Software vs AI

14 Aug 2026

home  >  english  >  blog

Software vs AI?. AI is software. Yeah, kinda confusing, right?.

For this conversation, let’s assume

• Software = deterministic output (same input produces same output)
• AI = probabilistic output (same input DOES NOT produce the same output)

How do companies make money?

In simple terms,

Raw materials from Earth + Human Effort = Usable Product.

If other people want to use your product, you created value. The value could be measured in terms of money, other products or services, or labour. This is a classic Labor Theory of Value (LTV).

Ex.

• Take clay from the earth, make a pot. If more people want your pots, you created value.
• Plow the land, sow the seeds, and make rice, corn, vegetables, etc. If more people want your produce, you created value.
• Dig the soil and refine it to find or create diamonds, oil, coal, etc. If more people want your materials, you created value.

This is the only way value is created in this world. This is how companies make money. This is the base layer. There could be many layers of companies on top of this base layer. You can trace the supply chain of all the companies back to this base layer.

Ex.

• Space X is in the business of making rockets. It needs the metals and fuel from the base layer to make rockets.
• NVIDIA builds the world’s most advanced chips. It needs the silica from the base layer to make chips.

How do software companies make money?

Every software helps a business by

• solving a problem
• improving a process

In mathematical notation,

A business’s profit without software = $X
A business’s profit by using software = $Y
Additional value created by using software = $(Y - X)

(As long as $(Y-X) is positive, i.e., the business made more profit using the software, the software companies made a profit)

Software company’s share = δ $(Y - X)

(where δ is the software pricing model either share or fixed price)

So,

Software will not exist on its own. It exists to help other businesses make more profit.

Why are software companies so successful?

Make software once. Sell copies of the software to as many customers as possible.

Ex.

• Microsoft creates a Windows Operating System once, and gets paid every time a computer is sold.
• Adobe creates Photoshop once, and gets paid every time someone purchases the software.

With the innovation of the Internet and the SaaS (Software as a Service) model and Cloud Computing, selling software to many customers has become much easier.

You could literally go from zero customers to a million customers in a day. This unprecedented growth was never possible with any other business, even today. You can’t make an extra million bushels of rice, or barrels of oil in a day. Can you?

In math, this is Exponential.

So,

Software made more money than regular businesses, because of the exponential growth factor.

Why AI companies may not survive?

By AI companies, I mean companies like OpenAI, Anthropic, etc, who rent the servers, deploy their AI models, and serve the AI requests from their users.

For every request from a user (like a simple “hello”), the AI model will run on the server and produce a response. No matter how efficient the model becomes, this is the baseline for AI business. For simplicity, I excluded the model training cost, caching, distillation, etc.

In math, this is Linear.

With the current state of AI companies,

Actual cost of an AI response = $X
Cost paid by the users = $Y
Profits = $(Y - X)

(As of today the profits $(Y - X) are negative and are subsidized by VC money and software companies’ balance sheets, in a race to capture market share.)

In the future, even if the profits become neutral + a decent markup for the software companies, that would be the ceiling of this business. Just like every other brick-and-mortar business. You want to make more money selling burgers, open more restaurants. Get it?.

So,

AI companies DO NOT produce software that scales exponentially. They produce a commodity that scales linearly.

What would happen to those AI companies?

Best case scenario:

The AI companies figure out a way to stay profitable and operate like an infrastructure provider (like electricity or internet).

Worst case scenario:

When the money dries up, the big companies with deep pockets and real-world business to support the money flow would buy them at a discount (very cheaply) and continue to run the companies with a decent profit margin, still like an infrastructure provider.

What happens to AI-Assisted Software Development Process?

AI has changed the software development process upside down. Most software will be created using AI, with humans in the loop to set the product specs, review the code and architecture, etc. That’s going to stay for a foreseeable future.

I hope the AI models will become more efficient, open source, and cheaper to run.

• The big companies would host the AI models on their own infrastructure, contained within their VPC, available via intranet to their employees only.
• The indie developers would probably host a lower-end open-source model for coding assistance on their own laptops/desktops

Where else would AI bring benefits to humanity?

I can’t comprehend all the possibilities.

• Faster drug discovery
• Early diagnosis of diseases
• Invention of new materials
• Nuclear-Fusion, possibly
• Real-time language translation
• Better weather modeling
• and many many more…

When we use AI as a tool to help solve real-world problems, it has the immense potential to literally transform humanity.

home  >  english  >  blog