Week 1 HW: Principles and Practices

Class Assignment

Application: SoilBuddy

With climate change set to disrupt global agriculture and the unsustainability of commercial monocultures threatening to render the arable land left unusable, humanity urgently needs solutions facilitating soil stewardship to fortify our food supply. Existing solutions are either too expensive for small-to-mid farms, worsening inequality in the developing world, impractical for massive commercial farms, necessitating guesswork set to be upended by the sea change wrought by climate change, or provide, at best, noisy, time-lagged feedback through remote sensing.

I propose SoilBuddy as a low-cost, real-time, large-area biomachine platform to supercharge the soil microbiome using nature’s own solutions. The crux of SoilBuddy is intelligent biosensing that incorporates human-readable signals, provides an attack surface for genetic circuit-based intervention, and can integrate with existing, non-biological soil monitoring/fertilization workflows. At the same time, SoilBuddy features means to intervene in the soil microbiome by orchestrating microorganisms, and modulating nitrogenation, oxygenation and ion levels in the soil.

Getting to brass tacks, I outline its sensor/effector features below, with core functions necessary for a MVP in bold:

Sensing

• pH –> Pigment
• Nitrogenation –> Multiplexed pigment / biofluorescence
• Oxygenation –> infrared pigment
• Pathogen –> checkerboard pigment

Effector

• pH (proton pumps)
• pH buffering (zwitterion secretion)
• Nitrogenation (modulating different metabolic circuits)
• Oxygenation (modulating different metabolic circuits)
• Pathogen defense (phage libraries, defense proteins, orchestrate natural predators)
• Microbiome orchestration (quorum sensing mlcs, chemokines)

Governance Goals

While I was inspired by the four tenets of biomedical ethics, I reckoned the most important goals relevant to SoilBuddy are Non-Maleficence and Justice. Given the echoes of the GMO debate from earlier in the century, and the proliferative capacity of bacterial systems, prioritizing safeguards against ecological contamination are crucial for governing our bacteria-based platform. At the same time, given the long history of agricultural intervention being wielded as a political tool to harm access to food and development, it’s crucial that we prioritize justice in the administration of SoilBuddy to ensure its applications prove to be a tool for good.

Goal 1: Non-Maleficence

In the context of SoilBuddy, a commitment to non-maleficence - articulated in the credos “first, do no harm” - involves a rejection of utilitarian logic to instead foreground safeguards against proliferation and mitigation of unintended consequences before the launch of SoilBuddy.

1.1: Counter-Proliferation

SoilBuddy must be governed to prevent runaway proliferation in wild ecosystems and spillover events should be minimized as far as possible.

1.2: Mitigation

Before applying SoilBuddy, practitioners in the SoilBuddy ecosystem must enact measures that mitigate the harm caused by both foreseen unintended consequences (ie. known risks of biotech in agriculture per se, such as gene contamination) as well as unforeseen unintended consequences (ie. put in place inherent safeguards or plans of action to manage unknown unknowns).

Goal 2: Justice

In the context of SoilBuddy, our construction of justice pertains to its distribution and development into perpetuity – I mean to open-source soil health.

2.1: Material Equity

Governance should focus on ensuring geographic and socioeconomic equity in access to the materiel of SoilBuddy, administering its distribution on the basis that food ought to be a human right, in line with the UN’s SDGs.

2.2: IP Equity

Governance should guard against the privatization of developments on SoilBuddy as a platform, recognizing that, again, access to food as a human right contributes to the common good of the species. Insomuch as SoilBuddy will start off open-source, its derivative art should also be accessible to all.

Governance Actions

Option One: Building-in Self-Limitation as a Technical Strategy (“Self-Limitation”)

Similar to how nuclear power plants exported under technical assistance programs are designed to make weaponization impossible, one governance action for industrial, academic and governmental practitioners in the SoilBuddy ecosystem to undertake would be stressing self-limitation when designing SoilBuddy: incorporating failsafes (eg robust, immutable genetic kill-switches) within SoilBuddy to mitigate the risks of unknown unknowns when it comes to mitigating accidental proliferation; and only including biosensor and effector modalities amenable which negate or greatly minimize the risk of deliberate proliferation.

Academic practitioners will need to self-police their research into SoilBuddy to avoid the development of high-risk features, as will industrial stakeholders, despite the potential upside of risky features in contravention of safety-first design principles. National governments must agree to avoid research into safeguard-free SoilBuddy platforms despite the agricultural productivity they may enjoy, not to speak of avoidig the weaponization of SoilBuddy to wage agricultural war. International governments will have to dedicate scarce resources, and scarcer funding to coercive as well as persuasive measures to promote this technical strategy, promoting global buy-in despite political differences as well as counterproliferative interdiction.

A great uncertainty remains as to whether such kill-switches remain biologically feasible, and robust to accidental/natural degradation (eg mutations, horizontal gene transfer). Pessimistically speaking, this strategy also assumes an alignment of academic-industrial complexes away from private interests, not to speak of strategic apparatuses of power. Given how anything short of absolute, collective alignment will rapidly corrode the incentive of collective security that motivates cooperation on this technical strategy, it is also unclear exactly how stable such an arrangement will be without active enforcement against rogue states and non-state actors.

Beyond the failure modes analyzed while elucidating the policy’s assumptions, it is also possible this technical strategy of self-limitation could dampen the research momentum required to make headway on a key technology of humanity’s food future; that, shirking risk altogether might prevent a substantially greater benefit that outweighs even the existential risk of SoilBuddy’s proliferation and inequitable distribution (ie. concentration within centers of comparative advantage) which could actualy be a more efficacious state of affairs.

Option Two: Mandating Equitable Access as an Explicit Policy Measure (“Legislating Equity”)

Switching gears from food, let’s consider the state of California’s Assembly Bill 685 as a piece of legislation enshrining access to water as a human right; in a similar fashion, option two would involve (inter)national legislation (and requisite enforcement) guaranteeing equitable access to the material and IP of SoilBuddy as the key to a more secure food supply across geopolitical borders and strata of development. This stands in stark contrast to the oligopolistic pricing of economic factor inputs and complementary agritech/biotech goods used in the production of food crop in certain major agricultural markets, as well as the fervent privatization of IP governing crop cultivation methods – much like the denial by John Deere of farmers’ right to repair their equipment, but in contraposition to Norman Borlaug’s methods of knowledge-sharing that underpinned the green revolution.

Given the bottlenecks manufacturing inequity – such as mass production, marketing and distribution – lie in industry, academia does not figure in Option Two as much insofar as we discount the role of university corporations that might own key patents and licenses withholding access to SoilBuddy platforms and developments thereof. Taking them to constitute industrial actors for the sake of argument, we thus find that the key stakeholders in Option Two are the industry players charged with balancing the profit motive and corporate responsibility, national regulators that must maintain the free market incentive to innovate with an interest in delivering on their duty to citizens, and international government faced with material limitations in enforcing violations of legislated equity despite national interests in shaping the competitiveness of their own agricultural productivity, or those of others.

Specifically, taking this legislation to be an international treaty guaranteeing mechanisms of cross-subsidization pegged to national income levels viz debt and requiring the registration of SoilBuddy IP with a INGO, we find that Option Two also requires the assumption of total cooperation to negate private incentives to cheat (in an economic sense). Beyond overlapping similarities with Option One’s modes of failure at the industrial and national levels, Option Two relies even more on state and international capacity for the enforcement of this law since there isn’t a structural incentive to ensure equity, as much as there is one to maintain technical non-proliferation due to salutary effects on SoilBuddy metabolic budgets, the avoidance of legal liability or sidestepping openings that might facilitate corporate espionage.

Apart from the failure of legislated equity to take off given the precedent of global pharmaceutical distribution, there is a distinct risk that this legislation may be weaponized for hybrid contestation on the international level. On the other hand, should there be comprehensive good-faith buy-in, the runaway success of mandated equity might paradoxically entrench corollary inequities in the distribution of R&D capacity across the Global South by eliminating the incentive to develop indigenous agritech capabilities, or entrench inter-nation trade dependencies, that if not adequately de-risked, could be disrupted by trade disruptions - innocuous as they might be, such as the maritime trade disruption caused by the Ever Given’s freak beaching in the Suez Canal.

Option Three: Engineering Markets to Incentivize Equity (“Incentivized Equity”)

Apart from altruism and compulsion, a third drive worth exploring is that of economic incentives that might appeal to naturally self-interested, rational economic agents. Similar to the ethos underpinning the FDA orphan drugs scheme, I propose the disbursement of strings-attached market incentives such as tax holidays, targeted research funding, export preferences, indirect subsidies and fast-track approval to govern the nonproliferative development and equitable distribution of SoilBuddy.

Here, a finer dissection of industry is necessary. We lump industry’s R&D side of house with academia in the form of the academic-industrial complex, which will naturally be incentivized by the targeted fiscal outlays promoting responsible SoilBuddy platform development to incorporate biosafety features in the IP that will head into commercialization; this can be verified by funding bodies during progress reviews and the tiered disbursement of funding tranches subject to the satisfactory fulfilment of performance metrics germane to non-maleficent and just SoilBuddy platforms. Next, industrial players overseeing commercialization, production and marketing will be incentivized by corporate incentives such as tax holidays and indirect subsidies to stay faithful to safety-first IP, as well as allow regulatory oversight of their products to ensure they meet the counterproliferative and equitable design features articulated above. On the international level, member fee waivers at INGOs as well as the enactment of diplomatic commons would promote good faith adherance to biosafety in the use and distribution of SoilBuddy, just as the UN has slowly but steadily driven development across the board, and the WTO common prosperity through responsible trade.

Option Three rests heavily on governmental capacity for the microeconomic to international administration of the scheme, from the efficient disbursement of funds, to the capacity for responsible statesmanship promoting the common good. Aside from the obvious failure mode should government failure (ie. rent-seeking) distort incentives for the responsible development and distribution of SoilBuddy, there is the crucial issue of national strategic interest in the performance of the biotech firms developing SoilBuddy in the high-income biotech exporters likely to muster the industrial and governmental heft to implement Option Three in the first place; importing nations might indeed suffer dumping should the subsidies provided for the responsible development and disbursement of SoilBuddy be anti-competitive. Thus, strong international governance of SoilBuddy’s trade is pivotal in maintaining the sustainability of SoilBuddy’s application worldwide.

Evaluation of Governance Actions

We evaluate the governance actions on their efficacy in meeting the governance outcomes articulated above, as well as policy desirability, feasibility and sustainability.

Selection of Governance Action(s)

I would select Option One (“Self-Limitation”) and Option Two (“Legislated Equity”) as complementary policies that offer both a high-uncertainty, high-reward as well as moderately-achievable option comprehensively guaranteeing biosecurity on both counts, of non-maleficence and justice.

Option One (“Self-Limitation”) alone would seem the most promising governance action barring its inability to directly achieve justice. It is also highly contingent on the biological underpinnings of SoilBuddy, which is an unfortunately uncertain, if rewarding bet. Given the shortfall of a single policy, we pursue either Options Two (“Legislated Equity”) or THree (“Incentivized Equity”) as a synergistic policy that, in a swiss cheese fashion, provides better coverage of biosecurity objectives.

The key differentiator between Option Two (“Legislated Equity”) and Option Three (“Incentivized Equity”) lay in the unsustainability of fiscal outlays at both the national and international levels as required for the latter, as well as the greater alignment of coercive incentive in the former to adhere to non-maleficence. As a whole, Option Two is attainable, though contingent on the tedious, if achievable negotiation of international interests to settle on a suitable piece of biosecurity legislation.

Crucially, both policies hinge on the fair assumption that there is sufficient technical as well as state capacity to see SoilBuddy through in the first place in a responsible manner. At the same time, we qualify that, pending further research, the technical research underpinning first-of-its-kind killswitches and safety measures required for Option One is more uncertain than international cooperation on the governance of a risky technology, which has precedent in the historical governance of nuclear energy, as well as more contemporary discussions of human germline editing and mirror life.

Ethical Concerns and Appropriate Governance Actions

This week’s class and this exercise as a whole have brought to mind two worries, one deontological and the other of a more utilitarian turn. Firstly, I questioned if the governance of biosecurity might err on the side of caution and unduly impinge on the freedom of sovereign individuals to pursue their SoilBuddy-related curiosities, or nations to self-actualize their agricultural potetial. Applying a Lockean construction of the tradeoff between governance and freedom, however, one realizes that the necessary surrender of a few freedoms pertaining to the riskiest of technologies, especially ones that might have existential implications for the species and planet as a whole, really is a freedom from, to channel Isaiah Berlin, a greater evil; the freedom to feed our individual or collective curiosities is contingent on the responsible stewardship of the commons which make it possible.

Secondly, from the standpoint of effective altruism, I wondered if governance might be enacting longtermist harm by impeding the pace of technological research necessary to mitigate the existential threat the species faces to its food supply. Again, on the balance of probabilities this more consequentialist reasoning enables, we conclude that the risk of massive ecological contamination in the image of an Atwood dystopia requires that we factor in guardrails against irreversible harm borne of an irresponsibly Pandoran scientific accelerationism.

On a metacognitive level, inasmuch as one might worry about the balance of caution and speed required to safely realize the promise of SoilBuddy for food security across the species, it is this very debate that constantly refines humanity’s construction of this negotiation between risk and reward.

Week 2 Lecture Prep

Professor Jacobson

1. Polymerase has an error rate of 1:10^6. Given the length of the human genome, ~3.2 Gbp, the number of expected errors (3200) is substantial; the error rate of polymerase is significant for the length of the human genome. The error rate of polymerase is mitigated by proofreading during DNA replication, as well as MutS-mediated mismatch repair after replication.

2. Taking the average human protein to contain ~400 amino acids in its primary structure, this means ~400 mRNA codons would be necessary to code for it. Given the degeneracy of the genetic code, working backwards and assuming a roughly equal proportion of different amino acids for the sake of argument, the codon table implies 4^(820) * 3^(120) * 2^(1220) = 1.3210^178 different sets of DNA triplet codes to encode the typical human protein. In practice, certain DNA triplet code sequences encode mRNAs that are more structurally stable, degradation-resistent, and amenable to translational regulatory elements necessary for mRNA translation and protein folding.

Dr. LeProust

1. Solid-phase phosphoramidite synthesis.
2. Due to accumulated errors, the yield for fully-synthesized sequences is impractically low, especially since they have to be purified and isolated. For long sequences, steric hindrance near the substrate as well as chemical degradation further reduce yield.
3. For similar reasons as above, direct synthesis’ stepwise inefficiency would cause the yield to be impractically low.

George Church

1. They are: Phenylalanine, Threonine, Tryptophan, Leucine, Lysine, Methionine, Arginine, Histidine, Isoleucine and Valine [1]. To use the Jurassic Park example at surface value, I think picking lysine was a poor choice for omnivorous/carnivorous dinosaurs since they were able to obtain the essential amino acid from their diet, or the environment. At the same time, it offers a broader lesson on building safeguards into our biotechnology; even if we replaced lysine with a xenobiotic amino acid, for example, simply relying on one failsafe doesn’t provide a reasonable amount of biosecurity.

[1] Lopez MJ, Mohiuddin SS. Biochemistry, Essential Amino Acids. [Updated 2024 Apr 30]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2025 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK557845/