Ghost Stations

While the students protests gain momentum with support pouring in from various quarters, the city sits on edge. 

Restrictions on movement have been implemented. Metro stations in the  sensitive areas have been identified and have been marked as no man's lands.

Metro trains pass through such stations without stopping. The pace becomes slower than usual though.

On my daily commute, I have to cross a few of such stations.

Most of the stations are below ground level; you get an eerie feeling about the whole atmosphere - dimly lit lights, not a single person to be seen waiting or walking on the platforms, the cordoned off stair cases and most importantly, the doors don't open. The feeling can be extremely unsettling. 

Mobile network also goes in a prohibited zone for a certain length of time. No calls, no internet access. 

I wonder and ponder that, though the commute rush is a matter of irritation, it would be worse if it were to disappear suddenly!

Cloud Computing: Beyond the Buzzword

One of the hot questions these days is: "Hey, are you on the cloud?" or "Is that website cloud-based?"

Someone would even say it's not worth it if we are not on the cloud. Reminds me of one of the advertisements that aired on television in the mid-90s with the tagline "mooch nahi, toh kuch nahi"!

And if someone says the cloud is unknown to him or her, then that person is not considered a homo-sapien!

Jokes apart, let us try to understand what exactly a Cloud is. Many of us use this jargon daily without knowing the core concept behind it. It has become extremely fashionable to use this term quite often in our daily conversations.

So, what is a Cloud? In simple terms, companies or individuals can rent resources on an hourly basis, instead of buying physical servers. This hourly model has changed the IT economics.

The modern cloud era began 20 years ago; Amazon Web Services (AWS) launched its cloud services in 2006, with a handful of services like

Virtual Servers (EC2)
Cloud Storage (S3)
Databases
Networking

On the current day, the list has grown considerably, and AWS has added more than 200 solutions to its formidable repertoire. Each serves a different purpose.

Now, a layman would say simply host the website on the cloud, and everything will be automatically taken care of.

Is that so? The answer is NO.

The cloud ecosystem has its own caveats. Agreed, the cloud comes with numerous solutions and has a lucrative by-the-hour pricing; however, there is much more concealed than meets the eye.

Suppose one is designing a solution that requires multiple components. Each component has a price attached to it. If you want to save your time by automating the backup process of files, folders, etc., you have to pay extra.

The monthly billing of resources also tends to be ambiguous. Since every service is priced differently—and many services interact with one another—the final monthly bill can be difficult to predict without careful monitoring.

So, when someone says the cloud is cheaper, please don't believe them. They are incorrect in their assessment.

People are also of the opinion that if a solution is on the cloud, it's safe and secure by default. Again, an incorrect assumption. The safety and security of the data that users are putting on the resources depend completely on the owners of the resources. The cloud will only provide the tools; it's up to individuals how much they are willing to spend on the tools to safeguard the resources. Additionally, one has to be on one's toes and run security patches from time to time.

The AWS Shared Responsibility Model explains the above, that security is shared between AWS and the customer.

Though AWS still continues to be the preferred choice when it comes to deciding about cloud infrastructure, it faces stiff competition from vendors like Google, Microsoft, and other smaller players.

Though it has become very cloudy, many end users are migrating back to traditional servers due to price ambiguity and other complexities.

So, the cloud is definitely not the silver bullet that one may think of it!

Heartbreak

The other day, I was on a busy metro station platform, waiting for the train. The platform was buzzing with people all around. Suddenly, my attention was drawn to a young girl talking on her phone, tears rolling down her cheeks. She looked deeply saddened and dejected.

"Seems to be a case of heartbreak. Maybe she had a fight with her boyfriend." One of the bystanders commented lightly.

At that very moment, it dawned on me, is heartbreak only confined to matters of love and longing?

There can be, and there are multiple reasons that actually break the heart of us humans.

• A failed interview session.
• Serious illness / passing away of a near one.
• Unfulfilled promises.
• A denied promotion
• Losing something important 

and many more...

Heartbreak has many faces; the magnitude varies

You Type a Prompt. What Happens Next?

Like millions of users, I have been using Generative AI - especially Chat GPT - regularly to cope with my routine tasks. Have we considered what happens between giving a prompt and receiving the required response? What actually goes on behind the scenes?

The entire episode is nothing less than a journey and a fascinating one!

Let us review the various steps of this journey:

The moment we type a phrase (known as a prompt) and press the Enter key, the Tokeniser kicks in!

Now, what is a Tokenizer? 

Think of it as a chef's chopping board knife! It chops the prompt into smaller parts - known as Tokens.

For example, if I give a prompt - "What is the capital of India?" - the Tokeneiser splits the phrase in the following format:

["What","is","the","capital","of","India","?"]

The tokenizer decides the most appropriate split based on its algorithm and vocabulary.

These chopped pieces are Tokens.

Then the Tokenizer refers or looks up in a pre-defined library known as the Vocabulary.

This library has a numeric reference for each token. Something like this:

"what"       → 125

"is"   → 864

" the"  → 5421

"cpaital"       → 13

"of"       → 99

"India" → 199

"?" → 09

As soon as matches are found, the Tokenizer converts the tokens to corresponding numeric values - Token Ids - to be more specific.

Then comes the Embedding layer. 

This layer converts all token IDs into a vector format.

Token: "India"

Embedding Vector (illustrative)

[0.83,-1.42, 0.91,2.13,-0.37,1.78,-0.55,0.24]

This list of numbers is called an embedding vector. It is the mathematical representation of the token "India."

The embedding vector is a learned numerical representation of the token that captures statistical relationships with other tokens

Transformer

The transformer can only understand vector-based inputs; it doesn't understand texts, unlike humans. It neither comprehends the token IDs. The Transformer then processes these vectors using its self-attention mechanism to understand the relationships between tokens, refine their representations through multiple layers, and predict the most appropriate next token. By repeating this prediction one token at a time, it gradually generates the complete response.

This response is again in the form of token Ids.

The predicted token IDs are then converted back into human-readable text by the tokenizer's decoding process, producing the response that appears on our screens.

The response is rightly termed as closure.

Contrary to popular belief, an LLM doesn't think of the complete answer first. It predicts one token at a time. After generating each token, it repeats the entire prediction process for the next one until the response is complete.

To us, an LLM appears to understand language. Under the hood, however, it is performing mathematics on vectors, predicting one token at a time until a coherent response emerges.

So, imagine millions of requests and responses going through the same pipeline daily!

 

The appearance doesn't maketh a spy!

 So..coming back here after a hiatus due to multiple reasons.

************************************

I went to the theatre this afternoon to watch "Alpha" - the latest installment of the YRF Spy Universe. The cast has done a decent job, though I feel the storyline could have been better. 

Bobby Deol eerily looked like Stephen Lang for most of the time.

I don't intend to write a review of the film - I am simply not qualified for that, and moreover, many knowledgeable critics have already done their share of analysis of the film.

So what am I doing?

Post the release, I have been hearing and reading lots of comments on social media about the comparison of spy-themed movies, especially from the Yashraj stable, with Aditya Dhar's Dhurandhar (parts 1 & 2). 

What is the need for such comparisons? Why can't undercover agents be portrayed as glamorous? 

Much before Dhurandhar or YRF Spy Universe, we had a certain Englishman, who went by the name of James Bond - always suave and dapper - and the audience used to simply love him for his antics. My father adored him; I also have liked him since my childhood.

The spies in the Kingsman franchise were always styled with bespoke attire.

Even before, there was Mata Hari, the Dutch-born exotic dancer, who used her charms and glamour for espionage during the First World War.

If we consider Hercule Poirot's or Rip Kirby's sartorial style (fictional characters), they were portrayed as stylish individuals. Though neither of them was a spy in the true sense, being detectives, they did have to resort to some sort of spying now and then.

Dhurandhar certainly did commendable business and caught the nation's fancy. It has been hailed for being "realistic". Probably it has been so; however, do we - the common people - actually know what an actual spy/undercover agent looks like? Why do people want all the spy movies to be Dhurandhar templated now onwards? Where is the fun or logic in that?

If a spy can complete his/her missions successfully, then does the appearance actually matter?

If people are so particular about realistic cinema, then they should watch movies like Ardh Satya (1983)/Kalyug(1981) and the like..

Masala films like Dhurandhar are meant to be enjoyed and discussed over lunch, but expecting such movies to become the benchmark for future films is a little too much.

For now, grab a bucket of popcorn and a glass of Coke and enjoy the show..