MAdB, una nueva forma de ver en detalle los efectos a corto plazo sobre la #salud del #ruido del #tráfico en la ciudad de #Madrid madb.netlify.app/

Token costs are why there will be no saas apocalypse / good dev tools are cached intelligence for agents! The popular theory goes: agents can write code, so they'll just rebuild every tool from scratch and hit raw APIs. no more dev tools, no more CLIs, no more software layers. just agents and endpoints! We just tested this and the data says the opposite. We benchmarked Claude Code and Codex on real Hugging Face Hub tasks (~1,000 graded runs), with two setups: the agent-optimized hf CLI vs the agent hand-rolling curl or SDK calls from scratch. Hand-rolling burns up to 6x more tokens on multi-step tasks and fails more often (84% vs 94% task success). And that's just dropping one abstraction layer. It would obviously be orders of magnitude more tokens and a dramatically higher failure rate if the agent tried to bypass HF altogether and rebuild model hosting, versioning, and distribution from scratch. Every time an agent re-derives a workflow from raw API calls, you pay for that reasoning in tokens. every single run. a good CLI compresses that entire chain into a few high-level commands the agent can't get wrong. In a world where everyone is complaining tokens are too expensive, abstraction is leverage: thousands of hours of design decisions your agent doesn't have to re-reason about at inference time. Good tools are cached intelligence for agents! So no, agents won't rebuild everything from scratch. they'll gravitate to the most token-efficient tools, because that's what their owners pay for. The software that survives won't just be accessible to agents, it will be accurate and cheap for them to drive. We're seeing it happen with HF, which is becoming the platform for agents to use AI: ~49M requests in just two months, and growing fast! huggingface.co/blog/hf-cli-f…

En 2022 escribí tres piezas sobre estrategias urbanas contra el calor. Descubrí que la tecnología más eficiente es el árbol. El aire acondicionado dispara la temperatura, la factura eléctrica, la contaminación y las emisiones. El arbol los reduce. elpais.com/opinion/2022-08-1…

Announcing OMT Router — a client‑side routing library that builds #routing graphs from @OpenMapTiles vector tiles and computes routes isolines in the browser. Zero‑backend, multi‑engine, AI-driven, and a custom control for easy @MapLibre integration. abelvm.github.io/omt-router/

I used to like Raptor Mini for fast, small questions or tasks with a huge context, but the model is drifting lately, and any silly question leads to infinite reasoning loops. Such a pity

This is the most accurate representation of #Claude's behavior when the session drags on

Si compraste una casa en 2010 por 200.000€ y la vendes a 300.000€ en 2026. Has vivido 16 años gratis.

El mundo es bien raro porque hay gente poniéndole a sus hijos Summer Dream y luego hay un gato en el veterinario que se llama José Antonio

The upcoming board:

Visualiza fácilmente los autobuses de la EMT Madrid y entiende sus dinámicas por línea, fecha y hora. abelvm.github.io/madrid-bus/ #Madrid #GTFS #Dataviz #Mapa

The insider trading suspicions looming over Trump's presidency bbc.com/news/articles/cge0gr…

EL DISCO DE NINE INCH NOIZE ES DE LO MEJOR QUE HA SALIDO EN AÑOS. DENME MAYÚSCULAS MÁS GRANDES.

Los libros populares son cada vez más simples de leer. Un análisis de New York Times bestsellers muestra que el número medio de palabras por frase ha caído de más de 20 en los años 30 a cerca de 12 hoy. En los años 30, muchas novelas populares tenían frases de 20-25 palabras de media. Hoy, en algunos bestsellers, apenas superan 10-12 palabras por frase. El gráfico de The Economist sugiere una tendencia clara hacia textos cada vez más simples economist.com/culture/2025/0…