What is WikiSim? My problem with answering this question is that WikiSim is not like any other website I know of. This makes it hard to describe succinctly.¹ Currently I have called it "A Public Commons for Democratic Sense-Making" but I still feel there's room for improvement and I'd be interested in any suggestions you have for improving it; please send me an email.

There is no mental model that I've found yet that neatly conveys the concept. It's currently like a mash up between Wikipedia, an online code editor like codepen.io, a data repo like ONS or OurWorldInData, a spreadsheet, and a games platform like itch.io.

I'll list all the previous ways I've attempted to convey what WikiSim is and the value it might afford our democracies, our fellow democratic citizens and how it might help us support our leaders and their advisors in accidentally saying smarter things.

WikiSim One Liners

At different times I have used the following one liners to describe WikiSim:

  • WikiSim holds community made open-source simulations, and “back-of-the-envelope” calculations, to help us to make better sense of our complex world.

  • WikiSim is a public space for data and simulations - a Wikipedia for imagining futures that actually add up.

  • WikiSim is a public digital commons to facilitate data-driven democratic sense-making and deliberation.

  • WikiSim is a Collaborative Knowledge Simulation Platform - a collaborative platform for creating, sharing, and exploring “back-of-the-envelope” calculations, and interactive simulations of our complex world.

  • WikiSim is an unofficial extension of Wikipedia for the things that aren't encyclopedic so don't fit into Wikipedia, Wikidata or Wikifunctions. For example data about national fuel prices, even the strength of gravity on Earth don't have a home in Wikipedia and can't be referenced in other calculations.

What's the Problem?

The world is complex, trust in the democratic process is falling, democratic politicians are failing to deliver on their promises as the economic (and other) myths which supported our democracies for a long time have shifted. Namely that if you worked hard, and got an education, you could live comfortably and raise a family.

We also have increases in resource depletion, various forms of pollution, mental and physical health problems, and most importantly disinformation.

However there is also a lot to be hopeful about and celebrate. We're making great progress in some areas but collectively we're not understanding how much progress we've made and how we might continue to make progress.

Our social media and legacy media platforms are driven by attention, and negative news drives that attention better than realistic news.

Additionally our political leaders from across the political spectrum are reporting it is harder than they imagined, or impossible, to deliver on some of their promises. This is the breeding ground for authoritarianism which is even worse than a bad democracy. WikiSim aims to help address that.

Lack of detail is a deeper part of the problem. We have phrases like "the devil's in the detail" and yet when it comes to public discourse on complex topics we have sound bites and mostly empty rhetoric like "take back control". This lack of detail leaves our democratic citizens totally exposed to manipulation and choosing things that sound good on paper whilst missing the reality of the actual benefits and harms a course of action will result in.

What is WikiSim's Proposed Solution?

This has 4 parts: respect reality not just rhetoric, break problems into smaller pieces, do it publicly, and trust people.

1. Rhetoric and Reality

Rhetoric and vision are important. But we also need promises that are deliverable or flexible enough to adapt as circumstances evolve. Geoff Mulgan rightly highlights the lack of public discussion of "how" a promise might be delivered: what "organisation of roles, structures, processes, money and data [are needed to] make the modern state work".²

For example we might experience a loved one waiting to see a doctor, and a politician promising to "fix the NHS" and "cut waiting times". Great! I'll vote for that. But how many doctors do we need? How many do we have? Are we training more or are attracting immigrants from other countries? If so, how long and how much will that cost? Or if we're "stream lining" services and regulations, what does that mean in practice? Will it actually be effective? What are the trade-offs? What was not in place to empower the doctors, nurses, managers, staff, and patients to do that already?

A lot of this is basic stuff, simple numbers, simple hypotheses of cause and effect, simple back-of-the-envelope calculations. And yet it's almost always missing from the public discourse. WikiSim is a place where we can start to build the foundations for those "how" and "what if" conversations. And yes LLMs might get there one day, but they're not there yet.³

2. Divide and Conquer

Problems broken into smaller parts are easier to solve. A moderately complicated question like "What happened to the GBP to EUR exchange rate around Brexit" took me about 2 hours to find the data, clean it, upload it to WikiSim, then reference multiple different pages, and perform new calculations on top, so that I could answer that question: Conversation Upgrade #1 - GBP to EUR exchange rate around Brexit specifically https://wikisim.org/wiki/1153v9.

These parts can now be reused as foundations for further calculations and simulations.

3. Do it Publicly

We now have a public resource (https://wikisim.org/wiki/1153v9) that anyone can use. They can reference it in other discussions, all the workings are open, and any mistake can be corrected. We didn't have either this specific content or this general capability before I built WikiSim. We do now. Before I published that page, how could we expect people to hold a reasonable discussion about it? At best they would take hours to find the data, clean it, process it and then it would sit in a private spreadsheet on their computer / cloud service and never be shared or improved on. We would have to trust them that it was right. And if they did share it, we would have to trust them that they would not change it, that they would not take it down at some later date, or that Google wouldn't edit or hide their Google spreadsheet. It would not have been easy to discover. It would have been hard to reuse. It would have resulted in others redoing the same work. With WikiSim all of those problems are now solved. And just like with Wikipedia, those issues of trust are solved at the systemic structural level as all the versions are open and immutable.

4. Trust People

With WikiSim, just like with Wikipedia, we also trust in people. Not naïve trust: some people have biases that need to be softly but firmly corrected for. As Jimmy Wales wrote in his latest book "People are born to connect and collaborate" (Rule #2) and "Trust in the power of reciprocity... give trust to get trust" (Rule #4). It's the trust that ultimately people, we, have it in our nature to help each other. But it does require a clear and positive purpose that everyone understands, hence this post aims to help people understanding why and how they can connect and collaborate through the WikiSim site to build a more resilient and flourishing democratic society.

WikiSim's Philosophical Roots

WikiSim has 3 primary philosophical roots: complex problem solving, mediums, and trusting humans.

1. Complex Problems Have Several Unique Characteristics.

Using the Cynefin framework, complex problems are distinct from complicated problems where one person, with sufficient experience, expertise and resources can solve it. Secondly they do not have a single solution.

You cannot approach complex problems the same way as complicated problems like rebuilding an engine where it's possible to find an optimal solution that one person can perform.

Instead complex problems require the input from many different experts. Some of these experts are not "traditional" i.e. someone stacking shelves is an expert on that and can tell you that your AI shelf stacking robot and thus the related government policy has the following flaws. But if you don't have a sense-making system that can take in those expertise, it will take longer to find the solutions, let alone more optimal solutions.

Note the "s" for solutions plural: another hallmark of complex problems is that they have multiple good solutions, it's not "best practice" as per the complicated engine rebuild, instead it's good practice. And there are heuristics that can be developed to guide what good practice looks like.

Another crucial element is that the solutions to complex problems will change over time. Complex systems adapt to the interventions, so a solution that worked last week might not work now, likewise something that didn't work previously, might now work. Nothing is truly constant, everything is shifting.

Complex problems are also easier to solve if you break them into smaller parts as a) they fit inside people's heads so we can reason about them and b) they show us all the areas we agree on so we can focus in on the areas of disagreement and formulate questions & experiments to try to find a way forward rather than sound bites, empty rhetoric and dead ends.

2. Better Sense-Making Mediums

Part of the collective sense-making problem is that we, the citizens that make up our democracies, are trying to make sense using mediums that were bad before and now totally dire. Before we had narrow, highly controlled pipes from centralised mainstream broadcasters. Yes we regulated to ensure that there was some public service and some neutrality but the reality was there was a lot of distortion, relevant material missing, and ignorance. Sometimes this was due to wilful manipulation but often it was because of the medium itself. There's only so much information you can convey in 15 minutes of TV news. And you don't know what someone has already watched so people often hear the same small number of facts repeated again. WikiSim proposes there are 2 existing but under-used mediums that can dramatically help us make sense of the world. One which is trivially simple and the other which is very rich.

The first is back-of-the-envelope calculations. These are things that a small number of people in our societies do with spreadsheets where they find some data, do some calculations and come to some insight, but they don't have an easy place to share that insight, or have it in a trustable open wiki. With WikiSim they do now. For example there's a calculation for "How many neurologists do we need in the UK, to deliver clinical care according to NICE guidelines" https://wikisim.org/wiki/1068. And that lets anyone interested in the topic to have an immediate answer which would otherwise take them many hours, or days of work to arrive at... assuming they didn't make a mistake. Now with WikiSim that's all open, verifiable, editable, improvable, immutable and referenceable.

The second medium which is far richer are simulations, also sometimes referred to as digital twins. These allow for a huge amount of information, potentially gigabytes of data, to be served to a user in an experience that they can have in minutes. Notice it's not just reading, it's an active experience with cognition: when you're using the medium of text to read through and understand an issue, as you read you are converting the words and sentences into a rich mental model, a simulation, that then allows you to run the different scenarios suggested by the author. This allows you to arrive at insights about how the system worked and currently works, and also how it might behave in the future given different circumstances. With a digital twin all of the constraints of the system can be modelled as data and relationships (calculations and code) which can allow you to much more quickly understand the system.

An example of this is the Energy Explorer digital twin / simulation where people with polarised views on solar power (either very pro-solar and anti-solar) very quickly understood its limits after playing it for a few minutes. It was fun, fast, fact-driven: let's do more of that for other issues.

The analogy I have for mediums is that watching or reading 15 minutes of news every day is like going to the gym for 15 minutes: it's good, but you'll not get anywhere close to being able to out run an Olympic athlete. But giving someone a back-of-the-envelope calculation that's contextualised is like giving them a bicycle for the mind: they'll stand a chance of keeping up with or beating the Olympian. And a simulation is like giving their mind a racing car.

3. Trust Humans

Already covered above apropos Wikipedia.

What's the Value of WikiSim?

Bridge Between Experts and Everyone Else

Just like Wikipedia acts as a fantastic bridge between experts and everyone else, WikiSim aims to do the same for public policy and governance. WikiSim aims to aid bridging the gap between the experts, i.e. the people who are doing the work to understand the world at depth and "the rest of us" who are trying to do our best to live in it responsibly. By reducing the barrier to accessing deeper insight it will empower more people to be able to contribute to solving our problems and making our democracies work better.

But Most People Won't Use It!

They'll just use AI LLMs, legacy media, or social media.

That's fine.

  • If some content on WikiSim acts as a source to help train LLM AIs give better results then it will serve its purpose.

  • If it acts as a source of a political advisor, or a journalist, the purpose is also served.

  • Even if only a small number of people use it, that is often enough in a democracy to tip a decision in the right direction. You might only need 0.1% of the population to change their minds, that's 1 in 1000 people. They might influence another 3 people each, that's 0.4%. Taken as a swing, that becomes 0.8% and factoring in say a 50% turnout would be a 1.6% vote swing. Again I have a bias: I don't mind what people vote for but I do want them to get what they vote for and be happy with it. However both of those things are often not the case: people don't feel like they're getting what they're promised or they're not happy with what they voted for once they get it.

Justified Hope

Most importantly, for those who do care about issues and trying to make sense of them, I hope WikiSim will provide a place for logical, justified hope. To move people out of paralysis by dissolving the illusion of intractable expert disagreement, and biasing politicians towards delivering on their promises. It will undermine apathy or an urge for simple answers to complex problems; ultimately avoiding less fruitful authoritarian leadership and instead moving us towards a more engaged, effective and flourishing democratic society better able to navigate complexity.

Notes

¹ This article is attempt 3 at explaining what WikiSim is. Attempts 1 & 2 are still being written because I thought they were too confusing... I know this means I'm not publishing things soon enough because if I confuse you, you'll just tell me.

² In praise of plumbing - Why British politics' lack of interest in how things work explains why many things don't. Geoff Mulgan, 2025-06-29.

³ See this article for a comparison of LLMs attempting, but currently failing, to answer a fairly simple question: How many UK Neurologists do we need for Multiple Sclerosis?

Originally posted on https://ajamesphillips.com/blog/what-is-wikisim-v1