Malaya Rout.png

The road less travelled

Large Language Models (LLMs) are revolutionising the way we work, read, learn, and live our day-to-day lives. It is to the extent that psychiatrists call something “AI psychosis”. It is a state where one loses touch with reality and becomes increasingly immersed in the virtual world of generative AI. LLMs are proving to be the…

Read More
Malaya Rout.png

A talk to motivate

I was invited to speak and motivate a newly joined batch of students in a college last week. I have delivered numerous technical talks, but never of the non-technical type. I shared with them the following twenty-one things I have learned over time. I know that it is easier said than done. I guess I…

Read More
Malaya Rout.png

The bias variance tradeoff

The bias-variance trade-off is crucial for achieving an optimally performing supervised learning model. Bias is the difference between the average prediction of a model and the actual value. The model with high bias oversimplifies the model and pays little attention to training data points, thereby leading to high approximations and, consequently, high errors on both…

Read More
Malaya Rout.png

Are we living the dream?

Are we living the dream? You live your dream when you have interesting things to do that require two lifetimes. You live your dream when you wonder why your lifetime couldn’t be more than a lifetime. You have a question. You find an answer within seconds. You have hundreds of pieces of content to read,…

Read More
Malaya Rout.png

I see, I hear, I speak, I read

I am amused at how we have started referring to traditional AIML as “traditional”. I am equally amazed at the presence of “traditional” LLMs. How fast do you want us to move? The so-called traditional LLMs are entirely text-based. The not-so-traditional LLMs are multimodal by nature. They handle images, videos, audio, as well as textual…

Read More
Malaya Rout.png

Just you and me – A rambling

You don’t like me when I am too small. You don’t like me when I am too big. Two decades ago, I never thought I would be so valuable to humankind. I know, sometimes you get jittery looking at my capabilities, fearing that the systems using me will replace you. But don’t worry. I am…

Read More
Malaya Rout.png

Are we already in the future?

We are a future away from the future. Let it come at a carefully treaded pace. It has been nearly two-and-a-half years since we saw the first large-scale and democratised usage of large language models (LLMs). I agree that the world has not been the same place. We have referred to the times as pre-LLM…

Read More
Malaya Rout.png

ML and LLM: Where do they Meet?

Today, I will discuss five key intersections between machine learning and generative AI. I have previously spoken about these topics at conferences, but I am writing about them for the first time. A large language model is a machine learning model A machine learning model is the approximate relationship between the input variables and the…

Read More
Malaya Rout.png

My LLM is the best

Two steps forward and three steps back / I long to run ahead / In the rat race / I long to run fast / In the rat race. The winning rat I want to be / The running rat I want to be. / The tireless rat I want to be /  The fearless rat…

Read More
Malaya Rout.png

The Payasams in our lives

He was six feet and 4 inches tall. Doors were not usually built considering people of his height. He was having to bend his head and get inside the house. The family always had a tough time finding the right shoes for him. He was precious to his parents and his elder sister. Kailash was…

Read More