Continuously hardening ChatGPT Atlas against prompt injection
OpenAI posted “Continuously hardening ChatGPT Atlas against prompt injection.” A quick update for the record.
Artificial Intelligence Blogs
Short posts on models, agents, APIs, and the companies building them
Search results: “2025”
130 entriesDecember 2025
Updating our Model Spec with teen protections
OpenAI posted “Updating our Model Spec with teen protections.” A quick update for the record.
Addendum to GPT-5.2 System Card: GPT-5.2-Codex
OpenAI announced “Addendum to GPT-5.2 System Card: GPT-5.2-Codex.” This one looks larger than a routine product note.
OpenAI releases GPT-5.2-Codex
OpenAI announced “Introducing GPT-5.2-Codex.” This one looks larger than a routine product note.
Developers can now submit apps to ChatGPT
OpenAI posted “Developers can now submit apps to ChatGPT.” A quick update for the record.
OpenAI releases OpenAI Academy for News Organizations
OpenAI announced “Introducing OpenAI Academy for News Organizations.” This one looks larger than a routine product note.
The new ChatGPT Images is here
OpenAI posted “The new ChatGPT Images is here.” A quick update for the record.
The "think" tool: Enabling Claude to stop and think
Anthropic posted “The "think" tool: Enabling Claude to stop and think.” A quick update for the record.
people use Claude for support, advice, and companionship
Anthropic posted “How people use Claude for support, advice, and companionship.” A quick update for the record.
We Used Codex to Ship Sora for Android in 28 Days
OpenAI posted “How We Used Codex to Ship Sora for Android in 28 Days.” A quick update for the record.
Donating the Model Context Protocol and establishing the Agentic AI Foundation
Anthropic posted “Donating the Model Context Protocol and establishing the Agentic AI Foundation.” A quick update for the record.
Podium: arming 10,000+ SMBs with AI agents
OpenAI posted “How Podium is arming 10,000+ SMBs with AI agents.” A quick update for the record.
The Walt Disney Company and OpenAI reach landmark agreement to bring beloved characters to Sora
OpenAI posted “The Walt Disney Company and OpenAI reach landmark agreement to bring beloved characters to Sora.” A quick update for the record.
OpenAI releases GPT-5.2
OpenAI announced “Introducing GPT-5.2.” This one looks larger than a routine product note.
Update to GPT-5 System Card: GPT-5.2
OpenAI announced “Update to GPT-5 System Card: GPT-5.2.” This one looks larger than a routine product note.
Advancing science and math with GPT-5.2
OpenAI announced “Advancing science and math with GPT-5.2.” This one looks larger than a routine product note.
What type of inference is planning?
Google DeepMind posted “What type of inference is planning? , Google DeepMind.” A quick update for the record.
Gensors: Authoring Personalized Visual Sensors with Multimodal Foundation Models and Reasoning
Google DeepMind posted “Gensors: Authoring Personalized Visual Sensors with Multimodal Foundation Models and Reasoning , Google DeepMind.” A quick update for the record.
Modeling the Arrows of Time with Causal Multibaker Maps
Google DeepMind posted “Modeling the Arrows of Time with Causal Multibaker Maps , Google DeepMind.” A quick update for the record.
Large Language Models Self-Discover Reasoning Structures
Google DeepMind posted “Large Language Models Self-Discover Reasoning Structures , Google DeepMind.” A quick update for the record.
Scalable Diffusion for Materials Generation
Google DeepMind posted “Scalable Diffusion for Materials Generation , Google DeepMind.” A quick update for the record.
Frozen Feature Augmentation
Google DeepMind posted “Frozen Feature Augmentation , Google DeepMind.” A quick update for the record.
Challenges with unsupervised LLM knowledge discovery
Google DeepMind posted “Challenges with unsupervised LLM knowledge discovery , Google DeepMind.” A quick update for the record.
Learned feature representations are biased by complexity, learning order, position, and more
Google DeepMind posted “Learned feature representations are biased by complexity, learning order, position, and more , Google DeepMind.” A quick update for the record.
Few-Shot Recalibration of Language Models
Google DeepMind posted “Few-Shot Recalibration of Language Models , Google DeepMind.” A quick update for the record.
Learning a Fourier Transform for Linear Relative Positional Encodings in Transformers
Google DeepMind posted “Learning a Fourier Transform for Linear Relative Positional Encodings in Transformers , Google DeepMind.” A quick update for the record.
Understanding Learning from Human Preferences
Google DeepMind posted “Understanding Learning from Human Preferences , Google DeepMind.” A quick update for the record.
Revisiting Energy Based Models as Policies: Ranking Noise Contrastive Estimation and Interpolating Energy Models
Google DeepMind posted “Revisiting Energy Based Models as Policies: Ranking Noise Contrastive Estimation and Interpolating Energy Models , Google DeepMind.” A quick update for the record.
Optimization and Evaluation of Fine-grained Jaccard Indexes for Semantic Segmentation
Google DeepMind posted “Optimization and Evaluation of Fine-grained Jaccard Indexes for Semantic Segmentation , Google DeepMind.” A quick update for the record.
Nevis’22: A Stream of 100 Tasks Sampled from 30 Years of Computer Vision Research
Google DeepMind posted “Nevis’22: A Stream of 100 Tasks Sampled from 30 Years of Computer Vision Research , Google DeepMind.” A quick update for the record.
Kalman Filter for Online Classification of Non-Stationary Data
Google DeepMind posted “Kalman Filter for Online Classification of Non-Stationary Data , Google DeepMind.” A quick update for the record.
Distributional Bellman Operators over Mean-embeddings
Google DeepMind posted “Distributional Bellman Operators over Mean-embeddings , Google DeepMind.” A quick update for the record.
Learning from One Continuous Video Stream
Google DeepMind posted “Learning from One Continuous Video Stream , Google DeepMind.” A quick update for the record.
An Impossibility Theorem in Game Dynamics
Google DeepMind posted “An Impossibility Theorem in Game Dynamics , Google DeepMind.” A quick update for the record.
Optimistic Natural Policy Gradient: a Simple Efficient Policy Optimization Framework for Online RL
Google DeepMind posted “Optimistic Natural Policy Gradient: a Simple Efficient Policy Optimization Framework for Online RL , Google DeepMind.” A quick update for the record.
A Robot Walks into a Bar: Can Language Models Serve as Creativity Support Tools for Comedy? An Evaluation of LLMs' Humour Alignment with Comedians
Google DeepMind posted “A Robot Walks into a Bar: Can Language Models Serve as Creativity Support Tools for Comedy? An Evaluation of LLMs' Humour Alignment with Comedians , Google DeepMind.” A quick update for the record.
Learning Silicon Dopant Transitions in Graphene using Scanning Transmission Electron Microscopy
Google DeepMind posted “Learning Silicon Dopant Transitions in Graphene using Scanning Transmission Electron Microscopy , Google DeepMind.” A quick update for the record.
Partition Tree Weighting for Non-Stationary Stochastic Bandits
Google DeepMind posted “Partition Tree Weighting for Non-Stationary Stochastic Bandits , Google DeepMind.” A quick update for the record.
Performance Prediction for Large Systems via Text-to-Text Regression
Google DeepMind posted “Performance Prediction for Large Systems via Text-to-Text Regression , Google DeepMind.” A quick update for the record.
Distributional reinforcement learning in prefrontal cortex
Google DeepMind posted “Distributional reinforcement learning in prefrontal cortex , Google DeepMind.” A quick update for the record.
Scalable Neural Network Kernels
Google DeepMind posted “Scalable Neural Network Kernels , Google DeepMind.” A quick update for the record.
TIPS: Text-Image Pretraining with Spatial awareness
Google DeepMind posted “TIPS: Text-Image Pretraining with Spatial awareness , Google DeepMind.” A quick update for the record.
Learning Planning-compatible Cognitive Maps with Transformers in PartiallyObserved Environments
Google DeepMind posted “Learning Planning-compatible Cognitive Maps with Transformers in PartiallyObserved Environments , Google DeepMind.” A quick update for the record.
Michelangelo: Long Context Evaluations Beyond Haystacks via Latent Structure Queries
Google DeepMind posted “Michelangelo: Long Context Evaluations Beyond Haystacks via Latent Structure Queries , Google DeepMind.” A quick update for the record.
Neural Fields as Distributions: Signal Processing Beyond Euclidean Space
Google DeepMind posted “Neural Fields as Distributions: Signal Processing Beyond Euclidean Space , Google DeepMind.” A quick update for the record.
Holistic Safety and Responsibility Evaluations of Advanced AI Models
Google DeepMind posted “Holistic Safety and Responsibility Evaluations of Advanced AI Models , Google DeepMind.” A quick update for the record.
Unsupervised Keypoints with Stable Diffusion
Google DeepMind posted “Unsupervised Keypoints with Stable Diffusion , Google DeepMind.” A quick update for the record.
Feature Likelihood Divergence: Evaluating the Generalization of Generative Models Using Samples
Google DeepMind posted “Feature Likelihood Divergence: Evaluating the Generalization of Generative Models Using Samples , Google DeepMind.” A quick update for the record.
Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA
Google DeepMind posted “Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA , Google DeepMind.” A quick update for the record.
Advancing Biomedical Understanding with Multimodal Gemini
Google DeepMind posted “Advancing Biomedical Understanding with Multimodal Gemini , Google DeepMind.” A quick update for the record.