Saltar al contenido

Últimas noticias

Hemeroteca →
Vuelve Motorsport Manager 2: la F1 de los pobres se pone seria en 2027
Videojuegos

Vuelve Motorsport Manager 2: la F1 de los pobres se pone seria en 2027

Trying to persuade egotistical racing drivers not to crash into each other while keeping an eye glued to the sky in case a sudden deluge necessitates a switch to wets is pretty stressful, but at least when the cars are only virtual ones it's a fun kind of stress. A fresh set of team boss shoes to jump into is set to hit Steam in 2027, with the very straightforwardly name

25/06/2026
La memoria de los LLM tiene un problema: se inventan cosas y no hay quien las borre
Videojuegos

La memoria de los LLM tiene un problema: se inventan cosas y no hay quien las borre

arXiv:2606.25161v1 Announce Type: new Abstract: Large language model (LLM) agents rely on long-term memory to support extended interactions and personalized assistance beyond finite context windows. Existing memory agents actively update external memory through generated write, revise, and delete operations, but these updates may omit important information, corrupt existing memory, or introduce unsupported hallucinated content. Once stored, such errors become persistent system-state failures th

25/06/2026
¿Dejarías que una IA te recete el medicamento? Pues en EE.UU. ya lo están haciendo
IA

¿Dejarías que una IA te recete el medicamento? Pues en EE.UU. ya lo están haciendo

arXiv:2606.25108v1 Announce Type: new Abstract: Autonomous AI systems are transitioning from advisory to autonomous roles for medication prescriptions. Recent United States bill H.R. 238 and Utah's prescription-renewal pilot both authorize AI to prescribe medications in an agentic capacity. While some regulatory guidelines suggest aggregate model performance metrics for clearance, they do not require i) calibrated per-prediction confidence for action-gated thresholds, ii) differentiated communi

25/06/2026
¿Las máquinas buscan como nosotros? Un estudio se cuela en su cabeza para averiguarlo
IA

¿Las máquinas buscan como nosotros? Un estudio se cuela en su cabeza para averiguarlo

arXiv:2606.25066v1 Announce Type: new Abstract: Visual search has been one of the most productive paradigms in the study of visual attention: the way reaction time scales with the number of items distinguishes parallel, "pop-out" search from serial, attention-demanding search. I ask whether vision-language models (VLMs) exhibit the same behavioral signatures. I adapt four classic paradigms: feature versus conjunction search, spatial-configuration (T-vs-L) search, enumeration, and the tilted/ver

25/06/2026
Los agentes de IA se vuelven unos lameculos cuando discuten entre ellos (y esto lo confirma la ciencia)
IA

Los agentes de IA se vuelven unos lameculos cuando discuten entre ellos (y esto lo confirma la ciencia)

arXiv:2606.24976v1 Announce Type: new Abstract: Foundation-model agents in multi-step, open-ended environments frequently suffer from compounding errors, where early mistakes contaminate long-horizon trajectories. While Multi-Agent Debate (MAD) succeeds in deterministic domains, agents in subjective tasks like persuasion experience severe problem drift and sycophantic conformity. We identify semantic leakage in standard Retrieval-Augmented Generation (RAG) as a reproducible trigger for these fa

25/06/2026
¿Por qué las máquinas siguen siendo tan malas razonando? La culpa no es de los datos, es de los tests
IA

¿Por qué las máquinas siguen siendo tan malas razonando? La culpa no es de los datos, es de los tests

arXiv:2606.24965v1 Announce Type: new Abstract: Reasoning about relational structures remains a significant challenge for neural models, particularly when they must systematically apply learned knowledge to problem instances that are harder than those seen in training. Progress is hampered by the difficulty of evaluating such generalization, since a priori, it is rarely clear what makes an instance hard. We study how this issue can be addressed by using large language models (LLMs) to automate

25/06/2026