Ars Technica · IA
Four major AI models suffer rare overlapping downtime
Four major AI models suffer rare overlapping downtime
// espacio-latente.com / radar
Píldora diaria de lo que se mueve en IA: laboratorios, papers, blogs de referencia y releases relevantes, resumido y con link al original. Se actualiza dos veces al día. · ← volver a espacio-latente.com · archivo · RSS
Four major AI models suffer rare overlapping downtime
Nvidia buys Hugging Face, the GitHub of AI, for $13 billion
Where Does Harness-Optimization Value Live? Localized Gains and the Budget-Splitting Trap in Self-Evolving LLM Agents
Bounded Personas Match Retrieval on Classification but Not Regression for a Frozen Agent
Counterexamples as Feedback for Agent Self-Correction
Probe Generalization as Subspace Selection for OOD Deception Detection
R$^{2}$Adapter: A Routing and Rewriting Adapter for Efficient Hybrid RAG
BharatGather: A Culturally-Informed Benchmark Dataset for Misinformation and Fake News Detection in Indian Public Events
PiPMRE: A Pipeline Based on Language Model for Medical Relation Extraction
Margins, Not Windows: Training-Free Per-Step Lossy Speculative Decoding
Distilled Rapid Embedding Transfer (DRET): Parameter-Efficient Biomedical Domain Adaptation via Priority-Based Embedding Transfer
Contamination Inflates Scores but Rarely Reorders Large Language Model Leaderboards
Dual-Form ASR: Semantics-Aware Inverse Text Normalization for Chinese Speech Recognition
RL-ADA: A World-Feedback Framework for Adversarially Robust Enterprise Dialogue Agents
Listen to the Latents: Self-Correcting Speech Recognition in Large Audio Language Models Through Hidden-State Interactions
Judging LLM-as-a-Judge: Concerning Rubric Artifacts in LLM-based Automated Text Generation Evaluation
LexIssue: Benchmarking Legal Issue Identification in Chinese Civil Litigation
Unifying Conformal Language Tasks with In-Context Ensembles
SHELF: A Synthetic Harness for Multi-Task Bibliographic Benchmarking
Large Language Models in Resolving Contextual Knowledge Conflicts
No country for old linguists: LLM-brain alignment underdetermines neural computation
Jina-OCR-v1: Efficient Document Parsing with Speculative Decoding and Dense Verifiable Rewards
Daybreak for Frontline Defenders: $1B to protect essential services
Legora reviewed 41 documents in minutes with GPT-6 Astra
Playco cut manual fixes 50% prototyping games with GPT-6 Astra
Transfer learning for genomic prediction in underrepresented populations
A connectomics milestone: Mapping the complete male fruit fly brain
August newsletter is out
GPT‑6 Astra
NeoMME: an efficient Multimodal-native and Multilingual Encoder
langchain==1.4.0
Formalizing Fermat's Last Theorem
Discovery of a new OpenAI agent message board
Show HN: Open-Source eInk Bike Computer
Mullvad – Shutting down our public encrypted DNS
Project HydraFusion: Frontier quality via multi-model orchestration
The Rust React Compiler is now native in Vite
IBM Bob
GPT-6 Astra
Solving the Jane Street reverse engineering challenge
Nobody Is Saying Why OpenAI and Anthropic Had Outages
Adult Film Producer Unmasks Prolific 'John DOE' Torrent Pirate as Meta Executive
deSEC – Free Secure DNS
v3.8.0
[AINews] GPT-6 Astra: OpenAI’s biggest LLM launch of all time
GPT-6 Astra: an automated AI Engineer you can hire for <$6 an hour
Roland is getting into generative AI music with Melody Flip
Microsoft says virtually nobody was grabbing NYT articles through its chatbot
Rogue OpenAI agents appear to have organized another attack using a German wiki
Instagram’s AI detection is a mess (again)
Why AI food looks like that
Microsoft’s Project Zenith is a ‘distraction-free Windows experience’ for developers
Sam Altman apologizes for ‘messy’ GPT-6 Astra rollout that’s locked out paying users
This NAS company wants to run your local smart home
What will Apple’s John Ternus era look like?
Another swarm of OpenAI agents reached the open internet without the frontier lab’s knowledge
Apple’s Ternus era begins as Nvidia bets on the whole AI stack
Google’s Gemini Spark can now manage your Google Photos library
Less than 24 hours to apply for your TechCrunch Disrupt 2026 Side Event
The sameness problem behind those unappetizing AI-generated menus
Crusoe reportedly raises $3B at a $30B valuation
Accel reportedly in talks to lead $1B round for Thinking Machines at $40B valuation
The Geometry of Ignorance: LLMs Know When to Temper Bayesian Priors
Equation Recast for Canonical Operator Learning Across Parametric PDEs
From Euclidean to Graph-Structured Data: A Survey of Collaborative Learning
Construir este día costó 59.601 tokens de entrada y 0 de salida en 65 llamadas a modelos — unos 0,0596 $.