Week 3b: Questions - Evolution, Complexity, & the Process of Innovation #8
Comments
|
The first paper highlights how successful new ventures innovate by combining pre-existing, proven components into novel business models, reducing risk and attracting investment. The second paper, introducing Assembly Theory, explains how complex systems evolve through the assembly of smaller components, with selection processes determining which configurations succeed. By combining insights from both theories, industries like AI or biotechnology could accelerate innovation by assembling existing technologies or solutions into more complex, functional systems. Assembly Theory’s focus on selection and efficient assembly pathways could guide the integration of these components in ways that optimize performance and reduce the time to market. How can the concept of higher-order modular recombination in new ventures, as discussed in the first paper, be applied to accelerate technological innovation in industries relying on complex systems, such as AI or biotechnology, by utilizing insights from Assembly Theory's approach to selection and assembly of components? |
|
The third paper proposes a theory of organization as combinations of modules, categorizing combinations as higher order (different modules with proven success) vs lower order (nascent solutions applied to new problems). It finds that higher order ventures have more successful IPOs with higher valuations, while lower-order combinations often get bought out mid way with a much shorter timeline. It advocates for higher-order recombination's superior facilitation of experimentation. Almost contrary to what this paper is saying, from the entrepreneur's POV, lower-order combinations could be seen as equally (if not more) attractive than higher order engagements. Since lower-order inventions involve shorter timelines, higher probability of being privately acquired (as opposed to IPO which involves more difficulty and uncertainty), and lower barriers to innovate. Then, how can we further incentive entrepreneurs to pursue higher-order ventures (longer timeline, higher risk, higher reward) instead? Are there ways we can quantify the tradeoff that entrepeneurs face in choosing between these two types of ventures? |
|
In Modularity, Higher-Order Recombination, and New Venture Success, the authors argue that ventures with higher modularity and higher-order recombination tend to be more successful in securing private capital and pursuing IPOs. This makes sense, as successful companies in emerging industries are often not the first movers but later entrants who learn from predecessors’ successes and failures. While I appreciate the efficiency of recombining proven modules to innovate, I worry that foundational, lower-order innovations are ultimately the drivers of long-term growth. The authors suggest public funding for these harder innovations, as private investors often avoid them due to higher risks. My question is whether public and private funding should remain separate to foster overall societal innovation or whether policies should incentivize blended investments. For example, can private capital be encouraged to support lower-order innovation through tax benefits or co-investment initiatives? Would this benefit both short-term efficiency and long-term foundational advancements? |
|
The emergence of successful entrepreneurial ventures often depends on their ability to innovate by assembling components into new, value-generating systems. Modularity, as emphasized in "Modularity, Higher-Order Recombination, and New Venture Success", allows ventures to mitigate risks by leveraging proven building blocks while exploring novel combinations. This approach aligns with the broader framework of complex adaptive systems, where higher-order recombination leads to rapid prototyping, reduced failure rates, and enhanced scalability. How can entrepreneurs and policymakers foster environments that support modularity and higher-order recombination? Specifically, what initiatives can be taken to balance the risks of foundational lower-order innovation with the scalability of higher-order invention? |
|
How can Assembly Theory's framework of modularity and assembly indices be applied to predict the success of entrepreneurial ventures? Specifically, how might the concept of co-assembly spaces help identify pathways for higher-order recombination in emerging industries such as biotechnology and artificial intelligence, and what role could policy interventions play in fostering these complex innovation networks? |
|
“Modularity, Higher-Order Recombination, and New Venture Success” looked at what makes a new venture successful. Essentially, the most successful ones adapt existing technologies and implement them in the market well, without entirely making something new from scratch. The study was conducted within the US, a country at the frontier of innovation, which relies on innovation to spur growth instead of imitation. Is this model applicable to new ventures in any country? I would argue not - it is necessary to evaluate new venture success differently in developing countries, which benefit most from imitation rather than innovation. The most successful new ventures are likely those that introduce existing technologies into the developing country’s market in a way that makes it palatable to the culture and overcomes market frictions like a lack of internet or different regulation. In this way, success comes from how well it is introduced into the market rather than from its new innovative elements. |
|
How can universities, funding agencies, and journals systematically encourage “knowledge expeditions” (i.e., researchers stepping outside their disciplinary boundaries to address problems in unfamiliar fields) so that the resulting changes lead to lasting technological impacts? Similarly, given that a significant portion of Nobel Prizes are awarded for more conservative surprises, how can the scientific community restructure or shift incentives to encourage leading researchers to make bold disciplinary moves and inspire more transformative discoveries? |
|
Both Assembly Theory (AT) and modular recombination in new ventures suggest that success emerges through the strategic selection and arrangement of components, whether biological, chemical, or organizational. The theory of higher-order recombination in entrepreneurial ventures highlights how combining pre-existing modules increases the likelihood of success, while AT emphasizes the importance of historical contingency and assembly pathways in generating complex systems. How might principles from AT—such as the interplay between novelty, assembly index, and selection—be applied to improve our understanding of venture success in dynamic markets? Are there limits to applying such theoretical constructs across domains? |
|
Cao’s, Chen’s, and Evan’s “Modularity, Higher-Order Recombination, and New Venture Success” leverages a dynamic word embedding space, which offers insight into shifting semantic similarities for some innovative companies. For example, in Figure A3, we see that Tesla originally referred to the inventor (Nikola) and was closely associated with nouns like remanence and nonlinearly. Following the release of the Roadster, we see a shift to components like fuel cell and lithium ion. It is now most associated with brands, such as Porsche and Volvo. Can word embeddings be used to quantify or measure innovation levels? For example, we could see if Tesla is closer to words like innovation, creativity, and patent than Apple. How meaningful would these semantic distances be in reflecting levels of innovation, if at all? These relationships might venture into the realm of pragmatics, so word embedding might not work. |
|
"Modularity, Higher-Order Recombination, and New Venture Success" distinguishes between higher order and lower order combinations. Lower order combinations are found to be riskier and less rewarding, but necessary for further higher order combination and long term economic development. What policy would best incentivize these lower order combinations, and how as a governing body could we identify combinations as such? Furthermore, with regards to creative destruction, the higher capacity of larger incumbents to absorb risk of failure would give them an advantage in pursuing both these higher and lower order combinations, whereas an entrant has little room for error. How do we ensure that entrants are still able to contest these incumbents in this capacity? |
|
In "Assembly theory explains and quantifies selection and evolution", Abishek Sharma explains how patent laws act as selection mechanisms in innovation, shaping how technologies evolve. Strong protections increase the assembly index, preserving complex discoveries, but may limit recombination and slow progress. Does strict IP enforcement foster deep innovation or hinder the exploration of new ideas? Assembly Theory can quantify the selection pressure of patents, revealing how flexible vs. rigid systems impact innovation. Do stronger patents (US, Germany) lead to sustained breakthroughs, while weaker protections (China, India) encourage faster diffusion but fewer radical advances? By treating patents as filters in an innovation assembly space, can we predict which economies will lead future technological revolutions? |
|
Is selection short-sighted? This week’s readings introduce us to the mechanism of biological evolution, technological evolution, and organizational evolution. In all these accounts, selection is the key theme. In the Assembly Theory, selection is the critical mechanism that distinguishes novel functional features from random fluctuations. In “The Mechanisms of Evolution,” it is the constant recombination and elimination of the outdated (and thus selection) that brings vigor to the field of technology. The goal of the new venture success paper is to identity ways for Entrepreneurial firms to survive election–to attract new venture capital, and obtain successful IPOs or high-priced acquisitions. What criteria guide the selection? In all cases, the standard seems to be functionality, or more generally, success. I wonder if such criteria might inhibit future innovation by narrowing the pool of resources available for recombination. In biology, new features/creatures can only come into being by combining things that are still alive. Each evolutionary step must produce something viable–some living creatures–at all stages. However, given the innovative power of recombination, will we get something really wonderful by combining contemporary organisms with organisms that got wiped out? Maybe those eliminated creatures fail to survive its time but some of its features can provide great innovative power for our time? That is to say, maybe selection is context-specific and short-sighted? In biology, the recombination of extinct creatures and contemporary ones might meet lots of practical challenges. But in the realm of technology and organization, can the attention to ‘failed’ technologies/organizational strategies bring innovations to the world? |
|
The third reading highlights modularity as crucial for entrepreneurial success, enabling startups to assemble proven components into novel systems. Higher-order invention, which combines established modules across diverse domains, allows ventures to mitigate risks and accelerate growth. Something I was thinking about is, nowadays in the era of a lot of remote work, where teams are globally distributed and rely on digital tools, this principle faces new challenges and opportunities. Remote collaboration can introduce hurdles like asynchronous communication but also can expand access to global talent and markets. How can modularity help startups overcome these complexities while leveraging the diversity and scalability of remote teams to innovate and thrive? |
|
In Arthur’s Combination and Structure, he describes the importance of modularity within the structural composition of all technologies. As a general rule, he states that “... it pays to divide a technology into such modules only if they are used repeatedly– only if there is sufficient volume of use ... what starts as a series of parts loosely strung together, if used heavily enough, congeals into a self-contained unit” (Arthur 37). I am particularly interested in the phrase “if used repeatedly” and “if used heavily enough” here. What exactly are the factors that cause an early-stage assembly of loose parts to eventually become standardized over time? How does repetition over time actually improve the cohesiveness of subassemblies in technology? In the context of AI, is it easier to see improvements due to repetition in less “physical” technologies, such as ML models or algorithms that learn through repeated training with newer input data? |
|
In Tuesday’s lecture, we discussed the difference between point solutions, application solutions, and system solutions for general purpose technologies. How does this relate to the difference between low-order inventions and high-order inventions discussed in “Modularity, Higher-Order Recombination, and New Venture Success?” For instance, does higher-order recombination limit or enhance the long-term trajectory of system-level innovation? |
Drone, cupcakeI don't know but something just isn't sitting right with me regarding Modularity, Higher-Order Recombination, and New Venture Success. I understand why it could be beneficial to differentiate high and low order invention, and look at distance between ideas. But to be honest, to me, the paper seems to be over-abstracting / overlooking some core set of principles and processes of venture capital and startups. It claims that higher order modules is something that can lead to "value-generating systems" and "solv[ing] a human problem," but seems to under-analyze / under-appreciate how to actually arrange those modules into value generating systems, or how to identify a human problem for those modules in the first place. I understand that's important to see what building blocks are strongest, but it seems to have taken the foundation and construction for granted. Maybe foundation and construction weren't the intention of the paper though, maybe this is an unfair critique. I'd like to understand (maybe I missed it in the reading) if there's any predictive power in the models created, and how strong it might be -- it seems widespread that people should apply existing successful strategies when building their business (modules), and one of those strategies might be doing things that no one has done before (distance), and that a good combinations of these modules can lead to adoption and success. But how do we practically implement those things? |
|
The second reading for Thursday, "Assembly theory explains and quantifies selection and evolution," explains how we can use assembly theory to explain and meaningfully measure the steps needed to build objects with higher complexity. This measurement process also implies that natural selection pressures led to the evolution of objects with higher complexity. Additionally, the third reading for Thursday, "Modularity, Higher-Order Recombination, and New Venture Success," tells us that successful new businesses will successfully integrate existing problems into innovative solutions. My question is whether we can expert firms to follow the same evolutionary pattern outlined by the assembly theory paper? Is there a model we can use to index higher complexity firms or predict innovations for firms currently solving important problems with dated solutions? How will AI accelerate this process? |
|
In industries where regulatory constraints really slow the speed of market entry or iteration (like drug development or fintech), how might those extended development cycles interact with the third paper’s key finding that ventures assembling “higher-order,” already-proven modules gain an advantage in attracting investment and achieving high-value exits? Does tight regulatory oversight alter or even invert the relative benefits of lower- versus higher-order invention, since “rapid experimentation” may be curtailed and public/quasi-public entities sometimes underwrite more of the early-stage, higher-risk invention? |
|
Thursday’s readings discussed concepts of modularity, assembly theory, and the combination of different types of knowledge to create meaningful mechanisms. I found it interesting to learn about how words carry multiple meanings, as we learned about dynamic word embedding models to understand how these models can trace how business and technological concepts evolve over time. Higher-order recombination, in addition, leverages proven components to create valuable systems, as seen with companies like Uber, whose success hinged on combining mobile apps, cloud scheduling, and GPS technologies into a cohesive platform. In contrast, selecting untested components often face higher risks and tend to experience less success in the long run. Thus, my question is, can dynamic word embedding models predict future opportunities for higher-order recombination by identifying patterns in historical contexts and semantic shifts? We learned it can trace idea evolution, but can we utilize this, or another model, to predict future opportunities? |
|
In “Combination and Structure” and “Mechanisms of Evolution”, specifically after the section "The Core Mechanism", the idea of technological progression in terms of separate bubbles and opportunity niches arise. through the creation of simpler technologies, more complex ones are formed, and as those are formed and become superior to previous technologies, there would be "avalanches of destruction"(185) of the previous technologies. |
|
Assembly Theory introduces the assembly index as a way to quantify complexity and connect evolutionary processes across physical and biological systems. Given the emphasis on the balance between discovery and reproduction dynamics in fostering novelty, how might this framework be applied to understand innovation in human-engineered systems, such as AI development? Could the assembly equation help policymakers or organizations better navigate the trade-offs between exploration (novel discovery) and exploitation (optimization of existing systems), and what limitations might arise in applying this model to such rapidly evolving domains? |
|
How do the rates of surprising inquiries in middle-growth countries differ from those in high-growth nations, and to what extent do these differences reflect cultural values, such as the emphasis on meritocracy and competitiveness in American culture versus collaborative or community-focused approaches in other cultures? Additionally, how do these surprising combinations of research contents and contexts, as highlighted by Shi and Evans (2023), influence the ability to drive impactful breakthroughs in these diverse cultural and economic settings? |
|
Correlated Novelties as Catalysts for Equitable and Sustainable Societies Considering the concept of "expanding the adjacent possible," how might societies balance between harnessing the novel possibilities offered by high-assembly innovations and addressing the systemic inequities in access to these innovations? In a world increasingly shaped by technological evolution and the dynamics of correlated novelties, how can we leverage assembly theory to rethink the role of history and selection in innovation? It would be interesting to know how correlated novelties serve as a bridge between technological, biological, and cultural evolution to create pathways for inclusive and sustainable development. I believe this discourse connects theoretical insights with practical challenges in fostering equitable innovation ecosystems even beyond borders as well. It challenges us to think creatively about how historical contingency and novelty can be used not only as tools for progress but also as a means of addressing societal gaps. |
|
From the reading "Combination and Structure” and “Mechanisms of Evolution” I have a question: How does the modularity of higher-order recombinations in organizational emergence intersect with the context-dependent manifestations of novelty, and how can dynamic methodologies—such as semantic models and novelty measures—be integrated to better understand the interplay between innovative processes, venture success, and the socio-economic transformation of complex systems? Furthermore, What are the implications of treating novelty as "objectively subjective" in the assessment of innovation, particularly in understanding how modular combinations of components in emerging ventures drive success, and how might this perspective inform strategies for fostering long-term socio-economic transformation through public and private investment? |
|
Incentives Behind Higher-Order Recombination First, given this combinatorial approach to innovation, how do we trade off between incentivizing innovation via monopoly profits derived from patents, and enabling new innovators to build on proprietary processes for higher-order inventions? Second, does the incentive structure encouraging entrepreneurs to pursue higher-order inventions cause systematic underinvestment in lower-order inventions that could expand the set of possible higher-order inventions? I.e. if we take a dynamic approach to innovation, is the funding for lower-level inventions as public goods through university or government research sufficient to attract the optimal human capital investment? |
|
What are the implications of Cao et al (2024) for technology diffusion? Their analysis finds that combinatorial innovation is the key to success, i.e. that the probability of receiving additional rounds of VC funding or reaching IPO increases when firms focus on "higher-order" inventions that combine different ideas, rather than "lower-order" ones. The drivers of this dynamic are not quite clear to me—does it speak to the taste of capitalists for higher-order inventions, the marginal return of capital for higher-order inventions, or something else? Moreover, what are the implications of capital allocation with a "center of gravity" on higher-order inventions? If lower-order ventures fail more often, will that hamper the diffusion of technologies into industries via point applications and the like? |
|
In Combination and Structure and Mechanisms of Evolution, we learn that technology evolves as a recursive system, where innovations build upon existing technologies, creating opportunity niches for further advancements. Historically, trade has enabled the exchange and integration of technologies across cultures, allowing societies to incorporate foreign innovations that made their own economy more efficient. However, the book also emphasizes that technological components must work harmoniously together. This sparked my curiosity: How can technologies from different cultural and historical sequences be effectively integrated, given that they may have evolved through distinct paths and innovation sequences, what key elements are essential to ensuring combination harmony, and how can we successfully create new technologies that take advantage of innovations from different invention sequences combined? |
|
Both readings emphasize modularity and recombination as key drivers of innovation in complex systems, whether in business ventures or evolutionary biology. Modularity, Higher-Order Recombination, and New Venture Success highlights how startups succeed by combining proven modules into higher-order systems, minimizing risk and enabling rapid iteration. Similarly, Assembly Theory shows how selection and historical contingency in modular systems allow for the generation of novelty while constraining unmanageable combinatorial growth. What I’m curious about is how these ideas apply to industries or systems that rely heavily on both lower-order (novel) and higher-order (proven) innovations. For example, in biotech or renewable energy, how do organizations balance the risks of exploring lower-order inventions with the stability of higher-order recombination? Can we identify cases where this balance has been especially well managed? And how might policies or market incentives encourage the kind of modular innovation described in these readings? |
and others
Post your (<150 word) question for class about or inspired by the following readings:
The text was updated successfully, but these errors were encountered: