Mathematical Modeler

How to Apply

Email a short proposal, roughly half a page, to Use the information contained on this page to tell us how you would "spread" the data in column G of tab 1 labeled "neurons" to the same heading in tab 2, and subsequently to the other tabs. The goal is to fill in all the family order and class rows respective to each group of organisms labeled in the spreadsheet tabs (link below). What strategy would you use to minimize the imprecision inherent to finishing the first pass in one week and refine it over the following month? We realize it will be very inaccurate, but it will give us the marble slab we need to further chip away at the switchboard. Included in the short proposal should be:

1 Week From Now

  • First-pass values for all cells + how this will be accomplished

  • Price

January 1

  • Refined values for all cells + how this will be accomplished

  • Price

Job Description


What animal groups account for the largest share of global suffering?

IW is seeking a mathematical modeler to estimate global animal populations, their total number of neurons, and other proxies for suffering. IW has this data for broader groups of organisms but needs to fill in figures for smaller constituent groups. For example, we don't know how many arachnids there are, but we know the population of their overall phylum (Arthropoda). The first step is to evenly divide the broader totals among smaller groups, and then find more accurate means of modeling.

IW is looking for highly experienced consultants to provide feedback, and/or architect a plan. We are open to help from less experienced modelers who can (a) help connect us with the right architect, and/or (b) do most of the labor vetted by the architect.



Provisional Stepwise Plan:

  1. Fill the “Neuron Count” column for the 7 broad groups of animals listed in Tab 1 of the spreadsheet (S1 below). We know the total biomass and the number of individual organisms within these categories. We'll pick a representative from each group with a typical mass and number of neurons to calculate how many neurons the entire group might have. Other methods, such as encephalization quotient, can be used as well.

  2. On the second and third tabs of that sheet, we will evenly divide this data into smaller groupings of organisms. For example, if there are X "Marine Arthropods," and 300 groups within Arthropoda, we'll assume that crustaceans' total population = X/300.

  3. We will then improve the accuracy of each data point over time as more population, neuron, and other data becomes available. In the meantime, we can begin to address questions using the model, such as: if we assume insects don't suffer, how might this affect the overall burden of potential animal suffering between the groups?

For details on the calculations read this.



To aid neuroscientists competing for the S-Prize, IW is seeking mathematical modelers and similar professionals to estimate the distribution of potential global animal suffering. Imagine a switchboard that displays different distributions of suffering as one toggles between competing neuroscience theories of suffering. The switchboard itself is theory-neutral, establishing an upper limit that can be used to compare the ramifications of each theory on global burden.

This project is vital to prioritizing what animals stand to benefit the most from a broad range of compassionate interventions. It will also help to prioritize what biotech research foci might have the highest impact. For instance, if we assume a theory that says Arctic Krill suffer, it might make more sense to study their physiology than other organisms by virtue of their vast population. The usefulness of such prioritization goes deeper, though. If we were to show that the greatest differences in potential suffering, as one toggled between different neurological theories, rested on conserved neurological structures, genes, etc then research on these mechanisms could be of greater utility. This will also help to prioritize model organisms that could be used to study the effects of gene drives in quarantined ‘happy bio-domes’.


Expediting the Project

We consider the development of the switchboard a long-term project, but we'd like to have a published minimum viable product released immediately. Ideal candidates will pitch IW on more bite-sized specifications, which can be released faster in an iterative manner. For example, would it make more sense to collect population data on each family within Chordata first, or to use the animal groupings in Table S1 below as an initial bucket for counting neurons, as currently listed?

Useful Sources


Note, the "Abundance" column in Table S1 below represents the number of individual organisms estimated based on conversions between the Mass column and the typical size of the organism. The article "How Many Neurons?" uses biomass and other sources to estimate the number of neurons in organisms for which the number is unknown, e.g. using springtail-sized organisms as a proxy.


image (3).png