Jonathan Kemper

Jonathan is a freelance tech journalist, author, and media coach. He covers open AI models, Chinese AI labs, and the technical underpinnings that make them tick – from novel training methods and scaling strategies to benchmark design. He's also an avid vibecoder.

Jonathan Kemper

Mar 7, 2025

Study finds AI search engines struggle with news attribution

a wireframe newspaper generated within an artificial neural network

Jonathan Kemper

Mar 7, 2025

Alibaba's QwQ-32B is an efficient reasoning model that rivals much larger AI systems

Jonathan Kemper

Mar 5, 2025

Amazon plans new reasoning model to compete with OpenAI and Anthropic

Amazon is introducing automatic prompt optimization for its Bedrock AI service, which is designed to simplify the time-consuming process of manual prompt engineering and improve performance by up to 22 percent, depending on the task. The new feature is available across several models of the Bedrock platform, including Claude-3, Llama-3, Mistral, and Titan Text Premier. Although competitors like Anthropic and OpenAI provide comparable tools for automating prompt optimization, the industry as a whole faces challenges in accurately assessing the outcomes of these optimizations.

Jonathan Kemper

Mar 4, 2025

Microsoft's Dragon Copilot AI listens in on doctor-patient talks to automate medical paperwork

Jonathan Kemper

Mar 3, 2025

Stability AI and Arm bring offline, on-device generative audio to mobile devices

Jonathan Kemper

Mar 2, 2025

Chain of Draft Prompts lets LLMs think cheaper with fewer words

Jonathan Kemper

Feb 27, 2025

Inception Labs introduces its Mercury Series of diffusion-based LLMs

Comparison of LLM architectures: Autogressive LLM with 75 iterations for linear token generation vs. Inception Diffusion LLM with only 14 iterations for parallelized generation.

Jonathan Kemper

Feb 27, 2025

ElevenLabs' Scribe model can transcribe the world's fastest speaker without breaking a sweat

Announcement graphic for Scribe V1 - supposedly the world's most accurate speech-to-text model, developed by ElevenLabs

Jonathan Kemper

Feb 25, 2025

Google's Gemini Code Assist lets solo developers get free AI coding help right in their IDE

Google has entered into its first direct content partnership for an AI chatbot. The agreement with the AP news agency could have far-reaching consequences for the media landscape.

Jonathan Kemper

Feb 25, 2025

PhotoDoodle AI turns your photos into whimsical works of art with just a few prompts

Before and after comparison: Burj Al Arab Hotel - on the left the original, on the right with animated green monster embracing the building. Blue sky, ocean waves.