that asking questions and receiving answers was the core experience.
That belief was incomplete.
Through observing the broader technological landscape—and, admittedly, through being married to someone in IT—I began to understand the scale of what is happening. AI is not simply a convenience layer; it is becoming foundational infrastructure.
Industries are reorganising around it. Workflows are being redesigned. The pace of change is accelerating to a point where even the way I communicate with AI today may become outdated within weeks.
This creates a constant demand: adapt quickly or fall behind.
***
Being Forced to Learn Again
Learning AI feels like returning to education, except the curriculum evolves continuously. There is no stable endpoint. Only ongoing adaptation.
This is where my journey into agentic AI begins.
***
Day 1: Introduction to Codex
My husband installed Codex on my laptop. At first glance, it resembles ChatGPT. However, the similarity is superficial.
Codex operates at a different level. It functions less like a chatbot and more like an agent—capable of executing tasks across systems, not just generating responses.
One limitation is immediately apparent: there is no practical mobile version. Access is tied to my laptop, which changes how and when I can use it.
***
First Experiments: Connecting Systems
Since I did not have a structured plan, I approached Day 1 experimentally.
Integrating with Notion
I connected Codex to Notion.
My first task:
- read my shopping list
- identify items
- provide purchase links
Codex navigated through my folders, interpreted the contents, and returned relevant links—from kitchen tools like a juicer to personal items such as a work bag.
This demonstrated something critical: context-aware action.
Reorganising My “Second Brain”
Next, I asked Codex to redesign my Notion workspace into a more effective “Second Brain.”
It generated a structured proposal for reorganising everything. I have not implemented it yet, but the capability itself is notable.
Here, AI moves beyond execution into system-level thinking.
***
Working with Email at Scale
The most impactful experiment involved email.
I connected Codex to Gmail and asked it to:
- identify the least opened emails
- focus on frequent senders
- analyse the last 100 days
It completed this accurately.
I then attempted to automate unsubscribing. This is where a limitation appeared—Codex could not perform that action directly, so I handled it manually.
However, the next step was far more significant.
I instructed it to:
- delete all emails from those senders
- include emails older than 100 days
It worked.
In approximately one hour, my inbox went from over 16,000 unread emails to a dramatically reduced number. More importantly, I did not spend that time performing the task. I delegated it.
***
A Shift in How I Work
This is the key realisation from Day 1:
I am no longer just using software.
I am assigning work to it.
As a chef, delegation is fundamental—assigning prep, coordinating timing, ensuring consistency. Agentic AI operates similarly, except the “team member” is digital, scalable, and continuously improving.
***
Closing Thoughts: Day 1
I began with the assumption that AI was a helpful assistant.
I am ending the day recognising it as something closer to a collaborator.
And I am beginning to understand that learning AI is not a one-time effort. It is an ongoing process of adaptation.
If this is only the beginning, then the trajectory ahead is not just interesting. It is unavoidable.
Comments
Post a Comment