There has been a quick start to the new year with a diversity of emphasis and some urgency from those around us, to locate suitable collaborative environments for project development and resource sharing – a Resourceforge
SDNA joined us for a viewing of the abandoned Second Home space on Mare Street in Hackney. They operate a hybrid media practice with participation workshops and composite image projection at it’s core. Others seek a place to meet where We are all on a quest for suitable spaces to work publicly and have brought in Spacegenerators to the conversation. They have been successfully supporting establishment of community climate centres around the country, facilitating direct responses to the crisis and helping negotiate rate relief in exchange for meanwhile spaces. London rents are out of all our reach so we should appeal for supported access to unused buildings, breathing some fresh ideas and energy into city centres.
We concluded the seasonal festivities with EX Networks where we spent some of the conversation describing our first steps toward home-brew AI development with local installations of Ollama and introductions to CUDA toolkit for RAG. The commercial offer of OpenAI and it’s many challengers is a cost of use that may be beyond your IP budget or comfort zone. EXN provided a high spec PC with plenty of GPU cores for our use with latest locally hosted LLM models such as DeepSeek, a much needed experimentation platform.
Our AI revolution should be celebrated at all scales. Rather than get caught up with cost of using corporate platforms, let’s take up every opportunity to explore options and develop grass-root installations stocked with local data for collective benefit and value exchange as a priority. Nor is it a walled garden that’s needed but open access to tools and techniques required for practice now.
RTF is our working ID for this research initiative and collaborations with EX Networks are ongoing, to dig out a pathway for ‘retrieval augmented generation’, agentic and workflow refinement that we can re-deploy as needed. Their specific needs for speedy tech response demands a lot of their compact engineer team, service that an in-house LLM should well enhance.
Our highly creative prompting for useful output should be accompanied by a healthy dose of skepticism and reserve. If you already collect and host your business information or generate research data, it’s value may not be something you are ready to trade just yet. The training of AI models is surging in sync with our rapid engagement with these groundbreaking tools. All this enthusiastic interaction is providing a rich harvest of valuable context and raw material for service providers to recirculate and capitalise on.
