Artificial intelligence
Hallucinations and language models
Why models can produce convincing but incorrect answers, and what retrieval and verification can contribute.
Read article ES · EN
An open notebook for knowledge
Ideas connecting data, technology and decisions.
I’m Sergio Gevatschnaider. I explore how economics, artificial intelligence and data science help us understand problems and make better decisions.
Articles · Open courses · Interactive labs
Connect to understand01 / SG
32archive articles
3learning pathways
2bilingual labs
ES/ENtwo languages, one perspective
Featured lab · New version
Know without revealing.
Enter Ali Baba’s cave and discover the intuition behind zero-knowledge proofs. Follow each round, compare the verifier’s view and experiment with randomness.
ZKPES · ENNo installation
Three starting points
Start exploring.
Models that make mistakes, networks that coordinate and communities that build value.
Blockchain and DeFi
DAGs and blockchain: transactions as networks
Directed acyclic graphs, consensus and alternatives to linear chains of blocks.
Read article ES · EN
Blockchain and DeFi
Cryptocurrencies, tokens and memecoins
A technological, cultural and economic perspective on speculation and digital communities.
Read article ES · EN
An idea, a pathway
Choose where to begin.
Each pathway connects readings and experiments. Mark your steps and continue at your own pace.
45 min · ES / EN
From a bridge to an intelligent network
Recognise graph structures and connect trails, consensus and recommendations.
Before you begin: Basic algebra; no programming required.
- Experiment with the bridges
- Compare DAGs and blockchain
- Explore connectivity and expansion
- Connect networks and recommendations
0 / 4
Start pathway →40 min · ES / EN
Understand and evaluate AI models
Distinguish fluency, reasoning, reliability and the effects of incentives.
Before you begin: Curiosity about AI and critical reading.
- Identify errors and verification
- Explore latent reasoning
- Analyse incentives and reward hacking
- Study synthetic data
0 / 4
Start pathway →35 min · ES / EN
Prove knowledge without revealing
Connect Ali Baba’s intuition with probability, commitments and ZKP applications.
Before you begin: Basic probability; every symbol is explained.
- Experiment with Ali Baba
- Complete three guided challenges
- Compare ZKP and commit–reveal
- Connect to DeFi and zk-rollups
0 / 4
Start pathway →Your progress is saved in this browser. Times are approximate.
Structured learning
Courses and materials.
Three pathways connecting foundations, practice and applications. External materials retain their original language.
Data science for business decisions
From a business problem to validation, models and decisions. Includes modules, notebooks, study guides and simulations.
Machine learningPythonBusinessSpanish materials
Open course ↗Repository ↗Blockchain and decentralised finance
Fundamentals, Bitcoin, Ethereum, applied cryptography and on-chain analytics in a campus organised by learning units.
DeFiCryptographyLabsSpanish materials
Open course ↗Repository ↗Big Data, graphs and machine learning
Learning materials and notebooks exploring networks, distributed systems and machine learning with graphs.
GraphsGNNColabSpanish materials
View materials ↗Repository ↗From concept to experiment
Think with simulations.
Change a condition, observe the result and try again. Each lab shows its available languages.
Laboratory 01
Ali Baba and zero knowledge
Step-by-step rounds, two viewpoints, cheating probability and hash commitments.
Open lab →ES · EN
Laboratory 02
The bridges of Königsberg
Walk the graph, change its bridges and discover when an Eulerian trail exists.
Open lab →ES · EN
Laboratory 03
Decision trees and Random Forest
A laboratory with 16 simulations on partitions, bagging, boosting and interpretation.
Laboratory 04
Random network models
Explore graph structures using the Erdős–Rényi model from the Big Data repository.
An archive of ideas
One question leads to another.
Search by concept, filter by topic or save articles for later. Historical notes retain their original publication context.
32 articles
Artificial intelligence
Hallucinations and language models
Why models can produce convincing but incorrect answers, and what retrieval and verification can contribute.
Read article ES · EN
Blockchain and DeFi
DAGs and blockchain: transactions as networks
Directed acyclic graphs, consensus and alternatives to linear chains of blocks.
Read article ES · EN
Blockchain and DeFi
Cryptocurrencies, tokens and memecoins
A technological, cultural and economic perspective on speculation and digital communities.
Read article ES · EN
Artificial intelligence
Computational creativity and AlphaEvolve
Artificial evolution, algorithm search and implications for scientific research.
Read article ES · EN
Artificial intelligence
Huginn: latent reasoning
An exploration of recurrent reasoning and internal computation beyond chains of words.
Read article ES · EN
Economics and decisions
Technology, intuition and the future of work
Asimov and Polanyi as a starting point for thinking about tacit knowledge, robots and work.
Read article ES · EN
Graphs and mathematics
From Big Data to Ramanujan: expander graphs
Connectivity, mathematical structure and networks that combine sparse links with strong expansion.
Read article ES · EN
China and technology
The car as code
China, software-defined vehicles and the architecture of automotive power.
Read article ES · EN
China and technology
Modelling Chinese strategic power
A perspective on capabilities, decisions and the shaping of future scenarios.
Read article ES · EN
Graphs and mathematics
Tesseract: beyond the visible dimension
Projections, hypercubes and visualisations for exploring four-dimensional structures.
Read article ES · EN
Language and society
Cellular automata and algorithmic persuasion
Models of local interaction for thinking about influence, opinions and their limits.
Read article ES · EN
Artificial intelligence
How TikTok predicts what you want to watch
Mathematics, networks and machine learning behind a recommendation system.
Read article ES · EN
Artificial intelligence
The AI paradox and model collapse
What happens when models learn from synthetic content and lose diversity.
Read article ES · EN
Language and society
Digital psychopaths: manipulation and objectives
Reward hacking, incentives and behaviours that divert a system from its intended purpose.
Read article ES · EN
Language and society
Klingon, the brain and language processing
Constructed languages and their relationship to the brain’s language mechanisms.
Read article ES · EN
China and technology
Laozi, shi and Chinese strategy
Cultural concepts for exploring strategic positioning and action.
Read article ES · EN
China and technology
Harmony and hardware: China’s path to technological power
Technological capabilities, industry and a perspective on power beyond GDP.
Read article ES · EN
Artificial intelligence
AI, agents and graph theory
Multi-agent systems and their use as support for scientific research.
Read article ES · EN
Artificial intelligence
AI that teaches itself
Learning, reasoning and strategies for improving models through new training processes.
Read article ES · EN
Language and society
Li Bai, Tang poetry and creative AI
An encounter between Chinese poetry, culture and artificial intelligence tools.
Read article ES · EN
Artificial intelligence
When biology speaks: AI and proteins
Language models applied to the design and understanding of biological sequences.
Read article ES · EN
Artificial intelligence
Flexibility and neural architectures: Qwen 2.5
Graphs and neural networks as a framework for exploring model architectures.
Read article ES · EN
Artificial intelligence
Robots, PacMan and the laws of learning
Reinforcement learning, interaction with the environment and agent adaptation.
Read article ES · EN
Blockchain and DeFi
Bitcoin as a reserve asset and technology
Economic and technological concepts for studying Bitcoin and its possible roles.
Read article ES · EN
Artificial intelligence
DeepSeek-R1: innovation and geopolitics
Reasoning models, access to technology and international competition.
Read article ES · EN
Artificial intelligence
Janus-Pro and multimodal artificial intelligence
Image generation and new tools in China’s AI ecosystem.
Read article ES · EN
Artificial intelligence
Designing materials with AI
Generative models for exploring new materials and desired properties.
Read article ES · EN
Economics and decisions
Memecoins: collecting and speculation
Psychology, perceived value and market dynamics in digital communities.
Read article ES · EN
China and technology
Copying and originality in the Chinese economy
Shanzhai, innovation and adaptation as forms of economic transformation.
Read article ES · EN
Graphs and mathematics
Chinese characters and graphs: language as image
Visual representations and connections between writing, meaning and networks.
Read article ES · EN
Artificial intelligence
Titans and Transformers: new forms of memory
Memory, context and architectures that expand model capabilities.
Read article ES · EN
Economics and decisions
Game theory and calculated chaos
A reading of political strategies through games, signals and uncertainty.
Read article ES · EN
No articles match these filters.
My research notebook
Economía y Ética.
My blog brings together notes, readings and developing reflections. It is where many of these ideas begin and continue to grow.
Visit the blog ↗Latest notes published on the blog. This selection is synchronised during updates and retains a snapshot when the source is unavailable.
- 2026-09-16The Red Queen of Artificial Intelligence ↗
Bilingual notebook entry. Open the blog and choose English.
- 2026-09-04The Hidden Graph Behind AI Agents: Coordination, Emergence, and Security ↗
Persistent memory, emergent protocols, collective capability, and attack paths: a graph-based reading of the 2026 OpenAI–Hugging Face incident, accompanied by three interactive…
- 2026-05-21Infinity Inside a Circle: Escher, Science Fiction, and Hyperbolic Geometry ↗
A visual and conceptual exploration of hyperbolic geometry, from the Poincaré Disk and Escher’s tessellations to impossible worlds in science fiction and embeddings used in artificial…
- 2026-04-12The Geometry of Uncertainty ↗
Gaussian Processes as a probabilistic framework for learning functions, representing uncertainty, and making efficient decisions in complex systems.
An interdisciplinary perspective
Sergio
Gevatschnaider.
I combine economics, data science and teaching to explore how new technologies transform decisions, markets and society.
This space brings together articles from my notebook, learning materials and interactive experiments. I am interested in connecting intuition with models, and concepts with examples we can explore.
EconomicsData scienceTeachingResearch
Explore my projects on GitHub ↗Let’s keep connecting ideas.
Find my projects on GitHub and my notes on the blog.