1. How traditional software works
If you have used a calculator, a banking app, a website, or a mobile game, you have already used traditional software.
Traditional software generally follows instructions written by a programmer.
Imagine a simple program that calculates a discount.
The programmer might tell the computer:
If the purchase amount is above ₹1,000, give a 10% discount.
The computer follows that instruction exactly.
In simplified form:
Input → Rules written by humans → Output
For example:
- Purchase = ₹800 → No discount
- Purchase = ₹1,500 → 10% discount
- Purchase = ₹3,000 → 10% discount
The programmer has explicitly defined the rule. The computer does not discover the rule by itself.
2. What makes AI different?
AI systems can learn patterns from examples or data.
Instead of writing every rule manually, we can provide a system with many examples and allow a machine-learning algorithm to find useful patterns.
For example, imagine we want a system that identifies whether an email is spam.
With traditional programming, we might try to write rules such as:
- If the email contains certain words, mark it as spam.
- If it contains too many links, mark it as spam.
- If the sender is unknown, increase the spam score.
But real-world spam is much more complicated. Spammers constantly change their messages, and there may be thousands of different patterns.
A machine-learning system can instead be trained using examples of:
- Spam emails
- Normal emails
The system looks for patterns in those examples and uses what it learned to make predictions about new emails.
The simplified process becomes:
Data + Learning algorithm → Trained model → Prediction
3. A simple comparison
Think about teaching a child to recognize a cat.
You could try to give the child a long list of rules:
- A cat has four legs.
- A cat has two ears.
- A cat has fur.
- A cat has whiskers.
But those rules are not enough.
What if the cat is sitting? What if it is a different breed? What if the photograph is dark?
Instead, you could show the child many examples of cats.
Over time, the child develops an understanding of the patterns that distinguish cats from other animals.
Machine learning works in a somewhat similar way. The computer is not literally thinking like a child, but it can learn useful patterns from examples.
4. Rules versus patterns
This gives us a useful way to think about the difference.
Traditional software
A programmer defines the rules.
Rules → Computer → Result
Machine learning
The system learns patterns from data.
Examples/Data → Learning process → Model → Prediction
This distinction is fundamental.
5. Does AI mean that humans are no longer needed?
No. This is a common misunderstanding.
AI systems still depend heavily on humans.
Humans decide:
- What problem should be solved?
- What data should be collected?
- How should the system be trained?
- How should its performance be evaluated?
- What should the system be allowed to do?
- How should mistakes be handled?
AI can automate parts of a process, but people remain responsible for designing, supervising and using these systems.
6. An important limitation of AI
Traditional software can make mistakes because programmers can make mistakes.
AI systems have another important source of errors: the data and patterns from which they learned.
Suppose a model is trained using poor-quality or incomplete data. The resulting model may produce poor predictions.
This is why the quality of data is extremely important in machine learning.
A simple way to remember this is:
AI learns from data, so the quality of the data matters.
7. Where do we see these differences?
You encounter both traditional software and AI systems every day.
Traditional software
- A calculator
- A simple alarm clock
- A basic tax calculation program
- A program that follows fixed business rules
AI-powered systems
- Spam detection
- Voice assistants
- Image recognition
- Recommendation systems
- Fraud detection
- Chatbots
- Generative AI
Many modern applications actually combine both approaches.
An application can contain traditional software rules and AI models working together.
8. The key idea
The most important thing to remember from this lesson is not that traditional software is "old" and AI is "new."
Both are useful.
The important difference is how the solution is produced.
Traditional software usually relies on humans explicitly defining the rules.
Machine-learning systems use data to learn patterns that can then be used to make predictions or decisions.
Remember
Traditional software:
Humans write the rules.
Machine learning:
Humans provide data and a learning process, and the system learns patterns from that data.
This difference will become increasingly important as we explore machine learning in the next lessons.
Quick Check
Before moving on, see if you can answer these questions:
- How does traditional software normally produce an output?
- What is the role of data in machine learning?
- Why can a machine-learning system make mistakes?
- Can an application contain both traditional software and AI?
If you can answer these questions, you have understood the main idea of this lesson.