Foundations & the Economics of AI
AI as a general-purpose technology, machine learning for business decisions, and deep learning — the mental models and economics that separate value from theatre.
Most AI courses show you a demo. This one hands you 19 working datasets, 256 questions whose distractors are length-matched so you cannot guess by shape, and 37 lessons that each end in something you have made.
AI as a general-purpose technology, machine learning for business decisions, and deep learning — the mental models and economics that separate value from theatre.
Transformers, tokens and attention without the equations; prompting as interface design; multimodal and synthetic media with trust boundaries.
The data backbone, build-buy-rent decisions, retrieval-augmented generation, and agentic workflows with autonomy ladders and human checkpoints.
Analytics discipline, marketing & CX, finance & control, operations & HR, and the portfolio method that separates pilots from P&L impact.
Why AI sticks (or dies) organisationally; fairness and labor impact; the EU AI Act, GDPR and IP; prompt injection and crisis response.
Synthesis: portfolio, operating model, investment asks, risk appetite — plus the 2026 frontier: jagged intelligence, physical limits, sovereignty, and agents.
The builder track for everyone: vibe-code working apps from plain language, wire no-code automations with approval gates, and learn to steer AI coding agents without being a developer.
Function-specific patterns in context: demand and pricing in retail, sensors and quality on the factory floor, and the high-stakes discipline of healthcare and government AI.
Hands-on generation: image systems and brand pipelines, cinematic video (Veo/Runway/Higgsfield), voice (ElevenLabs) and music (Suno) — with rights, consent and provenance built in.
The edge, held critically: test-time reasoning and when to pay for "thinking", embodied AI and the sim-to-real gap, and the AlphaFold-class wins that show where durable value forms.
Anyone who has to use AI at work and wants judgment rather than tool tips — analysts, managers, founders and students. It assumes no code, and the labs run in the browser.
No. Every lab exists at three levels, and the Core level needs no code at all. The Builder and Advanced levels use spreadsheets or a Colab notebook if you want them.
You work on 19 real datasets with verified answer keys, not toy examples. Every lesson ends with a quiz, every module with a checkpoint, and the assessments repeat so you can see movement rather than a single score.
€49 for twelve months of full access. Module 1 is free without payment, so you can judge the material before deciding.
37 lessons across 10 modules. Most people work through it in six to ten weeks at a few hours a week, and it is self-paced.
It covers the 2026 landscape — reasoning models, agents, retrieval, the EU AI Act timeline — and every statistic carries its source so you can check when it was measured.