An “evolvable script” is a codification of how an organisation does things that evolves over time. It was introduced in The Imagination Machine by Martin Reeves, one of our trustees. Our new Rituals page is an evolvable script for our organisational rituals, from our daily routines to our favourite tools. Writing them down in one place prompts us to improve them and makes it easier for our members to find them.

Our convening power is our ability to quickly identify talented people and convince them to join us. Our new Jobs page increases our convening power by showing all of the positions that we are currently recruiting for, as well as those we have recently filled. It also introduces our standard template for job descriptions. This makes it easier to describe new jobs and to advertise them consistently across platforms.

Our core values are the fundamental organising principles of the London Institute. A simplified version of our obsolete principles, they determine what we do, how we do it, and who we recruit. Think of our values as the axioms in a mathematical system, from which everything else is derived. Each one is expressed by a short sentence, and some link to more detailed summaries written for our website or the press.

We have made it easier to search through all our papers. Now, users can explore our papers through filters such as subject, author and journal. Relevant papers can be found in fewer steps, using a more intuitive interface. Our new navigation tool also makes it easier for us to spot trends in our research output, and to increase our awareness of our colleagues’ work. The tool is in the research section of our website.

Over the last year we have increased the range of sources that give money to the London Institute, including our first major philanthropic gifts. To help others understand where our money comes from, we created a new page that reports our funding news. It is now the entrypoint for the funding section of our website. It gives a human perspective on our different sources of funding, and what each gift is used for.

Until now, we grouped together our similar papers in folders. We have replaced this binary classification with a continuous distance between papers. In the child page of a paper, we show the papers that are most related to it, each with a score between zero and one to indicate its distance. The distance is determined by comparing semantic vector representations of the abstracts generated with a large language model.