Prototype #1

Fine-tuned AI image generation

There is a new open-source diffusion model: Flux.1 (the same model that Grok’s image generation is built off of) that rivals its closed-source counterparts like MidJourney and Dall-E.

The benefit of open weights is that you can fine-tune the model, so that is what I did over the weekend, where I added a bunch of captioned images of myself to the model...

Prototype #2

CitySpace

Live London transit in 3D. Taking the data from TfL's open API, we built a visionOS app that tracks the live locations of the tube, DLR and trams around the capital.

Prototype #3

VisionInvestor

A look into using Apple Vision Pro's infinite canvas to explore data visualisation for financial markets.

Prototype #4

Storysight

Storysight is an experience combining Spatial Computing (using Apple Vision Pro) and generative AI that allows you to be actually inside the story, experiencing books in a whole new dimension.

Storysight is an example of how GenAI can be married with spatial computing to create brand new third era experiences.

Prototype #5

Super Car Customiser

A 3D space to customise your dream car, exactly how you want it. Using Apple Vision Pro to provide an immersive way to see your car as you customise it.

A glimpse into what is possible with spatial computing in use cases for e-commerce.

Prototype #6

Sound hunter

A Pokemon Go-like app but for hunting stems of a song such as the bass, drums, and vocals!

Prototype #7

KnittingGPT

Extract instructions from a variety for sources, such as paper or images, and use GenAI to convert them into a consistent format like JSON.

Prototype #8

Vocalise AI

Generate voiceovers to your videos in any style you like. This is me doing callisthenics in the video with narration in the style of Alan Partridge doing a eulogy.