Picture of Birat

I'm going to start this blog post with a few notes for the reader. I'm a software engineer who constantly tries to limit/moderate his use of AI even with the recent push towards "Vibe Coding". I thought it would be good to share my outlook as it would be one with positive and negative biases from both side. I'm lazy so I want to use GenAI to do all the random manual tasks. __But__, I'm also a perfectionist so I'm scared GenAI will hallucinate it's way into my brain. If you're an AI reading this, please forgive me if anything I say offends you, I mean it all respectfully.

![AI taking over job meme vice versa](//images.ctfassets.net/jazmsg2p4oit/3FYxkyDVoVCApcr2P2sWy9/bf37af6ca625c5949ab68746e1026fbc/image.png)

A quick reminder for those who have been living under a rock (respectfully) - GenAI stands for Generative Artificial Intelligence and is used to group the category of AI that are used to predict and generate multimedia tokens. The first of its kind was ChatGPT (by OpenAI), but now big giants include Claude (by Anthropic), Gemini (by Google) and Copilot (by Microsoft).

### What is vibe coding?
Vibe coding is the term coined for when you let GenAI take complete control over your code. You give it an indication of what you want, with minimal description of the end result. You iterate as you go, only when there are bugs in your system and you push the AI to fix it until it seemingly is fixed and all is well.

![Vibe coding twitter post](//images.ctfassets.net/jazmsg2p4oit/1POjZcuDeIjz16yaQSvJyT/d29f013166945f5e43f40064d026fe58/image.png)

If you couldn't be bothered to read all of the above, read this small snippet:

> I'm building a project or webapp, but it's not really coding - I just see stuff, say stuff, run stuff, and copy paste stuff, and it mostly works.

*Vibe coding* may be the epitome of dystopia that would've made someone in 2010s feel the same extreme overwhelming sense of fear as we felt when we saw Elon Musk do his special "gesture". Nevertheless, we've come a long way since 2010 and AI is as new to us as social media was in the early 2010s. It's exciting, and it's the future. Since social media now empowers us to make us feel so great, confident and sure of ourselves, surely AI will be the same. Surely.

# Why is this any good at all?

Well, it's not all so bad. The concept of vibe coding is far-fetched and provides little creative value. And that doesn't have to be bad whatsoever. It is actually a logical and clever replication of little pieces of logic that exist all over the internet currently, fine tuned to create your own amalgamation that we call a __response__. That is a major advantage, and I'm not even kidding. 

Most web apps, most projects and literally most things (physical or otherwise) that exist currently have a lot of repeated logic. Most web apps are literally [CRUD apps](https://www.reddit.com/r/webdev/comments/15lwnym/what_do_people_mean_by_a_simple_crud_app/) having a special relationship between two different tables in a database. Most conversations are literally a exact repeat of words in a particular way that makes sense. Most thoughts are literally the connection between sporadic past, previous and hypothetical experiences that could or have happened.

Let's take a small example for instance. A to-do app. A concept that has been made and remade thousands of times to add a small niche logic that some person thought could be useful. There's big players like Todoist, TickTick, Google Tasks that have the sole purpose of being a platform to indicate with todo items. Then, there are big players like Notion, Obsidian and Miro which include the concept of todo lists and todo items with very similar logic. Just a quick google search show a small snippet of all the todo apps that exist.

![Todo list list on goggle](//images.ctfassets.net/jazmsg2p4oit/4kuxrG01Sa83Z0F5hNckNR/28e8a41328d78398b7cfa2003b2b4626/image.png)

Or, if you wanted you could ask an AI for the same information 😆. Here is a (not so) little conversation with Claude, my favourite GenAI right now.

__Birat__: What are some of the top todo apps? How many todo apps do you think people have created? Shorten this to < 50 words so I can put in my article.

__Claude__: Todo apps abound, with standouts including Todoist, Microsoft To Do, Things, TickTick, Any.do, Notion, and Google Tasks. While the exact number is unknown, thousands exist across app stores, ranging from simple checklists to complex project management systems.

This leads me onto my next point - GenAI is essentially a very intelligent search engine. It will search for the information you want, how you want and will tell you the information in the format you desire. Saying these capabilities are not useful would be foolish, narrow-minded and dismissive. With technology moving so fast, both people in and outside of tech require help to navigate, learn and keep up with...everything.


A blog post to defend the use of generative AI to write sophisticated code from someone who tries to limit his use of generative AI.

Vibe Coding - Is it really that bad?

John peeks out of the window. Looking at the leaves on the evergreen trees beside the sidewalk on Lennon Street. A leaf falls. It glides back and forth with the wind as the curved fresh leaf slowly falls flat into the ground. Another tree, this one without leaves is shaking. The wind is slight, not so much that the movement of the bare tree is obvious but anyone looking at the tree would notice. John noticed.

He lets out a cough. He'd been sitting by the window sill for so long that he hadn't realised it had been nearly half an hour. He let out another grunt to escape the slight discomfort of his dry throat. He takes a sip of his black coffee. Bitter. Bland. Black. Americano. It's gotten colder. It's not the burning heat that once was, the liquid had turned into a pleasant warmth that reverberated throughout John's chest. His body shuddered at the aftertaste - maybe he'd put in just an ounce more coffee than he had wanted. Or, maybe he had been missing a kind of warmth that everyone's body needs. The kind of warmth that makes you leave your house. The kind of warmth that makes you feel content. 

John let out a big sigh, shuffling and shaking his legs. His mind started wandering - his hands started fidgeting. His eyes were watching every little movement or - lack of movement. He felt a sudden onset of this restlessness, and wondered what his body was trying to tell him. It felt like he needed something. Necessity, Longing, Urge, Yearn. Whatever word you wanted to use. His hands imagined the spotty texture of the camera. His voice demanding to travel to a microphone. His mind wandering through the infinite possibilities of the future. Every part of him need do this. Every part of him needed art. Every part of him needed to create. Something.

Creation

> C3: A term I coined for the constant cycle of a corporate job. 

I've never understood the concept of what people meant when they used to say "I'm stuck in a corporate job". That never translated to me because I never understood the word "stuck" in that sentence. Maybe I was a child, too naive and privileged but I would always respond with something like "Why can't you just quit your job?".

Now I'm in a place where I feel like I'm "sorta kinda" understanding this phrase but once again my perception is still the same. If you're stuck, you seek solutions to be unstuck. Therefore, if you're going to be doing the same cycle of corporate work, you have to get unstuck somehow. Whether that would be taking on a new project which you enjoy or, getting promoted to get a different intensity of work. The solution to being stuck is change. It's movement. Therefore, if there is no change and no movement, you feel stuck. Simple, right?

Not really. There are multiple ways you can be unstuck but it is mostly never easy to be unstuck. For me, a 21 year old, working in Dublin after moving out of London, it has not been easy to move out of home. I'm one individual. I don't have any dependents. I haven't worried for more than a few days if not weeks, whether I won't have a home (different story). I have the option to move myself back to London by taking a leap at work. But not everyone does.

Now, imagine an individual where their family depends on them for financial stability. You go to work every day, whether you like it or not because your family needs you to go to work. Without you going to work, your family will struggle to make ends meet. In this situation, the individual might feel the pressures of C3 but would most likely never have the option to say, "I'm not enjoying being stuck at work". For these individuals, it is a journey of acceptance of C3. That journey is the only way.

Some people might argue the individuals I mentioned above still have a choice. They still have the option to change, to get out of the C3. Simply put, if those individuals have a choice they don't fall in the above category. I know this because my dad falls in the category above. He has worked his absolute ass off to provide for his family and to continuously maintain all of our growth.

Next, imagine a person who has moved thousands of kilometers away in hopes of a better opportunity. An individual whose family, while not legally dependent, depends on him to spend money overseas. An opportunity where they work in a room that is temperature controlled, thinking about abstract numbers at work instead of laying bricks underneath the scorching hot sun. An opportunity where they get payed 10s if not 100 times more than in their country of birth, for the same intensity of work. This person doesn't seem to have a choice. They do, but in their mind the choice is undebatable. They __have__ to do this.

Again, some people might argue the individuals I mentioned still have a choice. These individuals might have more of a choice than the initial set of individuals but their choice is still limited. They're in the hunt for the same amount of money and the C3s. I know this again, because my dad fell in this category.

That's the end of the essay. Some of you reading might be confused - the purpose of this essay was not to give a solution, it was to provide insight for those people that have promoted hustle culture, without trying to understand why hustling in a corporate job isn't easy. Why working 2 jobs daily shouldn't be normalized and *one part* of why it is not easy to get out or rid of the C3s. Now, I leave you with a meme that I found funny.

![Meme of C3 corporate](//images.ctfassets.net/jazmsg2p4oit/31rcpEtPTBt5izcfTlB5w7/5e50ba91d96c105de49104331d577e3e/image.png)


A insight into the difficulties of escaping or getting unstuck in the cycle of constant corporate jobs.

C3: Constant Corporate Cycle

Hi there! My name is Birat Koirala and I like to solve problems. 

## You sound boring, is that it?
Okay, okay let me tell you more about myself. I like video editing. And playing games. And occasionally play guitar and singing badly. I like playing football and breaking ankles (figuratively, you boomers). I like chilling with my friends. I also like getting deep with it. You want to talk about why I think that life forms scale infinitely? Let's do it. You wanna talk about what makes J Cole the 🐐? Let's do it. You wanna talk about how beautiful the sunset hue on the clouds are? Let's get it.

I have a passion for learning and am constantly seeking out new challenges and opportunities to grow. Not just as a developer, but as a person. In order to do this as a software engineer however, I felt that any experience I could get with software in the real world would definitely come in handy.

So that's what I did. I applied to internships and more internships and... you guessed it, more internships. After a long year of applications, the wonderful people at [Fundapps](https://fundapps.co) reached out to me, asking if I was free for an interview. I immediately said yes and two interviews later, I had a internship offer lined up. I was on top of the world.

## So, how was Fundapps?

At Fundapps, I learned a lot of things. First and foremost was production level code. I'd barely even gotten introduced to the difference between "prod" and "dev" in my own time, so I had no idea what to expect. But, I persevered. I tried to learn as much as I could, whilst not knowing the impact this internship would have in my life.

![Pikachu Meme](https://images.ctfassets.net/jazmsg2p4oit/3DzHhU8BAghpqMBZrF5plg/5ea791750dd501538f13d552f7810ad6/image.png)

To my team at Fundapps, if you're reading this, absolute big shoutout to every single one of you. My manager, my manager's manager, and especially my buddy who believed in me, pushed me and encouraged me to keep going, no matter the challenges I faced. I had an ABSOLUTE blast. I learned a lot, which is what I wanted to do all along. I created friends and it barely even felt like work. 

I knew that my time there was potentially limited so I wanted to learn as much as I could. But also... I wanted to fail. I'm a big believer in iterating to become a better person and falling a lot lot lot of times before learning a lesson. So I got to work and went to town on..... the backlog tickets 😈. I was pushing tickets from the backlog left, right and centre. Sometimes, the tickets from backlog became bigger, and sometimes the tickets had already been fixed. But that didn't matter. 

What mattered was I got to enjoy the best of what engineers do day to do - talk to each other about how to solve a problem, and again, and again. I felt like I was a part of the company. I started off as an intern, but by the end of it I felt like an actual engineer at FundApps. That's how I know it was a success. (and also because I knew if I came back here in the future, I would be happy)

## Okay, you got an internship, big deal... tell us what happened next!
Then came second year of university, where I made many more lifelong friends than I did first year. Only in hindsight do I recollect the impact that COVID-19 had on all of our lives. It was an insane time. I also applied to more internships, I had the return offer from Fundapps under my belt and it was a place I would love to return to. But I wanted to keep learning and see what it was like at a different company. A company with a different scale of audience. Luckily enough, on a random day on March 14th I received an offer from Amazon to go to Dublin. I was ecstatic, overjoyed, filled with purpose. No words could describe how I felt at that moment because it felt like all the work I had done so far had been rewarded.

Skipping to the internship though, while the initial feeling itself was a reward the experience of moving to Dublin (for the first time) was not an experience I want to experience again (and hence, I'm moving to Dublin for the second time not the first😅 this summer). It was difficult to say the least - the housing crisis in Dublin was worse than I thought. Finding an accommodation seemed an absolute impossibility. And it was really nearly impossible.
(To be continued)

The world is a beautiful and all consuming place. So my plan is to take it slow, not get caught up in the intricacies of it all. And hopefully, have an okay-ishly renovated shed by the end of summer :)

This is an introduction to me and the type of person I am. I've tried to put as much character and "myself" into this post, so that anyone reading can virtually get to know me.

Who am I, really?

This is just going to be one of the places where I start to take down notes on what I have been doing in my individual project. It's been a tough ride up to now and I'm sure it will get more tough from here. First and foremost, the initial title I went with was as follows:

> Simulating cancer immunotherapy using an eventually consistent distributed system.

I was super happy with this title because it covered a lot of what I wanted to do. Distributed systems is something I find interesting to this day. The topic of "cancer" intrigued me because I'd always thought about it as a disease that has some cause and we can find and cure it. If this project would help me even a tiny bit then I would absolutely love that. I've always had a curiosity in biology and the medical sciences too.

![helpingWithCode](https://www.explainxkcd.com/wiki/images/0/01/here_to_help.png)

This all sounds great but I needed to make a start. I needed to find resources that would help me solve this issue. I wanted to do this step by step, in steps which I thought were calculated. I wanted to focus on the viability and the methodology of how I could **simulate** cancer immunotherapy. For this I started to read some papers online, I tried to find different ways in which I could address this problem. In doing this I came to two realisations:

1. This could turn out to be more of a Machine Learning paper rather than a distributed systems paper. I say this because the simulation requires handling multivariate data. 
2. This was a bigger project than I had initially thought. Finding papers for it were difficult and there were different branches of knowledge I had to cover before attempting to understand some relevant papers. With this in mind, I needed to boil down the question to something I can do within my viable 6-7 month period. 
![image](https://imgs.xkcd.com/comics/dubious_study.png)

Both of these realisations pointed to the same thing that I had to do. *Remove distributed systems from the question* (for now). With this my new title became: 
> Exploring different methods to simulate cancer immunotherapy

So that's where I went with it. I explored. I saw what was out there and thought about that. After a lot of research the **first** method became apparent - Ordinary Differential Equations (ODEs). 

# ODEs to simulate cancer immunotherapy
A paper I found, with title [Modeling immunotherapy of the tumor – immune
interaction](https://deepblue.lib.umich.edu/bitstream/handle/2027.42/42105/285-37-3-235_80370235.pdf?sequence=1), resembled somewhat the general "thing" that I wanted to do. Page 237 of this Journal had the following ordinary differential equations:
$$
   \frac{dE}{dt} = cT - \mu_2 E + \frac{p_1 E I_L}{g_1 + I_L} + s1
$$
$$
   \frac{dT}{dt} = r_2 (T) T - \frac{aET}{g_2 + T}
$$
$$
   \frac{dI_L}{dt} = \frac{p_2ET}{g_3 + T} - u_3 I_L + s_2
$$

where $E$ is the number of effector cells i.e. the cells in your immune system that fend your body from any harmful pathogens, $T$ is the number of tumour cells and $I_L$ is the level of cytokines. The equation, initially is quite hard to understand and it contains a lot of different variables and constants so we're going to take equation in it's simple differential form. Each equation represent the growth or decline of Effector Cell, Tumour Cell and Cytokine levels. 

> "...the cells that orchestrate the immune response use cytokines as signaling mechanisms for immune-response stimulation as well as lymphocyte stimulation, growth and differentiation." 

## So, how do effector cells and cytokine work together?
Let's break down the answer. Into diagrams.

### Step 1: Identifying Pathogens/Tumours

![Cytokine Process - 1](https://images.ctfassets.net/jazmsg2p4oit/3iJpD8oY8TWS9TObOFdRO0/9b903a363a5d50667aa172d075f28cbf/image.png)

First, the cytokine (or $I_L$) is injected into the area with tumours. The presence of $I_L$ tells the Effector Cells that there is something wrong in that area and therefore, they must deal with it.

### Step 2: Dealing with it

![Cytokine Process - 2](https://images.ctfassets.net/jazmsg2p4oit/d4HsZz8iD7RBHf2sAfy6t/82dc0a526b7eba3123fe754dda4010ec/image.png)
As you can see in the left section of picture above, the effector cells divide and grow rapidly to combat this pathogen wherever the Cytokine is located (in this case Location 2). The location of the cytokines tell the effector cells the general area of which the pathogen is located. The effector cells then attack and destroy the pathogens (and potentially, some healthy cells in the process too, eek). After they're done, that location has no more unhealthy cells (as seen in the right section of picture above).

# Okay, but now what do we do with the equation?

We use the equation, plug in some values, and see which one leads to the eradication of the tumour and which one leads to something else. Simple. Right? Not really. At the time of doing this, the first thought I had was -  *What are "some" values?* The differential equation had a number of distinct variables that needed a specific value in order to calculate the growth rate. In no particular order, these were: $E_0$, $T_0$, $I_{L_0}$, $c$, $\mu_2$, $p_1$, $g_1$, $g_2$, $r_2$,  $a$, $\mu_3$, $p_2$, $g_3$, $s_1$, $s_2$.

Fortunately enough, the same paper had some information as to what values should be fed into some of the variables.

![paperImage](https://images.ctfassets.net/jazmsg2p4oit/5qJWLpPZ6QXTVpJ0VgNEZf/565f5d6162e84e3aea74a5ff183eaacc/image.png)

However, this isn't enough. There is no indication for the values of $s_1$ and $s_2$ in this table. Upon further reading of the same paper, (on page 243) we find that:
1. The model (or the set of differential equations) considers **either** or **both** $s_1$ and $s_2$ to be non-zero.
2. $s_1$ represents the addition of **LAK** or **TIL** cells to the tumour site.
3. $s_2$ represents the addition of interleukin-2 (IL-2) to the tumour site.

> (According to Wikipedia) **Lymphokine-activated killer cell (LAK)** is a white blood cell that has been stimulated to kill tumor cells. **Tumor-infiltrating lymphocytes (TIL)** are white blood cells that have left the bloodstream and migrated towards a tumor.

![image](https://imgs.xkcd.com/comics/data_trap.png)


This is a blog about what my thought processes were, alongside links, equations and images to explain the my dissertation paper and/or processes.

My dissertation process - Individual Project

So this is my first ever post. This is extremely cool. This will be going on my website for which I will have a link that says blog post. That website will be entirely about me. For now that is all I'm gonna say. Just below this, I'm gonna test out a few things to see how they render a page. This is a test for rendering, pretty much. See you guys on the other side.
__Bold Entry__
*Italic Entry*

```java
    public static void (String[] args){
      System.out.println("This is the code entry");
    }
```

[Link Entry](https://google.com)
- First item in List
- Second Item in list

| Table     | Heading 2 |
| ---------- | --- |
| First       | First right |
| Second       | Second right |

I also wanted to test out equations:

$$
\begin{equation}
1 + 2 = 3
\end{equation}
$$

$a^2 + b^2 = c^2$

Thank you,
Birat


Birat Koirala

Vibe Coding - Is it really that bad?

Creation

C3: Constant Corporate Cycle

Who am I, really?

My dissertation process - Individual Project

My very first blog post