Deepak Aatresh began his lecture by reflecting on the implications of constant technological innovation in the semi-conductor industry. The semi-conductor industry operates under the premise of ‘Moore’s Law’, where the number of transistors in an integrated circuit doubles approximately every two years, resulting in an increase of speed and performance for the same economic cost. Innovation in chip design has enabled the advancement of mobile technologies, which in turn has reinvented our relationship to the world, and as Aatresh argues, has created a positive change in the world by transforming the way we relate to it. Continuing with the example of mobile technologies, Aatresh noted that while a number of high-tech industries have benefited from ‘Moore’s Law’, enabling a low cost of computational production, the construction industry hasn’t taken advantage of innovation in computation. The design and production processes in the AEC industry still remain substantially similar to the ones in the past.
Aatresh’s company, Aditazz, is introduced as a range of cross-disciplinary knowledge and methodologies from the semi-conductor, software, and AEC industries, questioning pre-established notions of authorship in design methodologies, aiming to take advantage of current computational power to increase the quality and performance of the isolated built environment, while reducing its cost of production and waste[i]. He argues that even if the computational power has substantially increased over the years, the AEC industry has still to take advantage of it, enabling the development of a methodical approach where different agents in design can be intercommunicated and interrelated, in constant negotiation to achieve an optimized value state.
Negotiating Agencies in Design
In recent years, it has become evident that technological and computational platforms for interdisciplinary analysis are promoting collaborations between different agents in scientific inquiry. Within this social landscape, computational technologies are continuously transforming the agency in design practices. This agency in design practices is no longer fixed, or tied only to the designer, but is in constant renegotiation. Computational endeavors that started in the 1970’s and aimed to create platforms capable of encapsulating specialist knowledge, as well as describe and share the many forces involved in design practice[ii], have now become widely available. Facilitated by computational developments, the increasingly heterogeneous landscape of actors in design has started to shift modes of authorship and agency in design into compound modes and cooperation between data, human designers, and clients, distributing the authorship in design.
Some of the individual computational approaches explored by Aatresh through his company have been increasingly adopted by design culture: design optimization and simulation, data-driven techniques, and parametric design have started to introduce transformations into the practical design domain. Through computational approaches and software it has become possible to embed domain specific and specialist knowledge, allowing designers to access multi-disciplinary knowledge without most of the lower-level knowledge previously required. Where a number of specialists were required for a number of tasks like, structural simulation or energy consumption, we now encounter a diverse combination of designers, software developers, and engineers collaborating and pointing towards new disruptive models of synthetic collaborative practice.
Similarly, contemporary practice is highly interconnected, where different tasks encompassing a design are managed by different individuals or groups within equal or different domains. The collaboration between the many agents of design oftentimes is made possible through a set of digital software and different file formats. The agency of design is increasingly distributed between human and non-human agents, creating hybrid relationships among them. Oftentimes digital design systems like parametric design and BIM are becoming mediators between multiple agents in design; the combination of different domains of expertise has resulted more evidently in the introduction of scientific approaches to design practices, enabling the designers to quantify and analyze the implications of their decisions.
Parallel to the disruptive transformations in design practice posed by advances in computation and information technologies, designers are starting to reflect and reject the new roles of the designer in contemporary practice. The designers’ agency might be perceived as threatened in a similarly way as the computer engineers’ was when the first compilers and interpreters were being developed in the 1940’s, replacing the need of machine language and opening up domain specific and specialist knowledge to a broader set of users.[iii] However, the development of compilers and interpreters allowed the engineers to focus on things other than low-level machine language, enabling the evolution of the domain.
The implications of new agencies in design practices seem to point out to a collaborative design practice, where computational methods enable a negotiation between different actors in design. Within a computational design landscape in constant transformation, where design creation is no longer fixed to a single agent, what is the creative potential of computational methods, and modes of collective, and distributed agency that have arisen in contemporary design? What are the potentials of technological tools for a creative collaborative design practice?
Quantification of Design
Similar to simulations for chip design that allow the quantification of the performance of a design iteration, Aditazz builds design methods capable of developing multi-objective rule-based proposals, which in turn get quantified by optimization algorithms and functions that can establish a number of goals for a given design.
Aatresh’s project is motivated by the high value of the AEC industry, over 8 Trillion Dollars[iv] and its lack of material and economic efficiency: in recent years continuously representing a decrease in productivity and an increase of employment. Aatresh claims that by incorporating design methodologies that allow the integration and quantification of design implications, it is possible to cut down construction waste by at least 10%[v]; in addition to quantifying the performance of the design, Aatresh argues, that it is possible to cut down waste by adopting strategies similar to the manufacturing industry, standardizing building components, but not necessarily standardizing the design itself, but the manufacturing and material process, introducing a combinatorial approach to the assemblage of the proposals.
In the lecture, the work presented by Aatresh redefines the agency of the designer; the designer no longer has a direct authorship of the design. However, the agency of the designer is now proposed as the explicit development of a set of rules that constrain and inform the generation of design proposals, and as the developer of fitness functions, capable of quantifying and selecting the performative value of a design iteration.
Our training as designers instills an instinctive understanding of multiple and complex relationships between different agents in design, for example, understanding the relationship of media to design thinking. We understand the structures of thought that accompany traditional and digital representations, and value representation in aesthetic and design evaluation. Similarly contemporary simulation techniques can be thought as a type of representation, helping us evaluate and project or proposal; however, simulation in design has been primarily focused on a search for efficiency in the quantification of the intrinsic properties of the system and its design performance, for example energy usage, material use, or structural performance, transforming the mode of creation of the designer, from the developer of a direct single iteration, to the developer of a system for negotiation between multiple quantifiable values, capable of generating an iterative set of proposals. In such a system, the designer still preserves a subjective decision power, where once presented with a design iteration, he can manipulate the individual parameters, starting a new combinatorial exercise to create new iterations.
The limitation of computational design approaches to a combinatorial exercise between numbers and variables might generate an optimized iteration, but can still be mathematically described as a function of a probability, where the rules and fitness functions limit the combinations to a manageable number of possibilities: even with big data and cloud computing these combinatorial searches wouldn’t be otherwise possible unless they are limited by an intrinsic logic. Applications of optimization methods in design and their results largely resonate with the principles of Artificial Intelligence laid out in the 1950’s and 1960’s where the outputs of a system are only dependent on the symbolic representations and knowledge of a problem (or the world), and its interaction with a process based on a pre-defined and predigested logic.[vi] Parallel to the development of Artificial Intelligence, cybernetics proposed an opposite methodology for problem solving that relied on a constant embodied interaction with the world, creating self-organizing responses to interactions with outside and inside agents capable of building an aggregated, open-ended result; in this model, outputs are not expected, but are the result of the construction of knowledge and learning.[vii] In a design context, generative design systems have mainly focused on constraints, parameters, and rule-sets that are internal to the system itself, transforming the algorithmic process into a top-down quantification of performance, leaving untapped the potential for a constant interaction and learning process with the different forces surrounding a design. In this sense, design methodologies of quantification prevent the computational methodology from the creation of unforeseen relationships between the agents of design, limiting themselves to optimal functions of probabilities. According to Heidegger, technology goes beyond being only bound to its instrumentality: technology is a way of revealing the world[viii]. If technology enables us to create the inexistent, could combinatorial approaches in design called technological? Is it possible to develop creative systems through an interaction loop and agency in the world?
Re-development of computational design methodologies
The technological developments in communication and chip technologies that were constantly pointed out by Aatresh, have largely increased the power of ubiquitous computing turning ourselves into embodied computational devices of worldly quantification, recording traces of our interactions with the virtual and non-virtual world. While ubiquitous computing has transformed our relationship to the world, it hasn’t been adopted as an active agent in design methodologies; the potentials in design culture of ubiquitous computing and the continuous development of computation go beyond the, constantly mentioned by Aatresh, exploration of an increased computational power, they lie on a programmatic redevelopment of the design methodologies. While designers have previously employed the power of ubiquitous computing as a creative and political device for understanding the potentials of data collection technologies in place-making (for example, Laura Kurgan[ix]), current approaches to data-driven design, like Aatresh’s, have ignored the potentials of these technologies and its capacities to transform design methodologies, negotiating with the complex set of forces encompassing design.
While in the 1960’s the development of AI and Cybernetics resulted in their divergence into different fields, current approaches in artificial intelligence seem to represent a hybrid of both disciplines, on one end basing problem-solving on the development of ruled-based symbolic representations of the world, and on the other end incorporating deep-learning methods that are based on an iterative construction of knowledge from multiple interactions with the world[x]. Neural networks of perceptrons assemble knowledge of the world through an interaction with data and the programmer, enabling them to eventually generate new sets of rules and relationships in the problem-solving process, pointing out new possibilities for design practice to go beyond a combinatorial approach to computational design and suggesting the creation of systems that can learn from the world surrounding them. With the overabundance of data, computational methods for data processing, and data representation, is it possible to create design methodologies that incorporate a broader understanding of the world other than the traditional performative constraints? Can generative design systems embody a process that enables a negotiation between the designers, the computational design system, the networks existing outside of the system, and the emergent set of rules and relationships in the world?
The potential combination of exploration, analysis, and generation of new layers and networks of understandings of the world through outside data collection and ubiquitous computing, can extend the social, economic, material, and ecological, benefits of the design process presented by Aatresh, far from the system itself, to an enlarged networked interaction. This encompasses not only an interdisciplinary expansion of the domain itself, but also of the computational design tools and their structures of thought. While contemporary design software paradigms have chosen to facilitate functionality by creating high-level abstractions, encapsulating lower-level functionality, this practice has often-times resulted in ‘black-box’ paradigms, limiting the user to a set of constrained commands and functions. Furthermore, the ‘black-box’ paradigm has resulted in a lack of communication and connectivity not only between diverse software interfaces, but also across disciplines and agents of design. The creation of bridges across disciplines and methodologies, and ultimately a practice in the expanded field, is facilitated by intercommunication across software.
Through the creation of bridges between ubiquitous computing and the new data structures of the city it is possible to speculate about the development of a generative design system capable of interacting, learning and creating emergent proposals focusing not only on the performance of the design, but on the effects of the intervention in an ever-more interconnected world. Design practices can only then be relevant not just to the individual design system itself, but to a larger spatial network which can only then become a ‘disruptive’ positive change, a term marketed by Aatresh.
[i] Aditazz, “Aditazz Building Products Whitepaper”, 2014
[ii] Sutherland, Ivan, “Structure in Drawings and the Hidden-Surface Problem”, 1975
[iii] Edwards, Paul, “The Closed World: Computers and the Politics of Discourse in Cold War America”, 1996
[iv] Aatresh, Deepak, SMArchS Colloquim, Oct. 31, 2014
[vi] Edwards, Paul, “The Closed World: Computers and the Politics of Discourse in Cold War America”, 1996
[vii] Pickering, Andrew, “Cybernetics and the Mangle”, 2002
[viii] Heidegger, Martin. “The Question Concerning Technology”, 1977
[ix] Kurgan, Laura. “Close Up at a Distance”, 2013. Zone Books
[x] Winston, Patrick, “Artificial Intelligence”, 1992.