Notes on Apollo report on biodefense

By Linch @ 2022-07-23T21:38 (+69)

This is a linkpost to https://biodefensecommission.org/reports/the-apollo-program-for-biodefense-winning-the-race-against-biological-threats/

The Bipartisan Commission on Biodefense’s Apollo Program for Biodefense (invoking the moon landing project) is an ambitious proposal to radically change the landscape of biosecurity and pandemic defense with a bunch of technical innovations. It was recommended to me some time ago but I only got around to reading it in June. Below are my (mostly unfiltered) notes, including some fast/quick hot takes. The original intended audience was a) myself, and b) teammates (including future teammates) working on researching future megaprojects. But hopefully it can be hopeful to other EAs as well.

Key Takeaways

• I originally planned to read this report with an eye out for inspiration on what EA megaprojects can Rethink Priorities ideate, analyze, and/or inspire people to work on. But reading this got me thinking: Perhaps trying to fund/do parts of the Apollo Program for biodefense should be considered a premier EA/longtermist megaproject in its own right?
• The technologies they recommend run the gamut between a) prevention, b) detection, c) slowing down/stopping spread, and d) resilience/recovery. 
• The focus on detection seems higher than what I naively would guess is normal in pandemic preparedness circles.
• I note that some EAs worked on the report.
• Many of the recommendations might actually be enacted (!!). They were in Biden’s budget proposal this year)
• From the report and the surrounding context, I’ve overall updated positively on the value of policy/politics.
• The report itself is short and highly readable. If any of the sections in my notes sound interesting, feel free to jump ahead to that section in the report and read it in full!

Executive Summary and Introduction (Pgs 1-7)

Choice Quotes

“Even the most ambitious program (about $10 billion annually) would be a small fraction of the current cost of the COVID-19 pandemic and an investment in our health, economy, and national security.”

“The existential threat that the United States faces today from pandemics is one of the most pressing challenges of our time; and ending pandemics is more achievable today than landing on the moon was in 1961.”

“The COVID-19 pandemic has resulted in more American deaths than World War I, the Korean War, the Vietnam War, the Gulf War, the War in Afghanistan, and the Iraq War altogether.8 COVID-19 will likely cost the United States over $16 trillion.9 We spend billions preparing for other threats to American lives, which may or may not occur. Spending on biological risk reduction would be far less than the significant cost of continuing to let future pandemics devastate the United States again”

“We must stop fighting the last war. We need new strategies and defenses. Through The Apollo Program for Biodefense, we can make invisible biological enemies visible and take pandemic threats off the table by the end of the decade. ”

“Rogue states wishing to challenge American primacy could take advantage of the Nation’s disease-stricken state to test our country’s ability and willingness to maintain global order. 

“The visibility of our vulnerabilities increases the likelihood of biological attacks in the future,21 as do the continued breakthroughs in biotechnology that lower the technical barriers to producing biological weapons. The likelihood of an accidental release of pathogens from laboratories may also increase as nations build more high containment laboratories and conduct more biomedical research.22”

“The Nation has a history of taking on grand technological challenges in times of need, such as the Manhattan Project (to split the atom), the Interstate Highway System (to create a network of highways to connect the entire nation),26 and the Global Positioning System (to enable geolocation anywhere on or near the earth).27 Those efforts share similarities in scale, ambition, necessity, and difficulty of execution, and demonstrate our ability to engage in systematic, large-scale execution and funding of a goal-oriented and coordinated effort to achieve the technological capabilities the Nation needs.”

“Previous national grand challenges focused on singular goals, such as landing on the moon or harnessing the power of the atom. The Apollo Program for Biodefense would not be limited to a singular goal (e.g., a moonshot), but would achieve multiple groundbreaking technological advances with a single, overarching goal—to gain technological superiority over biological threats. We envision a time when people will look back and wonder how we ever let infectious diseases wreak havoc on society and how we tolerated seasonal influenza, let alone COVID-19 and biological weapons.”

RECOMMENDATIONS

• Implement the National Blueprint for Biodefense
  • Recommendations by the Commission in an earlier, 2015, report
    • TODO: Read earlier report and see differences
• Produce a National Biodefense Science and Technology Strategy
• Produce a Cross-Cutting Budget
• Appropriate Multi-Year Funding 

TECHNOLOGY PRIORITIES

• Vaccine Candidates for Prototype Pathogens
• Multi-Pathogen Therapeutic Drugs in Advance of Outbreaks
• Flexible and Scalable Manufacturing of Pharmaceuticals
• Needle-Free Methods of Drug and Vaccine Administration
• Ubiquitous Sequencing
• Minimally- and Non-Invasive Infection Detection
• Massively Multiplexed Detection Capabilities
• Point-of-Person Diagnostics
• Digital Pathogen Surveillance
• A National Public Health Data System
• An Integrated National Pathogen Surveillance and Forecasting Center
• Next-Generation Personal Protective Equipment
• Pathogen Transmission Suppression in the Built Environment
• Comprehensive Laboratory Biosafety
• Technologies to Deter and Prevent Bad Actors

Linch’s hot takes:

• I originally planned to read this report with an eye out for inspiration on what EA megaprojects can RP ideate, analyze, and/or inspire people to work on. But reading this got me thinking: Perhaps trying to fund/do parts of the Apollo Program for biodefense should be considered a premier EA/longtermist megaproject in its own right?
• It also makes it relatively clear to me why we may need nation-state actors to coordinate funding for reducing extreme worst-case pandemics – it sure seems like EA has far less than enough $s to do this well (in addition to mention all the other things longtermist EA wants to do). 
• The “high range” for their budget is 10 bil for 10 years, or ~100 billion total. If this can reduce biological x-risk this century from Ord’s 3% to 2.5%, this roughly translates to an x-risk reduction of 0.5% in absolute terms, or saving >40M expected present lives. At $5000/cost to save a marginal life, this 100B compares favorably (but is maybe surprisingly in the same OOM) as AMF (which would be >200B to save >40M lives).
  • Note that this might well be a large underestimate, as there will counterfactually be plenty of GCBRs that are nowhere near extinction-level that this program might avert.
• The report itself has a fairly classical EA argument: ~10B/year for 10 years to prevent another pandemic, COVID itself cost the US alone >16T, 100B vs 16T is >100:1 cost-effectiveness.
• The language of the report was interesting: a combination of US boosterism, anti-pandemic boosterism, and actually-valuable-to-me information.
• I privately note in several cases where the evidence in the summary/intro seem slated for one angle (e.g noting the ways pandemic risk/mortality increased but not the plausible ways in which it decreased).
  • I chalk up the difference in the report and what I’m used to as a result of just different cultural/communication norms. 
• It might actually happen???? 
  • And if this goes well, perhaps this will be one of the largest visible accomplishments in biosecurity in recent years.
  • https://www.statnews.com/2021/09/03/biden-wants-65-billion-for-apollo-style-pandemic-preparedness-program/ 
  • https://www.domesticpreparedness.com/updates/ten-priorities-identified-in-commissions-apollo-program-for-biodefense-addressed-in-presidents-budget-request/ 

Recommendations (Pgs 8-9)

• “Congress should provide multi-year appropriations to implement The Apollo Program for Biodefense. This funding should be commensurate with the goals of the program and aligned with the magnitude of the threat, as opposed to historical appropriations. Funding should also include multi-year budget authority to allow agencies to procure systems and medical countermeasures that take years to develop and produce. ”

Appendix A: Technological Priorities (Pg 11-25)

This appendix describes in more detail the technological priorities recommended earlier.  

• VACCINE CANDIDATES FOR PROTOTYPE PATHOGENS
  • It’s hard to do vaccination for novel pathogens ahead of time. However, there’s a relatively limited number of known virus families and lessons learned from developing one vaccine can transfer well to others (LZ: also there might be cross-immunity considerations)  
    • “Therefore, by investing in vaccines for at least one prototype pathogen in each of the 25 viral families known to infect humans, we could reduce the global burden of infectious disease while simultaneously preparing for the next unknown biological threat. These efforts would also help develop a strong and diverse research community, better prepare us to address new threats rapidly as they emerge, and prevent the need for difficult and blunt interventions”
  • “We should continue research to validate generalizability. When we need to use the same vaccine approach in the future, rapid entry into Phase 1 clinical trials will be possible by leveraging data from previous clinical trials.”
  • LZ’s hot takes:
    • I don’t see a clean line for how vaccine development can get us safety from x-risk-level biothreats, other than chained reasoning like “deterrence by denial”-style arguments
    • I’m moderately worried about infohazards in novel vaccine development. 
• MULTI-PATHOGEN THERAPEUTIC DRUGS IN ADVANCE OF OUTBREAKS
  • “To ensure that we have a multitude of drugs ready at the beginning of the next pandemic, we need to make investments in the development of multi-pathogen therapeutics—those that can be effective against multiple phylogenies of viruses.”
  • “Previous efforts to develop multi-pathogen therapeutics have largely targeted direct-acting small molecule antivirals. However, new modalities are emerging that may result in increased breadth and potency and which warrant extra investment, including host-directed antivirals and monoclonal antibodies targeting regions conserved across multiple viral species.”
  • LZ’s hot takes:
    • Similar issue as with faster vaccine development, though the infohazard stuff is less concerning
    • Also naively I feel slightly more optimistic about new broad-spectrum antivirals covering new territory than I do with vaccines. (but VERY far from a technical expert here!) 
• FLEXIBLE AND SCALABLE MANUFACTURING OF PHARMACEUTICALS
  • “Following the successful development of therapeutics and vaccines against a novel pathogen, they must be rapidly manufactured at scale, both initially for clinical trials and later for distribution to the public”
  • The report talks about how there’s currently a bunch of focus on developing vaccine platforms, but less on developing therapeutic platforms
  • LZ’s hot takes:
    • I see an even less clean case for why scaling pharma can avert existential risk than targeted defensive technological development
    • They briefly talked about recombinant proteins and cell culture, which I thought was interesting. A shower thought here is something like a lot of the work in cultured meat (that I spent a decent fraction of 2021 studying) goes into this, so there might be a few surprising synergies. 
• NEEDLE-FREE METHODS OF DRUG AND VACCINE ADMINISTRATION
  • Healthcare workers may be in short supply in a global pandemic, and many people are afraid of needles.
  • “The U.S. government should invest in the advancement of the aforementioned technologies which enable transdermal (microarray patches), intranasal, inhalable, and oral delivery of drugs and vaccines. We can deliver pharmaceuticals that use these methods by developing them for infectious diseases for which needle-based delivery is currently predominant”
  • LZ’s hot takes:
    • Same as above.
• UBIQUITOUS SEQUENCING
  • Current sequencing is already impressive. However, it does not allow arbitrary targeting. 
  • “Metagenomic sequencing, the reading of all genetic material from a sample, offers advantages that many other capabilities struggle to rival.”
  • “Despite continued advances, often outpacing Moore’s law, sequencing technology has critical bottlenecks to achieving the ubiquity, simplicity, and affordability needed.50 If realized, sequencing could become routine in the clinical setting, as well as in high-risk low-resource areas of the world, expanding access to the most capable diagnostic tool. Sequencing could serve as the diagnostic for diseases generally and permit novel pathogen detection early and beyond our borders”
  • LZ’s hot takes:
    • In relative terms, this seems extremely promising from an x-risk reduction perspective.
    • I have a few theoretical concerns about sequencing, but hopefully they aren’t big issues.
• MINIMALLY- AND NON-INVASIVE INFECTION DETECTION
  • Currently infection detection requires the host to be sick before we’re on the alert. The world doesn’t need to be that way.
  • “We are on the verge of the ability to detect whether the body is currently infected with any pathogen, known or unknown, through the interrogation of host biomarkers. Increasingly, we can also detect infection indicators non-invasively through advances in wearables52 and volatolomics.”
  • “Minimally invasive technologies (i.e., those that permit sample acquisition without pain, discomfort, inconvenience, or risk) would also facilitate molecular diagnostics for the identification of pathogens. ”
  • LZ’s hot takes:
    • First order-impression is something like “big if true”
    • It does seem very promising
    • Clearly combos well with metagenomic sequencing
• MASSIVELY MULTIPLEXED DETECTION CAPABILITIES
  • “Single-pathogen diagnostics require clinical suspicion and are not readily available […’ Multiplexed detection capabilities address these challenges and bring new benefits by simultaneously testing for multiple pathogens, resistance genes, biomarkers, and analytes in a single simple assay.”
  • “​​Syndromic panels via multiplexed PCR assays (e.g., those used to test for approximately 25 of the pathogens most associated with respiratory infections) are currently available in many parts of the world, but do not include most known pathogens. While adequate for most presentations of infectious disease, crucially, these panels do not cover less common and novel pathogens. Massively multiplexed panels can address these limitations by including virtually all known human pathogens and even detect novel pathogens based on conserved sequence homology (i.e., the ability to detect similar regions in a pathogen’s genetic tree).”
  • “New CRISPR-based massively multiplexed panels are particularly promising.57”
  • “We should prioritize techniques enabling the tests to move out of centralized laboratories, and especially those that can operate in resourceconstrained settings.”
  • LZ’s hot takes:
    • Clearly combos well with above two technologies
    • I see a relatively clear line for x-risk reduction via this plus the above.
• RAPID POINT-OF-PERSON DIAGNOSTICS
  • “Rapid point-of-person diagnostics, also known as point-of-need diagnostics, are tests that can rapidly identify an infection wherever the individual is located.”
  • “Point-of-person diagnostics should be considered public health instruments”
  • “Rapid low-cost tests also allow for repeated use, which can be essential for novel pathogens with unknown incubation time”
  • LZ’s hot takes:
    • I think I’m more bullish on easy-to-use diagnostics than medical countermeasures, and less bullish on them than the other early detection/early warning stuff
• DIGITAL PATHOGEN SURVEILLANCE
  • “By leveraging advances in machine learning, and in particular natural language processing,62 we can continuously track vast amounts of data and filter the noise to provide relevant information to public health experts. This information is useful to prompt further investigation, allocate resources, and inform clinicians and public health authorities about potential pathogens to consider in their routine work”
  • LZ’s hot takes:
    • Seems like one part of the pipeline for really good disease forecasting
• A NATIONAL PUBLIC HEALTH DATA SYSTEM
  • LZ hot takes: 
    • Seems good to do, I remember horror stories from covid like the Californian government telling my friend in Michigan to come back to California to report a positive covid test. 
    • Don’t see the x-risk angle tho
• A NATIONAL PATHOGEN SURVEILLANCE AND FORECASTING CENTER
  • “An integrated real-time national pathogen surveillance and forecasting center with advanced capabilities to detect and model naturally occurring, accidentally released, and intentionally introduced biological threats does not currently exist.”
  • “Aggregating diverse data sources in real-time and forecasting infectious disease outbreaks are necessary to prevent or rein in the spread of biological threats. ”
  • LZ hot takes:
    • This is near and dear to my heart, but my best guess is that the model as presented will probably be too slow to pick up on the “hot signals” of an otherwise lethally undetected pandemic
• NEXT-GENERATION PERSONAL PROTECTIVE EQUIPMENT
  • “Personal protective equipment (PPE) can be used to protect against a broad-spectrum of biological threats. However, the current state of PPE burdens its users, requires experience in proper usage, is seldomly reusable, is not widely available to all populations, and does not properly fit everyone (e.g., children).68 Additionally, since the primary goal of PPE is to prevent the wearer from becoming infected, not enough emphasis has been placed on preventing the wearer from infecting others. ”
  • “To develop the next generation of PPE, we should make innovations in the following areas: 1) reusable, sterilizable, and self-disinfecting equipment; 2) modular designs responsive to a wide range of threats, including those which go beyond biological threats; 3) personalization to ensure adequate protection, comfort, and attractiveness; 4) rapid production from widely available materials without supply vulnerabilities; 5) the ability to neutralize pathogens; 6) sensing capabilities to detect potential exposures; and 7) protection beyond traditional masks, respirators, gloves, gowns, etc., that safeguard the wearer without burden.”
  • “The government should invest in and incentivize the development of these PPE innovations through inducement prize challenges, intramural and extramural research and development efforts, advance purchase commitments and consistent acquisition, and use-inspired basic research programs, such as the Defense Advanced Research Projects Agency (DARPA) Personalized Protective Biosystem effort. ”
  • LZ hot takes:
    • I’m pretty excited about this
    • In addition to the use cases mentioned, if we have only a small number of very paranoid people use next-gen PPE during peacetime, this might be enough to stave off total civ collapse.
    • It might also be appropriate to use some of them in lab settings, in refuges, and in other high-importance or high-risk environments
• PATHOGEN TRANSMISSION SUPPRESSION IN THE BUILT ENVIRONMENT
  • “To reduce the effective transmissibility of most airborne, droplet, vector-borne, and fomitetransmitted pathogens, we should make investments in: 
    • • affordable air filtration and sterilization systems 
    • • deliberate design of airflows 
    • • self-sterilizing surfaces 
    • • easily sterilized materials, robust against harsh sterilization
    •  • robotic and autonomous integrated sterilization 
    • • fomite neutralizing technologies 
    • • integrated real-time pathogen sensing capabilities”
  • “part of a modernization effort, the federal government should invest in technologies to retrofit current infrastructure, such as HVAC systems and public transport, ”
  • LZ hot takes:
    • I’m pretty excited about this, it’s near my heart
    • In addition to the mentioned things, I’m also interested in UVC sterilization and resonance stuff. 
    • Obviously built-environment suppression is super relevant to pandemic-proof shelters
    • I’d also be excited for EA office spaces to prototype earlier versions of this
• COMPREHENSIVE LABORATORY BIOSAFETY 
  • “Our risk tolerance in laboratories worldwide74 working with biological threats should be comparable to that of air travel, where safety is engineered into the airlines and airports, and monitoring occurs constantly to detect and prevent human-generated and technology-based accidents. A constant focus on and prioritization of safety ensures that the complex and previously risky nature of flight can be undertaken safely”
  • “While training personnel is essential and the core of biosafety,76 insider threats should also be more seriously considered, and safeguards put in place to deter and prevent any malicious behavior.”
  • “Additional funding is necessary for the study of laboratory accidents and the development and testing of new capabilities and tools to achieve comprehensive laboratory biosafety systems. These should be tested in safe environments, continuously incorporated into current high-containment labs, and ultimately integrated into all biosafety labs. ”
  • LZ hot takes:
    • I think this is extremely important, maybe among the highest importance on this list so far
    • The report seemed a bit vague on the details of how we can get lab biosafety.
• TECHNOLOGIES TO DETER AND PREVENT BAD ACTORS
  • The ability to investigate, analyze evidence, and attribute deliberate biological events is essential for both deterrence and response to a deliberate or accidental threat.77 As tools are developed and the barriers to engineering pathogens continue to decrease, the number of possible actors may increase. Technologies are required to ensure safety is built in and capabilities developed in advance to prevent and deter action. 
  • Unfortunately, biological attribution, genetic engineering detection, and microbial forensic techniques have only made small strides since the anthrax attacks of 2001.
  • “Once developed, these capabilities could be broadly deployed and integrated into routine laboratory, clinical, and environmental settings as sentinels monitoring for engineered pathogens[...] With additional dedicated funding to research, develop, acquire, and operate such technologies, as well as maintain the relevant repositories, we could establish a robust and known capability to detect, analyze, and attribute biological threats.”
  • LZ hot takes:
    • Not sure why they saved the best for last. Clearly this is the most important. 
    • Existential-threat level pandemics don’t happen naturally all that often, clearly we must primarily be worried because of worries of human negligence or malfeasance. 
    • So figuring out ways to deter bad actors seem extremely important.

Acknowledgements

Thanks to Rethink Priorities for paying for my time to do all sorts of odd readings in pursuit of identifying future megaprojects. Thanks to Chris Bakerlee for suggesting that I read this report originally, and Renan Araujo for coworking with me enough for me to actually get the notes out. Thanks especially to everybody who wrote the Apollo report, and for the people working in worst-case biosecurity more generally. You guys are an inspiration to the rest of us, for envisioning and implementing the concrete projects we can accomplish, to safeguard the future better.