Hi, Jim!

Shorter timelines for transformative AI (TAI) would make me prioritise more interventions whose effects happen earlier. There will be more change soon if TAI happens earlier, and I believe effects decay faster when there is more change.

A best guess for the probability of an event has implications for the resilience of the best guess. If my best guess is that something is 50 % likely to happen, the probability of me updating towards it being 90 % likely to happen should be at most 55.6 % (= 0.50/0.90).

I recommend research informing how to increase the welfare of soil animals over pursuing whatever land use change interventions naively seem to achieve that the most cost-effectively. I have very little idea about whether increasing agricultural land, such as by saving human lives, increases or decreases welfare. I am very uncertain about what increases or decreases soil-animal-years, and whether soil animals have positive or negative lives. So I do not know whether saving human lives sooner increases welfare more or less than saving human lives later. Assuming that saving human lives increases welfare, I agree doing it earlier increases welfare more if TAI happens earlier.

This seems in tension with what you and @Wladimir J. Alonso say here.

Nevermind. I have been using your estimates for the time in pain as if they do not account for any considerations relevant for interspecies welfare comparisons. However, the sentence below made me think no adjustments were needed to compare your estimates for the time humans and shrimp spend in excruciating pain. So I mistakenly inferred you were accounting for considerations relevant for interspecies welfare comparisons. However, as you say in the same paragraph, you "hold this assumption as temporary until better evidence allows for a more accurate placement of each experience on an absolute scale".

In the Welfare Footprint framework, pain intensities are defined as absolute measures, meaning that one hour of Excruciating pain in humans is assumed to be hedonically equivalent to one hour of Excruciating pain in shrimps, if shrimps were capable of experiencing Excruciating pain.

the Welfare Footprint Framework is intentionally agnostic about correction values for interspecific scaling

In agreement with the above, I have been using your estimates for the time in pain as if they do not account for any considerations relevant for interspecies welfare comparisons.

In the Welfare Footprint framework, pain intensities are defined as absolute measures, meaning that one hour of Excruciating pain in humans is assumed to be hedonically equivalent to one hour of Excruciating pain in shrimps, if shrimps were capable of experiencing Excruciating pain.

This sentence made me think no adjustments were needed to compare your estimates for the time humans and shrimp spend in excruciating pain. So I mistakenly inferred you were accounting for considerations relevant for interspecies welfare comparisons. However, as you say in the same paragraph, you "hold this assumption as temporary until better evidence allows for a more accurate placement of each experience on an absolute scale".

Thanks for the relevant quotes, Mo! 

Thanks for sharing, Aaron!

Nitpick. The flyer says 2024 instead of 2025.

True! Employers paying more will more easily attract people who want to donate more, but I do not think this plays a meaningful role in setting pay.

Thanks for the post, Cameron! I strongly upvoted it. I think it is very valuable to have posts unpacking jobs.

Hi Hannah and Sophie.

The series evaluates the innovation, deployment, and animal welfare impacts of these technologies to help animal advocates and funders identify which developments should be endorsed and what actions should be taken to prevent increased animal suffering.

Will the series assess the cost-effectiveness of potentially promising interventions? There has been significant discussion about how AI will affect animals, but I do not recall any cost-effectiveness analyses.

Thanks for the post, Seema! Staying at a higher-paying less directly impactful job may make sense to donate more. Impact can be doubled by donating 10 pp more of gross income to interventions 10 times as cost-effective as one's own direct work.

Thanks for the analysis, Julia! I strongly upvoted it.

Have you considered estimating the cost-effectiveness of each of the 3 channels? This could help you decide which ones to scale up or down.

Meta-analyses of suicide prevention programs show effectiveness rates ranging from 10% to 40%, with significant heterogeneity based on:

• Intervention type (universal vs. indicated)
• Population (general vs. high-risk)
• Implementation quality
• Follow-up duration

I used 25% effectiveness as a central estimate. This assumes that among youths who would otherwise die by suicide, our intervention prevents death in 1 out of 4 cases.

Why 25%?

• Gatekeeper training programs (similar to our workshops): 10-30% reduction in suicidal ideation
• Crisis hotlines: 10-20% reduction in immediate suicide risk (short-term)
• Combined interventions: Potentially higher, but limited long-term data

Do you think the risk of the students reached via all channels was decreased by 25 %? What you write above seems to have in mind the direct channels (Suicide Prevention Workshops, and Crisis Hotline Youth Callers), but I suspect the effectiveness will be significantly lower for the indirect one (School Awareness Packages). This accounts for 86.1 % (= 10,875/12,635) of the people you estimate to have reached, so I guess you are overestimating the cost-effectiveness. However, not as much as suggested by that fraction. I believe you are overestimating the students you reached via the indirect channel.

A. School Awareness Packages (Indirect reach)

• Sent mental health resource packets to 145 schools
• Assumed 300 students per school (conservative estimate for Japanese high schools)
• Assumed 25% engagement rate (materials opened, read, or discussed)
• Total: 10,875 youths

You are assuming the materials were delivered to all students, but you only contacted at most a few representatives of each school? If so, did they all share the materials with all students? You can add a factor representing the number of students who received the materials as a fraction of all students of the schools you contacted.