The source code for the Append Us Apart project is now publicly available on GitHub [link].

After months of design and development, and a successful launch on the Vrije Universiteit Amsterdam premises, the Append Us Apart project created by Daan Wijnhorst is now open source.

Append Us Apart is a non-humanoid embodied AI installation where a user can interact with a Large Language Model through voice prompting and get their answers on paper to show the effects of Large Language Model output on the physical world.

Large Language Models have exploded since OpenAI provided access to their ‘ChatGPT’ model. This created a big wave of glorification of the model, but also criticism: makers were not transparent enough about the development process, the environmental impact is vast, users could attribute human features to the output and some questions about moral and legal implications remain unanswered. In ‘Append Us Apart.’, the visitor is asked to handle text generators in a different context and form than from their private devices. The visitors make a (displaced) collective effort to create a physical, permanent form of their Questions and Answers. Should you leave your piece as part of this collaborative effort? Do you see yourself reflected in the output? How does it make you feel?