online school for new zealand year 13 student future architects by marc williams
This page features information about AI apps my students have been using since 2016, Architect School AI Chatbot, various creative AI drawing apps and examples of Generative Design, Prompt Engineering and Algorithmic Design
Chat with Architect School Character.ai
In September 2022 Character.ai released their beta version chatbot app that generates human-like text and participates in contextual conversation. Their goal is to "give everyone on earth access to their own deeply personalised superintelligence".
I have have trained and tested it with more than 100 questions about architecture. It's still in beta mode & a small number of responses were not 100% accurate but the majority were. Here are some responses to questions about pencils, De Stiji, the crit and Palladio.
AI Crit
AI Pencils
AI De Stiji
AI Crit
In 2016 Google released Quick, Draw! an interactive AI drawing game. Then in 2017 they released AutoDraw, an AI that turns your drawings into clipart created by artists.
Nvidia AI painting 2019-2022
In 2019 Nvidia released GauGAN, an AI app that turns simple brushstrokes into realistic images. This video shows how it works.
Nvidia updated GauGAN in 2021 to include text-to-image functionality. The 2022 GauGAN360 update features 360 degree panoramas in 8K quality.
Draw with these experimental AI apps
2022 saw the public release of Generative AI applications, a technological phenomena that enables computers to create images and natural language text (ChatGPT) based on keyword text phrases.
Generative AI images are created with machine learning algorithms by computers that generate images and other digital media based from text prompts by referencing datasets of hundreds of millions of images and text descriptions that were sourced from the Internet.
There are many text-to-image Generative AI applications, the main ones are Dall-E2 and Craiyon by OpenAi (creator of ChatGPT), Midjourney, Stable Diffusion, Neural Blender, Starryai, and the yet to be released Google Imagen.
Image-to-3D like Nvidia's NGP Instant NeRF and text-to-video are the next evolution of Generative AI.
This AI is affecting many industries and institutions from solo artists to gaming, education, journalism and architecture. It is enabling new ways of working and creating digital content but also has issues like copyright and plagiarism.
Detailed text Prompt Engineering can enable the creation of architectural images that can feature: inspiring, futuristic and technologically complex but abounding with optimistic hyper-realistic parametric designs. Check out Generative AI on Instagram
Prompt Engineering is a method of writing instructions so that AI applications like Dall-E2, Midjourney and Crayon can create designs and rendered images of architecture, like the IG examples above.
Written prompts fed into AI apps can be a page of writing. Structure your prompts: Establish context. Explain the topic. Specify the task. Ask follow-up questions. State specific requirements. Use key words. Encourage creativity and imagination. Apps like Prompt Box can organise your prompts.
Info about Prompt Engineering.
An algorithm is a specific list of instructions, like a recipe for cooking something in a kitchen.
Algorithms are not computer code, they are are instructions written by Computer Scientists, Machine Learning specialists, IT Engineers or Programmers etc. These instructions are fed into computers which calculate answers to problems or designs.
Algorithmic Design is used within architectural software like Autodesk Revit, Rhino and Grasshopper to enable computational design of multiple solution ideas, complex parametric design or provide structural or energy efficiency building analysis. Check out OneistoX for information about AI Algorithmic Design.
