Duncan Purvis
Article
Duncan Purvis is a recurring person in the Astral Codex Ten archive, appearing 2 times across 2 issues between February 10, 2024 and January 01, 2025. The archive places it in contexts such as “Duncan Purvis, $30,000, for work on improving flu vaccines”; “ACX grantee Duncan Purvis is trying to improve existing influenza vaccines”. It most often appears alongside Blueprint Biosecurity, influenza A, Metaculus.
Metadata
- Category: People
- Mention count: 2
- Issue count: 2
- First seen: February 10, 2024
- Last seen: January 01, 2025
Appears In
Related Pages
-
- Blueprint Biosecurity (2 shared issues)
-
- influenza A (2 shared issues)
-
- Metaculus (2 shared issues)
-
- Operation Warp Speed (2 shared issues)
-
- Scotland (2 shared issues)
-
- 1918 (1 shared issues)
-
- 1940s (1 shared issues)
-
- 1968 (1 shared issues)
-
- 1977 (1 shared issues)
-
- 1DaySooner (1 shared issues)
-
- 2009 (1 shared issues)
-
- 23andme (1 shared issues)
External Links
Source Context
Recovered passages from the original issue text. When the raw archive preserved outbound links inside the source passage, they are listed directly under the quote.
Duncan Purvis, $30,000, for work on improving flu vaccines. Previous vaccines have included four strains of flu, but one strain recently died out. The World Health Organization, which coordinates flu vaccines, is planning to downgrade to a three-strain vaccine. But Duncan thinks there’s room to improve resistance to future pandemics by reserving the extra slot in vaccines for potentially dangerous influenza A strains. He plans to attend conferences, publish papers, and otherwise build a coalition to make this happen.
<1% chance it’s more than 100x as bad (unprecedented) If you multiply the 5% chance of an H5N1 pandemic per year by the 7% chance of severity ≥ Spanish Flu, you get an 0.35% chance of a Spanish Flu level pandemic this year - one in three hundred. That’s a little higher than base rates - the last pandemic as bad as Spanish flu was smallpox hitting the Indians circa 1500. If we don’t count that one (where would our conquistador equivalents come from?), then the last equally bad pandemic was the Black Death in the 1300s. So we seem to get that level of pandemic once every 500 - 1000 years; a 1/300 chance suggests a 2-3x elevated risk. The Spanish Flu killed about 50 million people. A second Spanish flu could kill more people (because the population is higher), or fewer people (because medical care is better). If we assume those two cancel out, and that a second Spanish flu’s death toll would also be 50 million, then a 1/300 chance of 50 million deaths = 166,666 deaths. In some weird probabilistic expected utility way, about as many people will probably die of H5N1 next year as died in the past year of the Ukraine War. You will have to decide whether this is a reasonable way to allocate mental real estate to different catastrophes. Other Considerations Even if H5N1 doesn’t go pandemic in humans for a while, it is already pandemic in many birds including chickens, getting there in cows, and possibly gearing up to get there in pigs. This will have economic repercussions for farmers and meat-eaters. The CDC and various other epidemiological groups have raised the alarm about drinking raw milk while H5N1 is epidemic in cows. There is an obvious biological pathway by which the virus could get into raw milk and be dangerous, but I haven’t seen anyone quantify the risk level. Epidemiologists hate raw milk, think there is never any reason to drink it, and will announce that risks > benefits if the risk is greater than zero. I don’t know if the risk level is at a point where people who like raw milk should avoid it. Everyone says that pasteurized milk (all normal milk; your milk is pasteurized unless you get it from special hippie stores) is safe. There are already H5N1 vaccines for both chickens and humans; pharma companies are working hard on cows. First World governments have been stockpiling human vaccines just in case, but have so far accumulated enough for only a few percent of the population. If H5N1 goes pandemic, it will probably be because it mutated or reassorted, and current vaccines may not work against the new pandemic strain. Some people have suggestions for how to prepare for a possible pandemic, but none of them are very surprising: stockpile medications, stockpile vaccines, stockpile protective equipment. The only one that got so much as a “huh” out of me was Institute for Progress’ suggestion to buy out mink farms. Minks are even worse than pigs in their tendency to get infected with lots of different animal and human viruses; they are exceptionally likely to be a source of new zoonotic pandemics. Mink are farmed for their fur, but there aren’t as many New York City heiresses wearing mink coats as there used to be, and the entire US mink industry only makes $80 million/year. We probably lose more than $80 million/year in expectation from mink-related pandemics, so maybe we should just shut them down, the same way we tell the Chinese to shut down wet markets in bat-infested areas. ACX grantee One Day Sooner is trying to help the FDA get more resources for Operation Warp Speed style pushes that could expedite approval of pandemic-related vaccines. ACX grantee Duncan Purvis is trying to improve existing influenza vaccines in ways that could make them more effective. ACX grantee Blueprint Biosecurity is working on pan-viral suppression techniques. Conclusions / Predictions All discussed earlier in the piece, but putting them here for easy reference - see above for justifications and qualifications. H5N1 is already pandemic in birds and cows and will likely continue to increase the price of meat and milk.