Matthias Bastian

Matthias is the co-founder and publisher of THE DECODER, exploring how AI is fundamentally changing the relationship between humans and computers.

Matthias Bastian

Mar 10, 2023

Playground lets you easily compare large language model performance

Comic drawing, many mouths stuck in a monitor.

Matthias Bastian

Mar 10, 2023

GPT-4 reveal: Microsoft won't comment on launch rumors

Matthias Bastian

Mar 9, 2023

ChatGPT success unsettles Google - report

Google CEO in front of a yellow video screen at a lecture, speaking into the camera.

Matthias Bastian

Mar 8, 2023

Deepfakes to be used by US military for propaganda

A soldier sits at a desk with his back to the camera in uniform and looks at a monitor.

Matthias Bastian

Mar 7, 2023

Meta's LLaMA leak shows secure AI deployment is more of a pipe dream

A white llama runs to freedom on a dusty road.

Matthias Bastian

Mar 7, 2023

Google PaLM-E combines language, vision and robotics

A wheeled robot with a robotic arm reaches into a kitchen drawer and pulls out a bag of chips.

Matthias Bastian

Mar 6, 2023

Microsoft brings "next-generation" AI capabilities to business products

Microsoft has released three new open source AI models in its Phi series: mini-instruct, MoE-instruct, and vision-instruct. These models excel at LLM reasoning and support multiple languages, but have limitations in factual knowledge and safety.

Matthias Bastian

Mar 5, 2023

German government critics fall for Jordan Peterson deepfake

Jordan Peterson speaking into a camera, wearing a suit. There is also a picture of Annalena Baerbock, the German Foreign Minister, in the upper right corner.

Matthias Bastian

Mar 4, 2023

Google's chat AI Bard not to replace search

Google Chatbot Illustration

Matthias Bastian

Mar 3, 2023

Microsoft's Kosmos-1 is a multimodal step toward more general AI

Systems like ChatGPT or Midjourney are experts in text and image. What happens when you combine these and other capabilities? Researchers at Microsoft are testing this with Kosmos-1, a model that combines image and text.