Published inGenerative AI in the NewsroomPersonalizing Newsletter Headlines: Testing LLMs Against Human EditorsA controlled study shows readers prefer AI-generated headlines 62% of the time — while keeping core editorial decisions in human hands3d ago3d ago
Published inGenerative AI in the NewsroomStructured Outputs: Making LLMs Reliable for Document ProcessingA guide — and prototype — for getting clean data out of PDFsDec 5, 2024Dec 5, 2024
Published inGenerative AI in the NewsroomTesting OpenAI’s o1 Models: A Look at Chain-of-Thought Prompting for Journalism TasksNew LLMs are touted as smarter problem-solvers — but how do they fare in real-world journalistic tasks like data visualization and…Oct 23, 2024Oct 23, 2024
Published inGenerative AI in the NewsroomDo People Want to Get Their News from Chatbots?What data from real-world LLM interactions tells usSep 12, 2024Sep 12, 2024
New release: Substack Python packageLightweight API wrapper for Substack dataApr 13, 2024Apr 13, 2024
LLM-generated labels for topic classificationCan synthetic data replace human annotators?Mar 17, 2024Mar 17, 2024
What should platform data collection look like now?Stronger tools for more difficult dataFeb 28, 2024Feb 28, 2024
