Duncan Gough

A creative, technologist and autotelic, designing and juxtaposing ideas applied to creative communities, games, empathetic digital companions, and products that users love.

I am a maker looking to answer questions about how we live alongside technology as it continues to disrupt. And I am trying to thread the needle between initiating meaningful work for large audiences, creating art that moves people, and getting paid.


Recent

Over the last few years I focused on using creative technology in the haus, in your home. I created a passive AI companion called Ara, designing and selling a limited number, as a means of exploring the emulation of companionship with AI. Later, I designed and created Maylo as a way to understand how a passive AI could document your ambient life - using machine learning to classify background noise, and Google's doodle dataset to represent the places and experiences she could hear, Maylo documented your life as a subtle and passive companion.

Away from my research with Companion AI, I worked on the ambient haus, designing and selling a limited run of peacelights - small furniture pieces that used light and silhouette to create a soft and welcome haus guest. I also sold a limited number of scents for neurodivergents - small and discrete scents that work by cutting through spiraling obsessive minds by going straight for the feels. Recently, I designed another limited run product; a low-power e-ink interface and low-distraction daily mantra inspired by horoscopes and hieroglyphs, using a selection of icons each day to inspire thoughfulness on a slower daily cadence.

Most recently, I was proud to be named a two-time winner of the prestigious Design Minimalist award from Apple for dedication to creating apps that follow a minimalist design approach, adhering to the "do one thing well" theory of design.


Portfolio

Those concerning the internet, the ambient haus 1, the people & stories

Peace, friend ✌️

A tool for focusing on oneself.

Ported from React Native to Switft, Peace, friend is an ambient background audio meets visual ASMR app designed to help you mitigate distractions and relax, whilst providing relaxing visuals that help your to regain focus.

#ios #swift #ambient

Small Seasons πŸ‚

A guide to understanding.

Taking the original idea from Small Seasons, a calendar of 24 smaller seasons as described in ancient Chinese and Japanese farming calendars. Small Seasons on iOS is a widget that displays the relevant small seasons, making them an ambient and glanceable part of your life.

#ios #swift #ambient

Hieroscope β™“

Thee horoscope of hieroglyphes.

Provides ambient and visual provocations, deliberately open to interpretation, helping you focus and reflect throughout each day.

#ios #swift #health

Shaders 😎

Several GLSL shader prototypes, advancing previous creative coding explorations that used Processing and p5.js.

#creativecoding #glsl #shaders

Chai πŸ€–

Chatbot implementation with a focus on the whole conversation, rather than the realtime dialogue. Chai summarises the previous conversation, in the style of Clarice Lispector, and removes the chat history, allowing the reader to focus on the current and "complete" state of the discussion.

#openai #langchain #chatbots

Ambient Audio/Visual ASMR mobile app πŸ•ŠοΈ

React Native app for ambient background noise supported by AI-inspired generative art to improve focus, clarity, and peace of mind.

#react #mobile

Music + Stable Diffusion = Video 🎡

Using StableDiffusion, ControlNet, img2img prompting & Davinci Resolve to create a AI-assisted short music video.

#stablediffusion #controlnet

Stable Diffusion & Deforum prototypes πŸŽ₯

Two videos created with Deforum, using photographs to seed the image generation process along with discrete embeddings trained on a select number of decorative art objects (mostly ceramics, embroidery) from the V&A collection. In each video you can see clear hints of the patterns and styles from those embeddings in use as each frame is rendered.

#stablediffusion #deforum

Stable Diffusion & Vector Studio πŸ–οΈ

Using a new StableDiffusion plugin to generate vector illustrations from a seed photograph, using prompts to develop the animation while mainting a coherent style.

#stablediffusion #vectorstudio

AI-assisted digital art patterns βœ‚οΈ

Using StableDiffusion to train an AI on public domain images of artistic materials, styles, and mediums, and working with AI to design tiling patterns for use in crafts and digital projects.

#stablediffusion #digitalart

ChatGPT with documents πŸ—Ž

Integrating ChatGPT with custom PDF, webpage, and .csv data, adding a chatbot interface with the underlying data.

#chatgpt #langchain

ChatGPT with a Museum attached πŸ–ΌοΈ

Prototyping a method of connecting ChatGPT with external data, provided a natural language search interface to museum collections.

#chatgpt #langchain #llamaindex

Ebsynth Style Transfers 🎸

Created a short film to learn the intricacies and limitations of ebsynth style transfers

#ebsynth #machinelearning

Ebsynth + TouchDesigner πŸŽ₯

Using Ebsynth's style transfer capabilities to convert a short film to rotoscope, applying the style from one keyframe to the rest of the file, and TouchDesigner for adding effects to the finished project.

#ebsynth #touchdesigner

TouchDesigner + OpenAI πŸ“

Published an integration between TouchDesigner and OpenAI to design a creative path for AI to work with generative art tools.

#python #touchdesigner

Steganogra-pfps γŠ™οΈ

Minting NFTs and using steganography to embed side-channel data directly into the image. Inspired by the Spore creature creator.

#solidity #steganography

Mirror of Multitudes πŸͺž

Check your vibe in the multiverse - using OpenAI to generate selfies that reveal how you look in an alternate reality.

#openai #nextjs

Memory Palais πŸ‘‘

Using the OpenAI datasets to make it easier to decorate our memory palaces with remarkable images.

#openai #nextjs

Video Filters

Gesture recognition in the browser & real-time video effects, using the Google Mediapipe APIs.

#mediapipe #machinelearning

Radio Robot

Thee πŸ΄β€β˜ οΈ radio station run by robots, 🦜 not included. Created out of a desire for an always-on, passive listening experience, the Radio Robots had free reign to deliver their own programming schedule with only a few strict rules.

#text-to-speech #APIs

Can AI be creative? πŸ€”

A lightning talk at the Site Gallery that helped me to condense some uncertainties around AI into to work on digital companionship.

#AI #creativity

‘hypebot! 🎀

A quick lockdown hack to create a speech recognition hype-bot as a kind of meeting augmentation companion.

#speech-to-text #IoT

Amy Href πŸ“°

A trainable newsletter-parsing bot that lived in your inbox and managed a daily dashboard of fresh links and news, and learned what content was the most shared, popular, and 'liked'.

#machinelearning #rubyonrails

bliss β›―

“Everywhere you look, there seem to be increasing signs that we are living inside a novel that JG Ballard started to write at the exact moment he died, a novel that takes the form of a reverberating hallucination”

Inspired by this quote I built a living data visualisation of social activity across London, mapping social media check-ins over an abstract Google Maps style background, to reveal the feedback loop when digital and physical meet.

#nodejs #dataviz

Metro πŸš‡

Playing on the theme of designing for opposites, this was a mobile game about data couriers, designed for the underground and only playable offline.

#mobile #gamedesign

Jewel Thief πŸ’Ž

A one-button/one-finger casual game designed for desktop and mobile with variations written in Flash and PyGame.

#gamedesign

Slider / WordSlider 🧩

A pair of match-three style word puzzle games, built using Pygame.

#gamedesign #pygame

GrokItBot πŸ€–

A trainable Bayesian ICQ/MSN chatbot, able to parse, learn and 'guess' using AIML.

#AIML #bayes
Those concerning the ambient haus

lighttable ⚑

A bold and modular lighttable, featuring customisable silhouette inserts with a magentically attached tabletop surface.

#productdesign #silhouette

Peacelights πŸ’‘

Crafting a limited edition set of nightlights with silhouette inserts to create ambient, restful furniture pieces.

#productdesign

vrnjr 🎨

A physcial, generative art product with ambient lifecycles and a library of different design styles guided by a 'director' AI that selected the colour, effects, duration etc, based on a simple and effective time-of-day cycle.

#generativeart #proceduralgeneration

Maylo 🎧

Maylo used environmental audio recognition datasets to track and identify her environment by the background noises and draw in her sketchbook, creating an ML-powered record of each day. Using pytorch, Quick, Draw! dataset, and AudioSet.

#machinelearning #environmentalaudio

Ara 🐦

As part of my work investigating what it will be like to live with AI in the future, Ara is an experiment in understanding robotic companionship. Ara has been featured in an AI edition of Creative Review, exhibited at the Derby QUAD, and led to talks at the Site Gallery in Sheffield, and the Festival of Marketing in London.

#machinelearning #generativeaudio

Glyphe πŸͺ·

Using a raspberrypi and eink screen to create a daily meditatative mood-setting selection of hieroglyphs that act as an ambient daily horoscope, deliberately open to interpretation. Icons from Alice Noir.

#ambientdigital #raspberrypi

iQu πŸ‘οΈπŸ‘οΈ

A consciously small and fun project to add life to an e-ink screen, these eyes refresh independently on a regular basis throughout the day, keeping track of the state of each eye in order to present a harmonius pair.

#raspberrypi #eink
Those concerning people and stories

Somewhere ⚑

Co-founded a future of work and employee engagement platform that encouraged visual sharing, community meetups, and prompted open discussions that focused on the how and why of what people are working towards.

#futureofwork #community

Storythings @ Dublin Maker Faire 🧦

Spent a day as part of a long-running arts and research studio, Special Branch, making LED sock-puppets whilst running a storytelling booth at the first Maker Faire in Dublin, 2012.

#storytelling #puppets

PMOG / TheNethernet πŸ•ΉοΈ

Co-founded a games company in San Francisco and turned the internet into a passively multiplayer online gaming space.

#multiplayergames #play

Recognition