Practice

We work along three lines that feed one another, each corresponding to one operation of a transistor.

Switching — Workshops

Our workshops are made for people working in arts and culture. No technical background is required.

Participant-led. People bring the real work in front of them and try things on it, rather than following material that has nothing to do with them.
Practice first. The task is the way in; the principle underneath is the by-product. Understanding of how AI works accumulates through using it.
Teach the fishing. Rather than hand over ready-made programs, we want participants able to bend AI to what they themselves need.

The sequence begins in everyday practice and moves towards judgement about AI: optical character recognition and proofreading, web archiving and organising source material, building a simple web page with AI, working with AI on long-form writing, then the command line and a first encounter with AI agents — before returning to a single question, which is where the humanities sit in all of this.

What we are after is not fluency but confidence: that someone working in culture can put AI to their own use without being captured by the language of technology.

Modulating — Advisory

We build AI workflows fitted to cultural and non-profit organisations.

These are among the organisations that put the important questions to society. We want to help them use AI well, so that scarce resources go to advancing the work itself.

We begin by understanding how an organisation actually runs from day to day, then help it get clear about what it really needs, and design the intervention from there. We pay particular attention to the long term: an organisation should end up able to run and maintain its own AI systems, drawing on open-source tools, rather than bound to a technology platform.

Amplifying — Research and publishing

We study how AI is actually being used across the local cultural field. Who is using it? How? Where are the bottlenecks? How is the ecology shifting?

The workshops and the advisory work are our field. What we trace and describe there — the associations that form when people and machines meet — goes back into public discussion, as writing and as talks.

Selected work

Our current advisory engagements include:

A local cultural organisation working on the ecology of arts administration. We are helping them automate repetitive administrative and financial work with AI, so that staff time goes to community work instead.
A dealer in twentieth-century antiques, who has to research and write a description for every object. We are helping build an archive that brings image and text together into a single object record, with AI assisting classification, draft descriptions and preliminary research that can be checked against its sources.
An advisory firm that has to keep client data private and so cannot use cloud AI. We are building them a locally run system that keeps everything in-house, and that can also pull the details it needs out of client replies for later follow-up.

Client names are withheld by policy.