This my personal website to share work, perspectives and interests. I work in architecture + urbanism, environmental science, and technology. I am also interested in nature, running, music, cognition, politics, and science, and more.
I work, design, write, lecture, teach, consult, and think on the themes below. Click headings below for summaries.
For an intro to the perspectives newsletters, click here.
I believe there is a severe materialization crisis ongoing in the modern world. This crisis has two dimensions: physical and ethical. Physically: abiotic, abiotic and systemic resources necessary for human flourishing and survival are being depleted faster than alternatives are available. Ethically: non-human non-consumable features of the planet are being destroyed despite their ethical and aesthetic value. For me, this is unbearably shocking and tragic. I cannot in any way understand how modern societies and educated individuals are not prioritizing this crisis.
There are three broad approaches to solving the material crisis - by improving at large scale the input, deployment, and output modes of material use. Input: resources can be extracted and processed in more sustainable ways. Deployment: production and consumption processes, and material transforming systems, can be massively optimized to generate more value from less stuff. Output: waste products, material artefacts, and systemic impacts can be greatly improved to ensure that material throughputs are not downgraded more than necessary.
The standard methods to improve the physical dimension of materialization are regulation, fiscal measures, and education to change behaviour. I believe there are better methods available that are based on deeper understanding of what humans are doing - how and why - in materializing their lives. Most of my work is trying to contribute to this emerging science and design of materialization, on two levels. Firstly, there is a need to study scientifically the link between matter and value (which I call the Matter-Value Problem): how do raw material resources accrue, and lose, material value, given that value is coterminous neither with exergy nor preferences? Secondly, there is a need to study the operational processes by which matter is transformed, to create finished goods, services and built environments that people want (which I call Resource Transformation Analysis).
I don't know what the solution to the ethical dimension of over-materialization is. In general, I would say that better understanding of the matter-value problem, and resource transformation analysis, readily leads to the conclusion that high material standards can be achieved in many ways with low total resource impact; and this creates both space and clarity for emphasizing alternative goals to human progress beyond material accumulation, including ethical standards that protect the non-human.
Regarding justice and materialization, I feel strongly that humans should not be left in poverty or without opportunity materially, as many people are, and that in some obvious regards, for example through education and family planning, social development must come before large change in materialization trends, implying a local trade-off. But I don't believe in a simple equivalence between, or forced integration of, social justice and solving the materialization crisis. In fact, I think incoherence around this leads to lack of progress on both fronts. Specifically, “sustainable development” may be a useful political framework across otherwise incompatible realms of discourse, but it is technically incoherent and operationally self-defeating.
Architecture has become a triumph of form and style, based on form-giving designers being rewarded for injecting market-valued identity into limited and impressionistic briefs. Architectural achievements today are often, at the building scale, participation medals for rich people, and, at the urban scale, baubles on a gnarled sometimes dying tree of social space. I believe that a counterbalanced model of architecture - in which form-finding design is rewarded for solving and optimizing for complex and functionally-precise programs - is essential and will have a resurgence. Without this, architecture cannot contribute to actual social goals, including solving the materialization crisis.
To do this, more architectural science is needed: what does architecture actually do, and how, and how can this be technically improved? At the present time, there is almost no stable theory, or even methodology, in architecture - not even an agreement on what these should be based on. This causes the growing problem of computation being used in architecture without a scientifically-grounded mapping of key variables, interdependencies, or methodological guidelines for how good built form is created - leading to either no useful results; yet more speculative forms; or engineering-heavy low-variable optimizations.
In general, I think the science that is required in architectural and urban contexts includes a domain of unique spatial knowledge, with characteristics that cannot be extracted from other types of knowledge. A useful reference model for developing the science of space may be linguistics where the main framing concepts - orthography, morphology, syntax and semantics are immediately useful. But: language science is not itself very stable, and has its own challenges in being subject to useful computation; space is different from language anyway; and the high-order complexity of the built environment is more multivariate, with higher stakes, than maybe any other domain.
There are two material loads through which architecture and the built environment relate to the physical dimension of the materialization crisis: the built form itself, and the induced consumption of the the built form. In reducing the material load of the built form, all three approaches to dematerialization (input, deployment, output) are relevant: how are materials sourced and what is the resource impact over their lifecycle, how are they used to construct built spaces, how do they perform over time, how they are recaptured after use, and so forth? In reducing the induced consumption of built forms, the deployment issue is dominant: how does the building or urban object generate, constrain, or guide material and energy consumption? And the specific challenges are more complex, for example: how does the building induce transport patterns, energy use patterns, and other productivity/health/activity and all other material consumption patterns of the building users? The sum of these two types of materialization is probably, by a large margin, the largest addressable flow of resources through modern societies and invites prioritizing a better designed built environment.
Marc Andreesen says that 'software is eating the world'. My view is that 'the world is eating software, but digesting slowly'. The key distinction is that software is not, at least in every case, merely aggregating different domains of human activity into extended versions of a digital paradigm, even if it looks like it is. For example, spatial disaggregation in architecture and urbanism - and more sophisticated aspects of spatial science, such as access routing, sightlines, offsets and other standard spatial dimensions, and above all the dependencies among them - are not things that yield to raw computation in live operation, they are stable facts with their own implications.
My interest is in the analytical and operational viability of digital application to new domains: it is likely to be ineffective to attempt to optimize non-digital domains without using as much domain-exclusive knowledge that is available and can be further found. For example, in the way that industrial engineering analogies applied to production leads to to workers- as-cogs, which is not effective (let alone ethical), digital operational frameworks applied to, for example, ridesharing, logistics, and housing lead to congestion, badly planned cities, and residences that resemble battery farms - unless specific spatial knowledge is included up front. It is true, in my view, that architecture and urbanism are not supplying much knowledge that can be used for computation, but what there is should be used - and this is an opportunity to stabilize the science, and find more.
As digital transformation shifts from highly localized forms of computation, to highly distributed and connected forms of computation, and as price of all aspects (compute, storage, connectivity, devices) goes down, it seems as if the paradigm of digital capacity as difference engine (computing things as input, generating output) shifts to a paradigm of universal network (linking things and computing their relations). It's at this boundary that platform business models are evolving which, even if it is not understood as such, accentuate the difference between digital and non-digital knowledge domains: platforms are basically mechanisms for optimizing the use of assets through network dynamics which is not otherwise optimized by computation directly.
However, I also believe, inversely, that digital transformation itself represents at least one distinct knowledge domain. For this reason, and given the huge potential contribution of computation and networks, it's not acceptable to govern digital evolution using social and regulatory knowledge that don't appreciate the distinct features of digital knowledge. Tethering airplanes to the ground, and making them run along rail lines, because early airplanes could crash and fly off course - is not the way to engage with new vectors of transformation in their early stages.
Broadly, digital transformation will progress in all directions - software will eat and be eaten - and it is my particular hope that computation in the built environment can help reduce the material loads of built form and consumption induced by built form, in particular by the use of discrete spatial and morphological knowledge in architecture and urbanism. This will be a contribution, potentially massive, to solving the materialization crisis of modern society. This is what I am mainly working on anyway.
Beyond specific change vectors, I think the most interesting question of computation is not the similarity between human and computer intelligence, but the differences. It ought to be far more interesting than it is that human cognition can do the things that it does, and I am generally of the view that human cognition has various transcendental properties and faculties that are not reducible to conventional physics or thermodynamics, let alone computation, and will need to be added to an overall model of cosmology on their own terms.
At a young age, I started feeling a sense of distress for the world. This is doubtless connected to distress experienced by me, and others around me, but I became increasingly puzzled by the meaningfulness of life, and distressed by loss - not particularly my own loss - even of trivial things. At the age of 14, I went on a transcendental meditation course, and soon after, I became interested in Zen Buddhism, and from there Indian and Eastern philosophy more generally, increasingly over time in Buddhism.
As part of this track, I started mapping the core concepts and practices of these traditions, comparing them and gradually getting used to their (to me) reasonable and viable mental outlooks. From there I became interested in all spiritual practices, giving a lot of credence to transcendental states and goals, and although I was and am very resistant to religion and mysticism; I am aware that a lot of value is achieved for many people through devotion, group inclusion and ecstatic states.
I have visited a lot of intentional communities including Findhorn, Erraid, Samye Ling, Rocky Mountain Shambhala Center, Ängsbacka, Burning Man, and while part of my interest in these places is the self-reliant or at least self-defining community creation, I find the focus on spiritual or at least non-production/consumption goals to be very valuable, even if a lot of the practices seem dodgy to me, the balance between the worldly and transcendental is unresolved, and communicable ‘progress’ for individuals and groups seems hard to define and communicate.
My Bachelor’s degree from Oxford is in Oriental Studies, specifically Sanskrit focussing on Buddhist material, with a thesis on Madhyamaka thought. The sophistication of the ideas in Buddhism, leading to and extending from this, makes it hard, for me, to be interested in conventional Western framing of philosophical questions. For example, epistemology, ontology, metaphysics and ethics are entirely linked in Indological thought, without any particular mystical effects; separating them seems highly artificial and stipulative.
Partly because of this background, and partly through general curiosity and desire for clarity - including a wish to get beyond cultural accretions of the Buddhist and Indic traditions (in particular the parts that are unclear to me, such as karma) - I have spent a lot of time thinking about what cognition is and how to characterise it.
Broadly speaking, I believe that cognition is part of the fabric of the universe: what essentially is going on is that at some self-creating boundary, subject and object come to be and mutual define. The precise process I call entification - how entities come to be, both subjects and objects, and I believe that this (including even identification, of the most basic entities, both subjective and objective) involves definably transcendental cognitive acts.
Cognition is certainly not fundamentally computational, and minds are not difference engines. I suspect that information is as much a fundamental property of the cosmos as energy, we will discover that cognition is part of of this pure state of matter.
I started learning piano very late, at age 15, inspired by a friend who was becoming a virtuoso at that time. I studied intensively for about 10 years, most seriously with Robert Milnes in London, practicing regularly between 4-6 hours a day, and performed a few times. I also studied and performed orchestral conducting at Oxford.
I am very interested in the performance traditions and pedagogy of piano playing, and followed with interest teaching genealogies among the great pianists.
The way that music works, how it generates emotion and why, what it really represents in cognitive terms, is fascinating to me. I don’t think it makes much sense to describe music as a kind of ‘storytelling’. Even while there is clearly a narrative structure, and identifiable themes - mapping to plot and character in stories - if music was really like stories, then music would be much more programmatic (i.e. having an explicit story attached), and indeed stories could easily be ‘read off’ from the music. This is very much not the case - music must be appreciated on its own terms - so something else is going on.
In general, harmony is something that clearly works profoundly on the average person without training, and it doesn’t seem to have much narrative definition in its own right, so studies of what music is really doing could probably start with that: how harmonies are perceived, understood, appreciated. My sense is that harmony, and music in general, are their own domain of human knowledge, and cannot be reduced meaningfully to another domain.
I am interested in running and try to run 70-100km a week. It’s clear that running and fitness have huge cognitive benefits beyond physical health.
I have done a lot of cycling, most of it on a custom-build Charles Roberts cycle touring frame (now stolen). Trips include: Northern India, various trans-Switzerland tours and trans-alpine trips, NYC, Catskills, Rocky Mountains, various long tours in Nepal, masses of touring in the UK, including multiple Oxford-Scotland return tours.
I have travelled a reasonably large amount. My first long trip at 18, was travelling overland to India from London. I have tried over time to use low-carbon forms of transport (I have no driving license), usually trains. I travelled overland to Nepal from Geneva, with my bicycle, cycling the first leg through the Alps, and doing various tours in Nepal while there.
Generally, I am interested in languages, linguistics, natural and basic science, maths and computation, and try to study core principles and learn basic techniques periodically. It does fascinate me and worry me that there is so much to know to have any sense of competence in the modern world, and I am on that basis concerned that the human realm is increasingly unmanageable. I find it very hard - to the point of impossibility - to engage with areas where there is no structural insight around which knowledge can be grouped, or rather synthesised. On this basis I find it very hard to know much about history, except in so far as it is the history of a specific knowledge or practice that is evolving.
Overall, I am coming to the view that knowledge exists in distinct domains, and as such science as an overall project is not the same as basic or natural science. If there is a general science, I think it must be information theory (in the sense of all possible information, not information encoding), and to a lesser extent system theory, but the correct way to perceive this is may be as a collection of information domains or dimensions. It's possible, as such, that such domains are infinite even within one universe.
I love film, and believe it is an essential medium for the modern world. I am very interested in the media, and how technology is changing it. I follow politics closely, at least in the UK and the US, and broad international trends. For reading, I have read a lot of fiction, including science fiction, but increasingly worry if conventional books are the best way to promote reading, and more broadly, knowledge sharing. Animated books, audio, various forms of video, seem good ways to share knowledge, and for reading itself, I am unsure if the classic book format is ideal.