Title: We never did find Dave Jumbo Year: 2023 Video: 1080x1080 MP4 25fps Length: 30s Made using: Blender, Audio2Face, Bard

The first - and only - artwork in the "Storytime" series that I had envisaged as ultra-short dreamy stories delivered by these hyper-digital, lop-sided characters that looked like they were hand-made out of clay, alluding towards something like a cyber-golem.

The style of the stories was very much based on Raymond Carver, William Burroughs, Tom Waits, and David Lynch. Or the paintings of Edward Hopper. I like stories where you're just dropped in to a largely recognisable world, but the story has a kind of illness to it that lingers with you afterwards.

I wrote the story, and modelled the figure in Blender. I used an early AI text to speech system called Bard that introduced all sorts of weird mannerisms into the audio that added to the haunting nature of the work. Then I fed that into NVIDIA's Audio2Face software - another AI system that takes an audio file and creates a facial rig mapping to a custom 3D model.

Back then, the process took days, and it is almost unimaginable that I would do it in the same way today as AI tools have moved on so much. But it has a strange manual clunkiness that I look back on with fondness.

Pieces like these pop up in my work occasionally as I experiment with different technologies and tools. To me, the lineage is always very connected and clear, but to others they can seem as jarring and dissonant with themes in my wider art practice. This is fine. Enjoy this little curiosity as rare find. A weird star in the sky that you can't help looking at and wondering how it got there.