←
Taal Levi7,563 posts
Taal Levi avatar

Taal Levi

@taaltree

Wildlife ecology and conservation, metabarcoding, eDNA, disease ecology, mathematical modeling. Prof of Fisheries, Wildlife, and Conservation at Oregon State U

◷ Joined June 2009▧ 928 Photos and videos
Sara Beery avatar
⟳ Taal Levi retweeted
Sara Beery@sarameghanbeery·Nov 10

🌍🔬 I’m recruiting PhD students for the BeeryLab at MIT (starting Fall 2027)! We build computer vision and machine learning methods for biodiversity monitoring at a global scale. Learn more about our research here: beerylab.csail.mit.edu/

Post attachment
Taal Levi avatar
Taal Levi@taaltree·Nov 9

Philanthropist-scientist partnerships are an underused win-win opportunity. We already have salaries, and some of us have sustained direct experience in the field, are well-connected, and our jobs allows us the flexibility to do service like this. Many of us would be happy to manage philanthropic giving for free as part of our existing jobs, reducing overhead to extend the impact. /9

Taal Levi avatar
Taal Levi@taaltree·Nov 9

Also in Malawi, I saw the limitation of mosquito-net philanthropy. There was a surplus of nets, enthusiasm for them was low because local concern for Malaria was surprisingly (to me) low, and they were regularly repurposed for fishing and had littered the rivers. This of course won't be true everywhere. They are a net good, but I don't think the EA community has hit the mark for where to be most effective. /8

Taal Levi avatar
Taal Levi@taaltree·Nov 9

I have a PhD student from Malawi who worked with @AfricanParks. They took over management of Nkhotakota Wildlife Preserve from the state, protected it, and brought back large mammals including almost 500 elephants. A very impressive project. We were helping to establish long-term post-translocation monitoring. African Parks is amazing, but their budget is too large for the small amounts that I can give, but I could make an impact on Nkohtakota specifically /7. youtube.com/watch?v=lOXrb7_m…

Taal Levi avatar
Taal Levi@taaltree·Nov 9

These principles apply to tigers in SE Asia. The region is rapidly developing and its population is peaking. If we can prevent tiger extinction now, we face a different conservation landscape in a few decades. I had been teaching a field course in Malaysian Borneo and decided to give to @PantheraCats Taman Negara tiger program in Peninsular Malaysia. /6 panthera.org/blog-post/taman…

Taal Levi avatar
Taal Levi@taaltree·Nov 9

Principles underlying my thinking are 1. extinction is an absorbing state that you can't come back from, 2. Political will and environmental threats change quickly allowing the recovery of species if we can just maintain them through a bottleneck, 3. Bet on the jockey. Find truly dedicated and effective people and give them support. Working in the field allows me to meet people directly and to get a sense for who is doing nonsense that looks good, and who has an executable and impactful plan. A former very smart colleague, Ben Phalan, left science to dedicate his life to saving critically endangered birds with Save Brasil. Here I bet on the jockey. /5 savebrasil.org.br/

Taal Levi avatar
Taal Levi@taaltree·Nov 9

One of the most immediately effective conservation strategies is invasive species removal on islands or fence construction to recover nesting seabirds like albatross. If you can supply funding, that's another island that gets seabirds back. Immediately impactful. I asked a colleague expert on seabirds who recommended Pacific Rim Conservation as a small organization who could put the sums of money that I could give to work. /4 pacificrimconservation.org/

Taal Levi avatar
Taal Levi@taaltree·Nov 9

I was attracted by the team. 100% Guatemalan and very dedicated with clear understanding of the threats. They also have on the ground nest protection and captive rearing of some of the last remaining Central American Scarlett Macaws as well as jaguar population monitoring, but most important was preventing further destruction to the last best forest in Mesoamerica. guatemala.wcs.org/en-us/Wild… /3

Taal Levi avatar
Taal Levi@taaltree·Nov 9

I had been working in northern Guatemala on carnivore research and learned that drug traffickers and illegal cattle ranchers had destroyed half of Laguna del Tigre National Park, which was the largest park in the Maya Biosphere Reserve. @Guatemala_WCS had a clear strategy for preserving the other half with permanent human presence and a community-run non-timber forest products concession so that a permanent stakeholder defended that land that the park service could not. /2

Post attachment
Taal Levi avatar
Taal Levi@taaltree·Nov 9

The tech/EA community would do well to seek out well-connected experts because solving issues is complex, and identifying impactful philanthropic strategies is not easy. I have used a large collaborative international network to make giving decisions. I chose to prioritize urgency, a clear plan for impact, and non-redundancy (something important that wasn't already attracting attention) /2

Taal Levi avatar
Taal Levi@taaltree·Nov 9

A few years ago I was given a large sum of money and asked to donate it to maximize impact. Because I am a wildlife scientist, I focused most giving in that space, but how to decide what maximizes impact? Here are some places I've donated to (~$650k now) and some thoughts for philanthropists / 1

I don’t know what I find more ridiculous: someone attacking donating money to *checks notes* mosquito nets, or the fact that they’re an academic that’s seemingly unconcerned with how terrible their argument is.

Taal Levi avatar
Taal Levi@taaltree·Nov 5

If the blue whale population growth trajectory continues, we'll be back to the pre-whaling baseline when my daughter is my age. I'd love to see that!

New assessment of the greatest of all time: Antarctic blue whales, showing how whaling reduced them from >200,000 to a low of just 418 before slow recovery at 4.5% per year to more than 4000 Rand et al. [incl. me] doi.org/10.3354/esr01557

Post attachment
Taal Levi avatar
Taal Levi@taaltree·Nov 3

This was almost definitely a bias report made by a graduate student teaching an online course, which is why they refer to both their supervisor and the department chair. Anyone can report anything they want, and it should surprise nobody that a graduate student instructor might be overzealous. It would reflect poorly on the university if the administration responded inappropriately, but we have no such information.

An Oregon State University instructor reportedly filed a bias report against a student for arguing that there are two biological sexes in a class assignment. The university says its Bias Response Team doesn’t punish protected speech, but you don’t have to punish students to silence them. The prospect of being subjected to vague “restorative processes,” or having your views documented in a university file, is chilling enough. And students are already reluctant to speak up. In FIRE’s College Free Speech Rankings survey, 52% of Oregon State students reported at least occasionally self-censoring with professors, while 41% expressed discomfort disagreeing with professors on controversial issues in written assignments. Professors can grade students on their arguments, research, and understanding of course material. But reporting them for their personal views is another matter entirely, and teaches students a dangerous lesson: Keep your opinions to yourself. @OregonState must scrap its Bias Response Team.

Post attachment
Taal Levi avatar
Taal Levi@taaltree·Oct 23

Feel free to DM me issues that come up. I want this to be useful.

Please share with the community. I'm looking for beta testers for CascadiaTools, an AI coding agent harness for ecological, wildlife, fisheries and conservation-genetic analysis. It is available as a @claudeai plugin. Our aims are to: (i) argue that the trust ecologists place in vetted R packages can be extended to analyses run by coding agents, if the agents call vetted code instead of writing it; (ii) state the design principles that make agent-produced analyses checkable; (iii) describe the architecture shared by all modules; (iv) report how each of 29 modules was built and independently verified, including the limits each verification exposed; (v) show how a researcher with no coding experience uses it, from question to publication figures; and (vi) outline how vetted libraries for coding agents could become a standard part of ecological practice. /1 cascadiawildlifelab.org/tool…

Taal Levi avatar
Taal Levi@taaltree·Oct 22

Please share with the community. I'm looking for beta testers for CascadiaTools, an AI coding agent harness for ecological, wildlife, fisheries and conservation-genetic analysis. It is available as a @claudeai plugin. Our aims are to: (i) argue that the trust ecologists place in vetted R packages can be extended to analyses run by coding agents, if the agents call vetted code instead of writing it; (ii) state the design principles that make agent-produced analyses checkable; (iii) describe the architecture shared by all modules; (iv) report how each of 29 modules was built and independently verified, including the limits each verification exposed; (v) show how a researcher with no coding experience uses it, from question to publication figures; and (vi) outline how vetted libraries for coding agents could become a standard part of ecological practice. /1 cascadiawildlifelab.org/tool…