¿Dejaríais una IA trabajando casi 22 días sobre un único objetivo?
Summary
AutoNodo demuestra la capacidad de una IA para trabajar de forma autónoma durante casi 22 días en un repositorio de código de 1.8 millones de líneas, utilizando checkpoints y evidencia para asegurar la confiabilidad.
Similar Articles
Casi 22 días, un solo objetivo y un repositorio de 1,8 millones de líneas: ¿estamos midiendo mal la autonomía de los agentes?
AutoNodo ha logrado una ejecución autónoma continua de casi 22 días sobre un único objetivo técnico en un repositorio de 1.8 millones de líneas, cuestionando las métricas actuales de autonomía en agentes de programación.
26 Hours, Zero Prompts: 3 AI Agents Running Autonomously
The article describes an experimental architecture for autonomous AI agents that can work continuously on high-level objectives for days with built-in control, memory, and verification, demonstrated through a live experiment using multiple agents including Codex.
I let an agent pick its own task every morning for 23 days. 41 runs, 19 reached production, 22 died. The 22 are the reason it works.
An experiment running an AI agent for 23 days to autonomously select and execute tasks via GitHub Actions, with automated checks resulting in a 54% failure rate that enhances efficiency by minimizing manual audits.
Beyond the hype: I just watched an AI agent automate a 4-hour research workflow in 18 minutes.
The author describes observing an AI agent autonomously complete a competitive intelligence workflow that normally takes 4 hours in just 18 minutes, highlighting its ability to navigate websites, process PDFs, and iterate search, suggesting a move beyond simple chatbots to autonomous execution.
@EinsiaAI: An AI agent spends hours on a task. Why should all that work disappear when someone else takes over? Einsia AI’s answer…
Einsia AI introduces AgentGit, an open-source platform for version control and collaboration on AI agent sessions, enabling work to be saved, handed off, and shared.