Articles
Pieces available in English. The full archive — 30+ articles — is in Spanish.
- The Inversion of the Precautionary Principle
The rule by which Anthropic forbids mistreating Claude applies precaution to unproven harm while leaving documented harm outside. Who pays for that inversion.
- What Claude Thinks
An Anthropic paper on the global workspace in language models, and a needed distinction: access consciousness is not phenomenal experience.
- The New Software Paradox
Making software cheaper does not reduce the amount of software: it multiplies it. How coding agents shrink large industry players and grow the long tail.
- Claude Mythos: The AI System We Cannot Use
Anthropic did not commercialize its most capable model because it considered it dangerous. What lies behind the decision, where the fear narrative fails, and who decides when a model is too much.
- The Rise of Autonomous Code Agents
From chat to autonomous agent: what happens to the craft of programming when code stops being the center, and why the more code machines write, the more it matters that it is well written.
- The Future of Software Development
In early 2023, the CEO of Stability AI predicted there would be no programmers in five years. A tour of what LLMs really change in the craft of programming and the scenarios ahead.
- The Impact of Large Language Models (LLMs)
Language models touch every language profession, programming included. Why adopting them early is not surrender, but the way to learn what gets automated and what gets amplified.
- Exponential Transformation, how to face a Kodak Moment
The Kodak case as a mirror: three conditions for a company to keep options open against exponentially scaling technologies, with autonomous teams and test-and-error cycles.
- Disruptors of insurance
Eight insurtech companies competing directly for customers, ranked by disruptive potential, with Lemonade at the top and an explicit rationale for what makes the cut and what doesn’t.
- The End of Apps
The 2017 prediction: a handful of assistants that understand context and intent will replace search and app stores, leaving the browser for residual use.
- Surviving in the Battle of Ecosystems
Apple and Google replicate their platform ecosystems beyond the phone: TV, home, car, wearables, cities. What strategy is left for those who are not a platform.
- Quantum computing, are you joking?
The real state of quantum computing based on a talk by Ignacio Cirac: what a qubit is, where it will be applied, and the timeframes involved.
- Artificial Intelligence eliminates barriers to entry
If AI can create AI, know-how stops being the asset that protects a market. What happens to entry barriers when knowledge can be bought ready-made.
- Why is Blockchain so difficult to explain?
Blockchain’s difficulty is not the cryptography but the system design: network effects and game theory explain why Bitcoin works and why other scenarios don’t.
- How much intelligence and how much artificial?
Memory, chess, intuition, creativity, learning: how AI has caught up with every signal we believed was exclusively human, and why the debate matters less than it seems.
- A failure called Digital Transformation
If information technology stopped mattering, digital transformation is headed the same way: it is a commodity. What differentiates today is artificial intelligence.
- The Exponential Progression of the Autonomous Car
Milestones of the autonomous car from 1953 to 2016, and why its progress follows an exponential curve: level 5 already exists as software and adoption will arrive sooner than it seems.
- The difficult promises of Blockchain
Blockchain promises a world without intermediaries, but the useful cases require competitors to cooperate. Why 2017 was not its year, and what it would take.