Why shouldn't bubbles just burst?

Oct 10, 2025

If you saw the title and wondered, “Why am I reading about bubbles?”, I envy your internet usage because there is still a child-like wonder in your curiosity. If you thought about much serious stuff like the state of our economy, I still envy you because you care about stuff. Anyway, I envy you for reading long-ass articles on the internet.

This one has three parts. Part one, where I try to wriggle my way through an analogy for seeing the world through bubbles. Quite abstract and something I had real fun cooking up.

Part two is where I use this analogy to squirm around AI. It’s okay if you just rolled your eyes when you read ‘AI’. Even I was rolling my eyes while writing this (now my eyes hurt). Of course, it has to be about AI. Everything on the internet has to do with AI right now, in some way or another. You can’t write anything on the internet without connecting it to AI, or else the content police will come knock on your door. Also, my mind just jumps around how I feel about it anyway. So these are mere attempts at understanding my own views on them.

Finally, part three, because there needs to be a conclusion of some sort.

Also, I am writing this after four long months. Went cold turkey in June and now I am trying to restart it. Thanks everyone who has been reading what I write. It’s honestly quite encouraging.

If you are new here, thank you for checking it out. I write about too many things. Most of them are quite interesting. You can read some of my older ones and maybe consider subscribing? ¯\_(ツ)_/¯

If you feel this article is a bit too long, take a break before part 2 and comeback later to part 2 and 3.

So let’s get into it.

Part 1: Bubbling systems

I think one of the earliest things that we saw getting created with our hands and even got to be part of that creation were soap bubbles. Splashing water, making frothy bubbles. Probably one of the early lessons on ‘hey, you can make stuff with your hands’.

If you think about it, making bubbles is the closest you get to experiencing magic. The kind you find in some of those fantasy stories. You blow air into your hands, and out of nothing a transparent yet iridescent sphere comes alive.

A bubble is formed when three things come together. A material surface, usually liquid, with a pretty strong and elastic surface (one with enough tension). Something to stay enclosed in this boundary, usually air or some type of gas. Finally, a good amount of energy to stretch this film from within and allow the air to stay inside.

My point being, bubbles are simple yet quite complicated, and they are all around us.

This complexity just increases exponentially and blows out of proportion as the number of bubbles increases and they are connected to each other.

The best shape for a single bubble to exist is a sphere. The second best shape is to be connected to another one. When there are two connected bubbles, they form a straight wall between them. Make it three, and the shape gets even more complex. It’s so difficult to understand the shapes of interconnected bubbles that it was only in 2022 that mathematicians were able to calculate the shape that three interconnected bubbles will form. The funny thing is, bubbles love to form clusters and these complicated shapes.

Now here is the big analogy time. I think everything is a bubble. Our world is a complex set of interconnected bubbles, forming an intricate…foam structure.

Keeping the individual at the centre, there are two types of bubbles. The categorisation is based on the source of the bubble. Bubbles that have their origin outside of us, and the ones that are formed within us.

These externally originated bubbles include things like the places we have lived and where we live. The house, the neighbourhood, the apartment, the state, the country, and so on. It also includes how we travel. The car is a bubble; the different types of public transport are their own bubbles. The environment we exist in becomes its own externally originated bubble, just like the economy we live in, the technology, the media, and so on. Essentially, external bubbles are places, people, concepts, and things that have their origin outside our minds and inside someone else’s.

Then there are bubbles that come from within. Our identities, personal philosophy, mental models, personal ethics, morality, beliefs, relationships, even our fears and aspirations.

For each individual, the bubble behaves differently. It has its own reality. Time, space, knowledge, and even laws of nature behave differently in each of these bubbles. Also, it changes from person to person.

Some of these bubbles are big, some are small. Some are extremely short-lived, while some survive as long as the foam exists. Some may live longer even after that. Some bubbles burst in this interconnected foam structure, leaving space for other bubbles to pop up. While some just shrink down. Many smaller bubbles may even combine to form a bigger bubble. If you look close enough, a bubble may not be even just one bubble; it could be a collection of many smaller ones.

Sometimes bubbles give us a sense of safety. A momentary bliss to live in this world and not go completely mad. Just like how the cozy corner of our house is an external bubble that makes us feel safe, our identities and beliefs are an internal bubble that allows us to do things and sleep at night.

In other cases, the bubble stops us from being something or somewhere. A thin, burstable wall of exclusion. A transparent yet real wall with the sole purpose of keeping things outside. Gender, caste, nationality, beliefs, and so on. You get the drift.

The stretchy surface of these bubbles expands to a limit, and the elastic surface that forms the boundary gives up. The perfect example for the system archetype: ‘limits to growth’.

On a side note, this is my contention with people who say there are no definite boundaries and everything is fuzzy. There are boundaries. It’s just that the bubble is clear and transparent. Now and then, the iridescence of the bubble gives us just a small glimpse of the multicoloured boundary. I think the inability to see the boundaries is because of our inability to imagine complicated related systems. Maybe seeing it as a set of interconnected bubbles makes it easier.

Of course, none of this is a novel thought. If anything, this is just an unholy Frankensteining of concepts like systems thinking, filter bubbles, echo chambers, cultural capital, and so on.

Part 2: Intelligent (?) bubbles

If you find the word bubble on the internet, more than 90 percent of the time, it will be followed by the word ‘burst’. I think the alliteration makes us gravitate toward this pair of words. Right now, the word that precedes ‘bubble burst’ is AI.

Last month, the alpha male of the AI cult, Sam Altman, said that the AI bubble is real and it is about to pop.

“When bubbles happen, smart people get overexcited about a kernel of truth,” Sam Altman.

The ChatGPT, antiAI, and Singularity subReddits went nuts. Some said he has lost relevance, and some even welcomed the bubble to burst. In true late-stage Silicon Valley hype machine fashion, the conversation around AI saw an uptick again.

(Tin foil hat warning) Again, like everything around AI, it just feels like the discourse on the AI bubble is quite planned out by the handful of people controlling this technology. The conversation is no longer around stealing from artists, it isn’t around energy consumption, it isn’t around privacy, it isn’t around misinformation, and the conversation is definitely not around studies proving AI’s impact on our ability to think. We are talking about money that is going to evaporate soon.

It didn’t take too long, and soon after Sam Altman said AI is a bubble, others like Jeff Bezos have also called it a bubble. What’s interesting is, whenever any of them say it is a bubble, they are talking about others and never about themselves or their company. In fact, if the economic bubble bursts, the ones who are going to benefit the most are these handful of people. These megaliths, like Microsoft, Amazon, and Google, will be the ones to survive the burst and then control the stable world. Conspiracy over.

Warned you to get your tinfoil hat.

The argument for the AI bubble is that AI companies are going through crazy hypervaluation. Money, including VC funds, government investments, and even retail investments, is being pumped into anything and everything AI. A déjà vu of the dot-com bubble. These numbers, with crazy amounts of commas, have been pumped into all aspects of AI. Infrastructure, components, talent, and even marketing. I was surprised to see a humongous billboard of ChatGPT in Mumbai. Just as I was looking at it, a gust of wind blew, and it sounded like someone screaming, “irony”.

Keeping the economy to the side for a minute, it’s all quite a wreck anyway. What I am trying to see here is a sort of cultural bubble around AI. Even that seems like it is about to reach its limits.

What is a cultural bubble?

Cultural bubbles are all the different cultural elements, like jargons, memes, arts, discussions, discourse, behaviours, rituals, practices, content, and even ideas or beliefs, surrounding a particular piece of overarching thought. The thought becomes a clear film that holds all these cultural elements inside, and the energy to blow the bubble is the effort put in by people in creating and maintaining these elements. These culture bubbles are generally a cluster of interconnected bubbles that originate both inside and outside of an individual.

Since all these cultural elements are part of the air that’s inside these bubbles. Even when the other side of the discussion comes into the picture, that is, the ones who say AI will solve all the problems or even the ones who are not able to pick a side (looking at a mirror here), the cultural air is being pumped into the bubble. Soon, the bubble will reach its limits and it will burst.

So once more, how is this a bubble?

Like the bubble theory states you need three things for bubbles to form. An elastic film or surface, air or some gaseous element inside, and the energy to blow this into the elastic film.

If you take the different places AI is being used (or rather promised to be used) as the soapy surfaces for all these different bubbles, the air inside is all the memes, jokes, discussions, content, and even the hope surrounding it. The energy is not just people who are trying to put AI in everything, but also people who are warning against it.

Take AI in therapy, for example, this article from HBR titled ‘How People Are Really Using Gen AI in 2025’ says therapy is the most prominent use case. On papers and forecast charts, this looks picture perfect. Why pay a therapist 1500 to 2500 rupees for one session when you can get ChatGPT Premium for 400 rupees and let it be your therapist? Along with helping you do your work, learn new things, and apply for jobs, it can also listen to you vent and read through the whole world of stolen data to give you a diagnosis, all the while making you look like you are a Japanese cartoon character.

Many aggregator apps of the therapy world, basically uber for therapists, tried (a few are still trying) to push AI therapists. I remember seeing some of these therapist-supplying companies making a ChatGPT wrapper with some cringy name and then some posters asking employees to use it. Six months later, the same people are saying, robots can’t understand humans as well as other humans. What changed? the cultural elements. Slurs like ‘clankers’ and ‘askholes’ entered the vocabulary, ChatGPT sycophancy ruled the middle of this year, and now AI-induced psychosis being the latest addition.

You might say therapy is a stretch, too abstract, too ambiguous. Okay, then let’s take the basic purpose of AI. The one that people have been shouting from the rooftops on LinkedIn, the holy grail of the modern world, and the first word of a baby born in 2025:

“PRODUCTIVITY”

From the team that brought you ‘Shrimp Jesus’, ‘Ghiblified war images’, and ‘near-perfect scams’, comes ‘workslop’.

Workslop is essentially the daily corporate work equivalent of social media AI slop. Stanford Social Media Lab and BetterUp Labs have come up with this jargon while trying to understand why AI is not solving the problems it had promised to. In the article titled ‘AI-Genrated Workshop Is Destroying Productivity’ the researchers mentions that;

“We define ‘workslop’ as AI generated work content that masquerades as good work, but lacks the substance to meaningfully advance a given task.”

Workslop, a new phrase and thus a cultural element, added to the bubble.

This is just one side of the equation. The other side says, well, the opposite. That AI truly is going to redefine productivity itself.

Not to sound like a complete doomer, AI has proven to be useful, or rather, has been quite useful for some time, especially at work. There are people who use complex workflows to manage their daily work, and there are also those who like autonomy and use these AI tools sparsely but effectively. In a world where 9-5 has become more of a side hustle, AI helps people survive (a bit).

Personally, I use AI in some or other shapes and forms. From adding fluff to random documents when the feedback that I get is to add more pages, to figuring out where to start if I want to start learning critical theory.

At work, Copilot helps me ramble on and convert these rambles to coherent design docs and user stories. Taking notes on Teams was a pain. Copilot does that for me, and I have receipts when I want to say “but you said something else in the call”.

The other day, I tried Copilot on Excel to help find the optimum strategy for taking the remaining leaves I have. I gave it my remaining leaves, all the leaves I have availed till now, and holidays that are coming up. Then I asked it to suggest the best leave strategy, keeping in mind I don’t want to take more than two leaves a month, optimise for long weekends, and my mental health (four weeks without a break in the middle is a no-no). I thought this would be simple and straightforward. Yet, even after three attempts, it just gave me extremely bad results. In one case, it completely ignored everything and just suggested that I take all my leaves in one month and make it a complete vacation. Maybe Copilot needed a month-long vacation from doing all these random tasks people have been asking it to do.

Sidebar: if you find these last few paragraphs a bit too salty and not balanced enough for your taste, it’s just that we are being forced to adopt these AI technologies like they were helpless little kittens on the roadside that promised to cure cancer, end world hunger, and fix the global mental health crisis. Turns out it has just become an excuse to lay off workers and burn the planet, all the while actively trying to make reality obsolete.

Nonetheless, the content inside these bubbles is increasing at a tremendous rate, and the limits of growth for these bubbles are being reached at a crazy pace.

Part 3: So what?

Just like the economic bubble of AI is ripe for bursting because of the hype machines, the cultural bubbles are also there.

But we like to imagine things to be dramatic and not boring, especially when it is to do with destruction.

I don’t think we should see AI as a single bubble, even with the definitions of a bubble from an economic point of view. Yes, the tech companies injecting AI into our veins are hypervalued on the stock market, but there is no burst of any kind that’s going to cease their existence. That’s why we have to look at these as sets of intricately connected bubble systems.

The AI bubble, if you look closer, is made of many economic as well as cultural bubbles. The image generation bubble, the AI writing bubble, the AI therapy bubble, the AI coding bubble, the AI productivity bubble, and so on. When you take a step back, these bubbles are also connected to a larger system with the environment, power consumption, cheap labor, and so on.

Like any bubble system, the AI bubble is also not immune to the system archetype Limits to Success. As it keeps expanding, the elasticity of the bubble will buckle, and it will burst. But since it is a set of connected bubbles, if one bursts, the others need not necessarily burst as well.

So no bubble burst is going to wipe AI out of our lives. The bubbles settle down, and few will remain; maybe even new ones will pop up. They will then control the transparent wall of an iridescent exclusion.

So no, the AI bubble burst is not going to be dramatic. Like the apocalypse, it is going to be utterly boring.

These economic and cultural bubbles and their eventual burst have become so common to us. The dot-com, housing, crypto, NFT, and now AI. I think, for the last twenty-five-odd years, humans have been scraping the bottom of the barrel for some hope, and a handful of humans have figured out how to sell this idea of hope and make money.

I started this thought with quite a cynical approach of looking at everything around me as bubbles that are destined to burst. But the cynicism is also an iridescent spherical bubble that invokes a childlike wonder.

I guess the fun is not just when the bubble forms, but also when it pops.