How to Learn AI - Learning Curriculum
TL;DR
If you want to learn more about AI you reached the right place - you don’t need a PhD to learn the basics of AI. Start by building a mental model of the "Blackbox": BEFORE - IN - AFTER the Black Box is a big deal already to look behind the scenes.
Why Learn AI Now
- Understand the Black Box: move from magic to mental models.
- You press a button and get a result - but what is really happening in AI under the hood to produce this?
- Improve results: tune your skills and level up your game.
- Spot real use cases: with experience you make better decisions where to use AI - and where it might even harm.
- Manage risk: privacy, bias, hallucinations, prompt injection.
- Future-proof your career: this is just the beginning.
If AI feels overwhelming, that’s normal. The ecosystem moves fast, the math can be heavy, and the marketing is loud. The way through is a clear mental model and some guideance for navigation.
The Basics: AI Timeline & AI Glossary
How did we get here?
Explore the AI Timeline with the big influencial events listed in chronological order.
What are all these crazy terms?
The Glossary has you covered - including a quiz to validate your knowledge.
Mental Model: BEFORE → INSIDE → AFTER the AI Black Box
Think in three phases: before, in, and after the black box. Many of the concepts described here are listed with details and interaction in the AI Visualizations section.
BEFORE the AI Black Box
- Tokenize: text → tokens (results in Vocabulary).
- Architecture: Transformer with layers; parameters are learned.
- Open Weights vs Closed Weights
- On Prem vs Cloud
- Text, vision, multimodal
- Embeddings: tokens → vectors that capture meaning.
- Model type: text-only, vision, or multimodal; open vs. closed weights; on‑prem vs. cloud.
- Training data + objectives, watch our for bias in training data.
INSIDE the AI Black Box
- Giant Mathematical Function
- hundreds of layers and billions of Parameters - learned during large scale training
- high-dimensional space
- Non-deterministic by design
- Probabilities: predicts the next token distribution, not “truth”.
- Attention: each token looks at others to decide what matters.
- Sampling: temperature & top_p control creativity vs. determinism.
AFTER the AI Black Box
- Context window: how much the model can “hold in mind”.
- Prompt patterns: role, task, context, constraints, output template.
- Tools: function calls, retrieval (RAG), code execution, image tools.
- Guardrails: safety filters, validation, traceability and evaluation.
This mental model allows you to visualize structure and flow - and allows you to improve your knowledge in areas that need improving.
Dive deeper: Dedicated Articles & Interactive Visualizations
Explore these sections in dentro.de/ai to gain insights on dedicated subjects that accomplish the Learning Curriculum.
Articles in the AI Blog section
- How Large Language Models like ChatGPT Work
- AI Model Lifecycle - Analogy Explained with Cars
- Analogy to understand Open Source in generative AI Models
Interactive AI/ML Visualizations and Videos
Interactive Websites
- Tiktokenizer
- Moebio Mind
- Transformer Explainer
- BertViz
- LLM Visualization
- LLM Architectures
Videos
- Large Language Models explained briefly
- Deep Dive into LLMs like ChatGPT
Fun Challenge
- Bouncing Ball Hexagon
- Vitruvian Robot