We are improving our AI systems’ ability to learn from human feedback and to assist humans at evaluating AI. Our goal is to build a sufficiently aligned AI system that can help us solve all other alignment problems.
Recommended
Making AI work for everyone, everywhere: our approach to localizationOpenai
Inside our approach to the Model SpecOpenai
Operation "Doppelganger": Russian influence activity targeting UkraineOpenai
Everyday work with ChatGPT WorkOpenai
Romance scams: AI-enabled romance scam workflowsOpenai
Will Hurd joins OpenAI’s board of directorsOpenai
How to Use Google NotebookLM: The AI-Powered Research AssistantHeygen
Multi Face Swap Video: How to Swap Multiple Faces in a Single Video with AIHeygen
Inbox zero: What it is and how to actually get thereZapier
The Coana Approach to Reachability AnalysisSocket
Linux Foundation: Why Open Source Needs Research NowThenewstack
Magma Brings a Systems Approach to Wireless NetworkingThenewstack
Multi-token Residual PredictionModal
How sync. uses Modal to lipsync 100 hours of video a dayModal
State of DevSecOpsDatadoghq
State of Cloud SecurityDatadoghq
Cost-Based Hybrid Query Optimization in MotherDuck - MotherDuck Research PapersMotherduck
Fine-Tuning with Cerebras AI Model Studio Launchpad - CerebrasCerebras
Why Deep Learning is the Best Approach for Speech Recognition - Deepgram Blog ⚡️Deepgram
Partnering with Shutterstock to accelerate our AI research effortsSynthesia
