⚙️ Evolution of Software — 1.0 → 2.0 → 3.0
👨💻 “From humans writing code to AI writing code!”
🧠 1️⃣ Software 1.0 — The Era of Traditional Programming
In this stage, humans explicitly write the rules, and computers simply execute them.
💡 Definition: Humans manually write code to construct programs
💻 Method: Rule-based logic (e.g., if ~ then ~ else)
📂 Role of Data: Merely a reference (used as input values)
🤖 Examples: Excel, Photoshop, Word processors, accounting programs, video game logic
🔍 Characteristics: Predictable results, but poor at handling complex patterns (e.g., image recognition)
🧩 Developers directly define how to solve problems through explicit logic.
→ “Logic-based programming”
🧩 2️⃣ Software 2.0 — The Machine Learning Era
Now, humans provide data, and computers learn the rules themselves.
💡 Definition: Computers learn rules from data rather than being explicitly programmed
💻 Method: Input → Model (Neural Network) → Output
📂 Role of Data: The most valuable asset — models learn from it
🤖 Examples: Speech recognition, image classification, recommendation systems, self-driving cars
🔍 Characteristics: Can solve complex problems, but results are hard to interpret (black box problem)
🧬 Developers design data pipelines and model architectures, while the computer performs the learning.
→ “Data-driven programming”
🧠 3️⃣ Software 3.0 — The Era of Generative / Self-Improving AI
AI now understands human language commands and writes code autonomously.
💡 Definition: AI comprehends human language (prompts) and generates code or knowledge
💻 Method: Humans describe tasks in natural language → AI generates code/text/knowledge
📂 Role of Data: Pre-trained data + real-time feedback learning
🤖 Examples: ChatGPT, GitHub Copilot, Claude, Google Gemini
🔍 Characteristics: Humans describe what to create, and AI builds it directly
💬 AI is now “software that creates new software.”
→ “Programming by conversation.”
📊 Software Evolution at a Glance
| Category | Main Actor | Core Element | Representative Technology | Examples |
| SW 1.0 | Human | Explicit Code | Traditional Programming | Excel, Photoshop |
| SW 2.0 | Model | Data Learning | Machine Learning / Deep Learning | Image Classifiers, Recommender Systems |
| SW 3.0 | AI | Language Understanding & Generation | Generative AI (LLM) | ChatGPT, Copilot |
공부공책 