Let's explore the full post-ChatGPT story together:

From AlexNet (2012) to autonomous AI agents (2026) — every model, every breakthrough, every training cost.

• Filterable by category
• Detailed descriptions and paper links
• Training cost comparisons

Open the AI Milestones Timeline

Find the milestone that surprises you most. Be ready to share!

Plus: Microsoft (Azure OpenAI), Amazon (Bedrock, Q Developer), xAI (Grok)

8+ providers × multiple models × different pricing = paralysis

You need a systematic evaluation based on:

• Task type — generation, review, documentation, testing?
• Data privacy — GDPR, on-premises requirements?
• Budget — API costs scale quickly
• Reproducibility — open vs. proprietary trade-offs
• Integrations — IDE, CI/CD, code review toolchain?

Discussion: Which AI coding tool have you already tried? What was your experience?

1. Free for students — GitHub Education program
2. Best IDE integration — built specifically for VS Code
3. Industry standard — 70% of developers use VS Code, most popular AI tool
4. Model-agnostic — supports Claude, Gemini as backends (since 2025)
5. Full spectrum — inline completions → chat → agent mode

We'll also explore: direct API access (Week 6), Claude Code (Week 9), Cursor, and multiple models for comparison (Week 7).

Step-by-step:

1. GitHub Education — Sign up at education.github.com with your @studmail.htw-aalen.de email
2. Install VS Code — Download from code.visualstudio.com
3. Install extensions — "GitHub Copilot" + "GitHub Copilot Chat"
4. Sign in — Use your GitHub account (Education pack auto-activates)
5. Verify — Create a .py file and watch for ghost text suggestions
6. Python extension — Install the Microsoft Python extension

Why? Unified API, thousands of models, not locked to a single provider.

1. Create account at huggingface.co → claim student benefits
2. Generate API token (read-only) at huggingface.co/settings/tokens
3. Clone starter repo:
   git clone https://github.com/L3GJ0N/hf-llm-uv-starter
4. Create .env file with your HF_TOKEN=hf_...
5. Try it out:
   • uv run hf-llm list-models
   • uv run hf-llm ask --model [name] --prompt "Explain recursion"
   • uv run hf-llm chat --model [name]

Your mission: Recreate the Stack Overflow plot using Python + AI tools.

• Find the data programmatically (Stack Overflow Data Explorer, Stack Exchange API, ...)
• Reproduce the chart: monthly question counts + rolling average
• Extend it: Overlay markers for key AI breakthroughs from the timeline

Two observations you will make:

1. This is surprisingly doable with AI assistance
2. You will hit real obstacles (data access, parsing, bugs)

Before you start — answer honestly:

• How difficult do you think this is? (1-5)
• Time estimate for the script only?
• Time estimate for full SW dev cycle (structure, tests, CI/CD, docs)?
• Confidence level? (1-5)

Take the survey now:
https://forms.gle/XqmQDD1KvqYXfWFs5

The gap between your estimates and reality = a core lesson of this course.

• Teams: 2-3 people
• Tools: Must use an AI coding assistant (Copilot recommended)
• Deliverable: GitHub repository with:
  • Python code reproducing the chart + AI milestone markers
  • Record of every prompt you used

Why record prompts?
Prompts are design decisions. They deserve version control just like code.

For every AI interaction, capture:

For now, find your own natural approach. We'll formalize this later.