Simon Willison
Understanding ChatGPT Work
Understanding ChatGPT Work
// 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
Understanding ChatGPT Work
Inside Meta’s push to put robots to work in data centers
Accelerating LLM Inference via Vector Index Based Output Embeddings
SciReC: Diagnostic Evaluation of Multimodal, Multi-Turn Relational Reasoning with Adaptive Interaction
Sledgehammer or Scalpel? A Fine-grained Adaptive Framework for Implicit Hate Speech
The Effect of Emotional Context on Large Language Models' Endorsement of Premature Decisions: Comparing Emotional Vulnerability Across Six Commercial Models
PACE: Publisher-Adaptive Content Extraction via Agentic Automation
UIC-AIHealth4All at ArchEHR-QA 2026: Answer-First Evidence Grounding for Clinical Question Answering
Select, Don't Train: The Benefits of Modular Entity Disambiguation with LLM-Based Selection
XHotpotQA: A Benchmark for Cross-Lingual Knowledge Composition in Multi-Hop Question Answering
INSPIRE: An Internalize-Then-Improve Approach for Example-Driven Mathematical Reasoning
A Survey on Rubric-Guided Reinforcement Learning for Language Models
How Do Linear Probes Emerge? A Circuit-Tracing Framework with Concept-Targeted Attribution
Trajectory-Level Speculative Decoding for Diffusion Language Models
When Tokenizers Fail: Byte-Level Chunking for Zero-Shot Transfer to Low-Resource Languages
Knowing Before Answering: Decoding Language Models for Reliable RAG
First Make It Playable, Then Make It Good: Staged Interaction Learning for Small Dialogue-Game Agents
Below the Noise Floor: Bimodal Seed Collapse and Distinct Failure Modes in Small-Model Knowledge Distillation
Load-Bearing Context: The Question Damage Score for Evaluating Context Reliance in Linguistic Reasoning
Informational Antilocality and the Locality Bias in LLMs
Compositional Failure in Audio-Visual LLMs: Late-Layer Prior Dominance Under Cross-modal Conflict
Representation of syntax in LLMs through the lens of linear distance and similarity-aware entropy
I turned my security cameras into an automatic bird identification system
Playa Phone
A walkable ASCII cyberpunk city in one HTML file [video]
ChatGPT Work Tool and Skill Reference
ravynOS: Pre-alpha open-source OS based on Darwin, FreeBSD, Apple open-source
Since it was stripped of planetary status, Pluto’s defenders have been fighting
Weave (YC W25) is hiring ML, AI, product, & design engineers
Cheap GPS jammers are filling the world with navigation dead zones
Apple caught off guard by AI demand for Mac Mini and Mac Studio
Debian won’t ban AI code from its Linux distribution
New York Governor Kathy Hochul thinks AI should be ‘less evil’
ChatGPT to face tougher regulation in the EU
Instagram cracks down on AI accounts pretending to be human
Harvard Law dropout raises $6M for Blue Voice to build a ‘Harvey for police officers’
Clipto uses AI to search terabytes of video and is now valued at $250M
Nvidia’s $3.5B MediaTek bet reveals its plan for tackling Big Tech’s AI chip buildout
Meeting note-taker Circleback adds a free tier to attract more customers
The US is building barriers around drones and robots, but China has scale to get around them
Marginal Coverage Credit Reduces Redundant Exploration in Parallel State-Entropy Optimization
Quantization-Triggered Backdoors in Language Models: Cross-Quantizer Transferability and the Validation--Deployment Gap
DAMP: Decay-Aware Mixed-Precision Recurrent-State Quantization
A Deeper Analysis of Block-Sparse Featurizers
When Muon Meets Task Interference: A Spectral Perspective on Continual Learning and Model Merging
Dandelion: A Spherical Flower for Neural Simulation of Planetary Dynamics
Self-Explainable Multi-Label Graph Neural Network for Correlated Evidence Attribution
Curvature-Aware Radius Shrinkage for Adaptive Nearest Neighbor Classification
More Data Cannot Break a Symmetry: Identifiability by Design
Unsupervised Continual Learning with Growing Self-Organizing Maps and Synthetic Replay
SegBench-GC: Testing Segmentation Invariance in Multi-Step Offline Goal-Conditioned Reinforcement Learning
SafeStep: An Interactive Demonstration of Semantic Communication for Pedestrian Safety Monitoring
RiskBlend: A Multi-Signal Framework for Test Input Prioritization in Machine Learning Regression Testing
Construir este día costó 54.939 tokens de entrada y 0 de salida en 53 llamadas a modelos — unos 0,0549 $.