My Mission for 2026 · The Kodex

WHAT WENT WRONG?

In the last post, I talked about the importance of having a mission.

It has been 3 months past the new year 2026.

I am still confused like ever before on what to do because I have mixed interest and I am not able to focus on anything.

What is the problem?

Initially I started with one goal, but it got diluted with different goals.

My initial goal was related to AI Music, but I changed plan in between to focus on a topic related to my current office work which is related to LLM based pipeline development which is not my area of interest.

But the side effect is that I achieved literally nothing in both areas.

So right now I want to git COMMIT my MISSION for 2026, no matter whoever tries to convince me out of it or try to show "better" alternatives.

WHAT IS MY MISSION FOR 2026?

10 years ago I have watched this video by Thomas Frank Why You Shouldn't Tell People About Your Goals - College Info Geek .

I can agree with it from personal experience.

Over the last 10 years, I had a lot of goals starting from grade 10 to bachelor's to master's, but I used to just simply speak about my goals/dreams with random people so I will feel "completed" and "satisified".

But at the end of the day, I have not achieved anything meaningful.

The only reason I am here is because of God's grace alone.

But on the other side I want to keep myself accountable no matter what, so this is my last try again making my goals public.

I want to see it as a public experiment with two goals:

(1) By default it is known that sharing goals publicly will defeat the purpose, but instead seeing goal as an ending point can I start to see as a starting point and combine with hardwork, will it lead to success?

(2) If it fails/succeed I don't want to take this blog down no matter what as it will serve as a reference point for others so they can take whatever lessons they want.

Here is my mission:

I want to get into AMAAI Lab, Singapore University of Technology and Design as Research Assistant.

So as I said let's take it as a starting point for taking action rather than a heavy weight I need to carry around to prove people around me that I am something special.

Here is the ideal profile for research assistant mentioned in the Prof. Dorien Herremans' website:

Solid expertise in deep neural networks, especially using PyTorch
Strong interest (or hands-on experience) in machine learning applied to audio and/or text
Passion for — or prior work in — music generation systems
Excellent programming skills in PyTorch
At least a Bachelor's or Master's degree in Computer Science, AI, or a closely related field
The ideal candidate brings a strong machine learning foundation, curiosity about sound and music computing, and enthusiasm for collaborating with PhD students and postdocs. You'll help combine individual contributions into powerful, integrated systems that drive high-impact publications.

LET'S BUILD THE MISSION PLAN

AVOID MY PREVIOUS MISTAKES AT ALL COST

Again I will tell another mistake I did.

In the past, I collected a lot of resources (books, papers, YouTube videos) to study something, but I never sit down to study anything properly.

This time I don't want to do that mistake of collecting resources.

My job is not to be a librarian, but to be a learner.

Before building the mission plan, let's put one key thing under watch.

That are my distractions.

X.com and YouTube videos.

Both are great sites if properly finetuned, but as I said in my previous blog:

We are in a mission, any action catering to distraction is a self-sabotaging act (which is painfully irreversible).

So here are my rules:

If I have to use X.com a particular day, I need to hand code 500 lines of code (anything) on that day.
Always keep the Google Chrome YouTube extension "UnHook" in ACTIVE mode, to hide all the distracting elements in YouTube.

LET'S KEEP IT MINIMAL AND AI-NATIVE

Let AI tell me what to learn on a daily basis
Prompt for ChatGPT:

I want to get into AMAAI Lab, Singapore University of Technology and Design as Research Assistant.

Here is the ideal profile for research assistant mentioned in the Prof. Dorien Herremans' [website](https://dorienherremans.com/content/job-opening-research-assistant-large-multimodal-models):

You'll work directly under Prof. Dorien Herremans in the AMAAI lab, contributing to cutting-edge research lines in multimodal generative AI for music/audio. Check out some of our recent projects here: https://dorienherremans.com/biblio (including work on models like Mustango and other generative systems).

- Solid expertise in deep neural networks, especially using PyTorch

- Strong interest (or hands-on experience) in machine learning applied to audio and/or text

- Passion for — or prior work in — music generation systems

- Excellent programming skills in PyTorch

- At least a Bachelor's or Master's degree in Computer Science, AI, or a closely related field

- The ideal candidate brings a strong machine learning foundation, curiosity about sound and music computing, and enthusiasm for collaborating with PhD students and postdocs. You'll help combine individual contributions into powerful, integrated systems that drive high-impact publications.

Help me craft a realistic study book separated as chapters and sections, instead of daily study plan. 

First give the table of content for all the chapters. Split the book into two: (1) Fundamentals (2) Latest research done at AMAAI lab.

After that I will ask ChatGPT to write down one chapter at a time.

Hope I will be able to daily update my regular study sessions.

I want to end this blog with a prayer:

O God be with me and guide me. Amen.