This startup’s new mechanistic interpretability tool lets you debug LLMs
Goodfire wants to make training AI models more like good old-fashioned software engineering.
Topics: ModelsInfrastructure
Topic feed
Models
Model launches, frontier labs, and foundation-model research.
Google DeepMind Blog
AI co-clinician: researching the path toward AI-augmented care
Google DeepMind is researching the path toward an AI co-clinician that could work under physician authority to assist doctors and patients, enabling new models for AI-augmented care.
Topics: Models
Entities: ModelsGoogleGoogle Deepmind
Mitchell Bryson AI Articles
Nobody's locked in anymore - Mitchell Bryson
In the space of 48 hours, OpenAI broke free from Microsoft exclusivity to land on AWS, GM replaced Google Assistant with Gemini in 4 million cars, and Harvard dropped ChatGPT Edu in favour of Claude. The pattern is unmistakable: switching AI providers has...
Topics: ModelsInfrastructure
Entities: ChatGPTOpenAIModelsInfrastructureMicrosoftChatGPT EduGoogleGeminiClaudeGoogle Assistant
Simon Willison AI Filter
Release: llm 0.32a1
Access large language models from the command-line
Topics: Models
OpenAI News
Where the goblins came from
How goblin outputs spread in AI models: timeline, root cause, and fixes behind personality-driven quirks in GPT-5 behavior.
Topics: Models
Entities: Models
Simon Willison AI Filter
LLM 0.32a0 is a major backwards-compatible refactor
I just released LLM 0.32a0, an alpha release of my LLM Python library and CLI tool for accessing LLMs, with some consequential changes that I’ve been working towards for quite …
Topics: Models
Simon Willison AI Filter
Release: llm 0.32a0
Access large language models from the command-line
Topics: Models
OpenAI News
Building the compute infrastructure for the Intelligence Age
OpenAI scales Stargate to build the compute infrastructure powering AGI, adding new data center capacity to meet growing AI demand.
Topics: ModelsInfrastructure
Entities: OpenAIModelsInfrastructure
OpenAI News
Cybersecurity in the Intelligence Age
OpenAI outlines a five-part action plan for strengthening cybersecurity in the Intelligence Age, focused on democratizing AI-powered cyber defense and protecting critical systems.
Topics: Models
Mitchell Bryson AI Articles
The biggest bet in tech just got called - Mitchell Bryson
OpenAI committed $600 billion in compute spending betting that revenue would keep doubling. This week, the Wall Street Journal revealed it missed its own growth targets and its CFO is publicly questioning the plan. Meanwhile, Google just poured $40 billion...
Topics: Models
Simon Willison AI Filter
A quote from OpenAI Codex base_instructions
Never talk about goblins, gremlins, raccoons, trolls, ogres, pigeons, or other animals or creatures unless it is absolutely and unambiguously relevant to the user's query.
Topics: Models
Mitchell Bryson AI Articles
The model war is over — the device war just started - Mitchell Bryson
Three stories from the same week reveal the AI industry pivoting to its next phase. DeepSeek slashed model prices by 90%, confirming that raw intelligence is commoditizing fast. OpenAI responded not by cutting prices but by announcing plans to build its own...
Topics: ModelsAgentsInfrastructure