Boston-based, NAO Lead, GWWC board member, parent, musician. Switched from earning to give to direct work in pandemic mitigation. Married to Julia Wise. Speaking for myself unless I say otherwise. Full list of EA posts: jefftk.com/ânews/âea
Jeff Kaufman đ¸
Karma: 17,759
most average only 1 year lifetime or less with decreasing efficacy over time, have to replace the entire system when itâs spent, adds to electricity cost when in use, and you need to install a lot of them for it to be effective because when run at too high power levels they produce large amounts of hazardous ozone
I think this is true for many options, but not the Aerolamp:
It is built around the Ushio Care222 B1. This is a long-lasting design, rated for 10,000 hours before falling below 70% output. Thatâs one year of 24â7 usage, five years of working hours, or much longer if run less often.
They do use electricity, but at 11W itâs a negligible 1-2¢ per hour.
Even one Aerolamp cleans a lot of air. You can model efficacy with Illuminate.
Well filtered lamps do not produce a lot of ozone. Unless your building is incredibly well sealed ozone levels would go up if you opened a window. If youâre very concerned you can run an air purifier that includes an activated carbon layer, which many do.
Thanks for writing this! I think this is a direction that it would be valuable for more people to move in, on the margin.
On the other hand, as someone who went pretty far in this direction and has since backed off some, I think there can be some pretty strong trade-offs here that I donât see you getting into, around putting oneself in a position where you might spend time to save money in ways that are not actually worth it.
Letâs say you have a directly useful job, or you are earning to give in a field where your long-term compensation is going to track your overall productivity. These are both situations Iâve been in, and I think theyâre reasonably common? There tends to be a lot of opportunities to choose between more work time or less spending. Ex:
-
If my house needs repairs, itâs generally much cheaper for me to do it myself rather than hire someone, and Iâve done a lot of this over the years. To the extent that this is something I enjoy doing, itâs not a bad hobby! But more recently, Iâve been spending more of my âhobbyâ time on kinds of extra work for my org (tasks that I find less draining than my usual work). If a big repair came up, I think it would likely be actually a large mistake for me to put a lot of time into resolving that myself if that meant doing less of my primary work.
-
Say Iâm going to take a week off of work to spend with family. My work has unlimited vacation, so I could choose to take an extra day off on each end so I could travel by bus instead of spending more on a plane. But since itâs better for me to put more work in, being willing to spend the money on the plane is better.
-
How big an area they cover depends on how long the sight lines are: the more air the light can travel through the more it can clean. Itâs not linear, though, and there are different effects for different pathogens because some are quickly inactivated with low levels of UVC and others need more. The modeling tool Illuminate can be very useful here; hereâs something I wrote up when I was figuring out what made sense for the dance hall: Assessing Far UVC Positioning.
There are very few providers, and hardly any of them sell an off-the-shelf product. You usually canât just buy a lamp to try it outâyou have to call the company, get a consultation, and often have someone from the company come install the lamp. Itâs a lot of overhead for an expensive product that most people have never heard of.
This has changed! You can buy an Aerolamp for $500. I have one for my own use, and my dance organization uses four.
[EDIT: expanded this into a post]
What is our estimate for how logarithmic the utility functions of common EA charities are?
The argument is less about how the value of the marginal dollar falls off, but instead about how smoother funding is usually much more valuable to projects. Imagine Iâm trying to decide between donating $100k today and $100k in a year, or $200k today. If I expect everyone else together will be donating $1M today and $10M in a year, then I should probably pick the $200k today. The idea is that the charity can probably much more productively use $1.2M today and $10M in a year than $1.1M today and $10.1M in a year.
EDIT: added something like this to the post
They want to work at mission-driven organizations like SecureBio, ⌠Unfortunately, these organizations are often too busy working on the problems themselves to invest in younger, inexperienced trainees.
Flagging that while SecureBio has hired mostly senior people, we do have a relatively junior job posting open at the Nucleic Acid Observatory for a Bioinformatics Engineer. Weâd love to see more applicants, and weâre offering a significant referral bonus if you know anyone who might be a good fit!
Weâre not currently looking for engineers right out of undergrad (â2+ years of professional experience in industry or academiaâ) but itâs also not five to ten years skill building first. And in our case itâs true that they âdonât need another degreeâ.
Good question! Reworded a bit:
Far-UVCâs strongest effect on viruses and bacteria is to damage their proteins, so they canât infect us. This same protein absorption is what makes far-UVC safe for humans: unlike the UVA and UVB in sunlight that penetrate deep enough to cause sunburns and skin damage, far-UVC is so strongly absorbed by proteins in the dead outer layers of our skin and eyes that it doesnât reach the living cells beneath. Because of this shallow absorption, far-UVC has minimal effects on the human body, and is safe to install in our homes, schools, and workplaces
I thought a lot about frugality when I was getting involved in EA (ex: Living Frugally So We Can Give Away More, from 2010), but I think I (and some other early EAs) could be shortsighted here. For example, in retrospect I think it would have been really valuable for @Julia_Wiseđ¸ and me to meet other EAs in person in the UK, but we didnât go there until 2014. And only then because we could tack it it onto travel for my sisterâs wedding.
A focus on minimizing spending can also be a distraction from other ways of increasing your impact. For example, when I wrote that post I was earning (all numbers in 2025 dollars) $106k in a research group. Two years later when I realized I should be trying harder to earn money and Carl Schulman suggested I join Google, my starting salary was $149k and in my first full calendar year I earned $301k. Very quickly I was able to donate more than I had been earning before. A focus on increasing earnings would have resulted in more donations.
On the other hand, I do think some frugality is really valuable. If we had let our expenses grow proportionately during the period when I was earning $700k+ I could easily have become trapped earning to give, but frugality (âa low personal burn rateâ if you want to appeal to startup folks) allowed me to leave Google to join an early-stage biosecurity project that spun out into a non-profit that still canât afford to pay super well. And it has allowed me to take a voluntary salary reduction, allowing the non-profit to get more done with the same funding.
Overall, I think it would probably be good for EA to be moderately more frugal, but to be very aware of the downsides in burnout and turning people away.
You might also be interested in the top comments on Free-spending EA might be a big problem for optics and epistemics (posted at the height of the FTX-funding era) for some discussion on the pros and cons of EAâs more frugal past.
Aww, thanks!
Better yet, donate.
This takes you to a GWWC page that says:You can donate to several promising programs working in this area via our donation platform. For our charity and fund recommendations, see our best charities page.
But if you click through those links you can see there isnât a biosecurity fund, and the list of biosecurity-related programs you can donate on GWWC are just NTI and CHS. Which are not bad options, but as someone in the field theyâre not where I think funding would go farthest. It would be really great if there were a bio fund, or a bio evaluator!
(Disclosure: I run a project that would plausibly be funded by such a thing)
@fezzyđ¸ one important difference (especially important to me!) since we last talked is that the NAO just opened a position for a software engineer; consider applying? https://ââsecurebio.org/ââcareers/ââbioinformatics-engineer-2025/ââ
I wouldnât recommend getting into cybersecurity as a path to biosecurity. From the outside, the main thing that transfers in the actual work is the mindset, but a very large majority of skills in both areas are domain specific. On the other hand, I do think there are valuable things that need doing in cybersecurity, and biosecurity efforts nearly always also have cybersecurity needs.
Iâm guessing support for upper room UV would be even higher than far UVC because there is very little exposure of people for upper room?
Maybe? Or maybe people would be worried about whether weâd deploy it properly.
And in-room filtration would be high support as long as itâs not loud?
I donât think anyone has an objection to in-room filtration, and this is something Iâd be happy rolling out without polling the community. But the real problem isnât the noise (though that is significant, and Iâm excited about the Big Quiet Fan project), itâs the sheer number of purifiers weâd need around the room to handle ~200 people.
The smallest one I know of is the kidsâ Flo Mask.
Thanks! Iâll pass this along to Simon who leads our swab sampling work.
On reusable respirators, theyâre worse to significantly worse for intelligibility than disposables. If youâre only talking about what to wear when not talking (ex: listening to talks) then this doesnât matter, but if youâre considering masking for 1:1s or group discussions it becomes pretty important.
This is probably a regional thing: I donât see the 3M 4251 or other disposable respirators for sale in the US. My guess is the cost difference youâre seeing is due to comparing a US-market (6200) vs UK-market (4251) masks on UK Amazon?
Definitely! Storing replacement filters in addition to the masks themselves seems good.
If youâre trying to flag something engineered, I think the general public makes more sense than people who work with animals. What we do at the NAO (we really need to write up a page on this!) is visit busy public places, put out a sign, and ask if people are interested in swabbing their nose for science. People swab their own noses, drop them in a shared container, and we pay a small amount per swab. Itâs under IRB, and if you were looking to do something similar we could share our IRB documentation?
also use it to not spread random colds /â flu /â covid
If your goal is not spreading (âsource controlâ) then you should wear one that filters the output (no valve). I think the ElastoMask Pro is a good option there: manufacturer, Amazon.
Efficacy drops pretty slowly: the mask body itself shouldnât degrade, and the filters mostly degrade by being clogged with dust (which in normal environments is at least 6m). You can get extra filters cheaply.
how bad is it if I sneeze inside
Not an expert in this, but my understanding is itâs gross but not dangerous. You could consider also having some disposable masks to wear when youâre sneezy?
Thanks for clarifying! I do think in a context like this one, where people are thinking about why offices etc donât install far-UVC, your friendâs phrasing is likely to confuse people. For example, if I recommended someone not buy a car because it only had a one-year lifespan, I think theyâd be grumpy if they later learned I meant it would only last one year of 24â7 operation. When we talk about âlifespanâ weâre normally bringing in assumptions about expected usage.