Hmm, I think there’s some sense to your calculation (and thus I appreciate you doing+showing this calculation), but the $6.17 conclusion—specifically, “engagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum, which suggests the marginal cost-effectiveness of running the EA Forum is negative”—strikes me as incorrect.
What matters is by how much engaging with the Forum raises altruistic impact, which, insofar as this impact can be quantified in dollars, is far, far higher than what one would be willing and able to pay out of one’s own pocket to use the Forum. @NunoSempere once estimated the (altruistic) value of the average EA project to be between 10 and 500 million dollars (see cell C4 of this spreadsheet; here’s the accompanying post). That is far higher than the actual dollar cost of running the average project. (Indeed, if one is funded by EA money, then one’s generation of altruistic dollars needs to outpace one’s consumption of actual dollars—and by a large multiplier, if one is to meet the funding bar.)
Going back to Nuño’s spreadsheet: If I make the arrogant assumption that I’m within an order of magnitude of Ben Todd, impact-wise, then that means my lifetime impact is at least 10 million dollars. Assuming linearity (which isn’t a great assumption, but let’s go with it for now) and a career length of 40 years, this means my impact over the past 4 years has been ≥1 million dollars.[1] In that time, I’ve spent maybe 500 hours on the EA Forum.[2] Meanwhile, I’d say that the Forum has contributed greatly to my intellectual development, i.e., added at least 20% to my impact. (The true percentage may in fact be much higher, because of crucial considerations that the Forum has helped me orient toward, but let’s lowball things at 20%, for now.) This would imply that my impact has been amplified by at least $200,000/(500 hours) = $400 per hour spent on the Forum. ([Insert usual caveats about there being large error bars.]) Contrast with your $6.17.
(I did this calculation on myself not because I’m special, but because I know what the numbers are for myself. I’d guess that the per-hour bottom line for other Forum users would be ~similar.)
We can now go one step further, and estimate the Forum’s “altruistic dollar generated per actual dollar spent” multiplier to be at least 400⁄6.17 ≈ 65. Embarrassingly, I don’t know how this compares against today’s funding bar,[3] but seems very plausible to me that it’s above.
(Nonetheless, people may still not pay $6.17/hour to use the Forum because $6.17/hour is a non-trivial cost considering people’s actual incomes. Additionally, people are just used to being able to browse the internet for free, and so I suspect many wouldn’t do the expected value calculations and reach the “rational” conclusion that they should in fact pay.)
That is, 500 hours consuming the Forum’s content. I’ve also spent time writing on the Forum, but if we model the Forum as a two-way market, with writers and consumers, and say that it’s the consumers who benefit from being here, then it doesn’t make sense to include my writing time. (Also—and perhaps more relevantly—I don’t think writing time gets counted by the Forum’s analytics engine as engagement time if it’s spent mostly in a Google doc.)
Further detail: What really matters is what the multiplier is on the margin (i.e., what it is for the last dollar being spent on a project), rather than what it is for the project as a whole.
Thanks for sharing your thoughts, Will! I agree what matters is the additional altruistic impact caused by engaging with the Forum. However, I think my point holds as long as people have accurate views about how to maximise their altruistic impact.
For example, if you believed “factual impact of your marginal hour on the Forum”—“counterfactual impact of this hour” < “impact of donating 100 $[1] to the organisation or project you consider the most cost-effective”, and using the Forum costed 100 $/h, I think you would have a greater altruistic impact by your own lights by spending less time on the Forum, and donating the savings. Do you agree?
Analogously, if the user spending the marginal hour on the Forum believed “factual impact of their marginal hour on the Forum”—“counterfactual impact of this hour” < “impact of donating 6.17 $ to the organisations or projects they consider the most cost-effective”, and using the Forum costed 6.17 $, I think they would have a greater altruistic impact by their own lights by spending less time on the Forum, and donating the savings. In this case, if the marginal user-hour costed 6.17 $ to the Forum team[2], I believe they would also increase altruistic impact in the eyes of the user of the marginal hour by spending less time generating engagement on the Forum, and donating the savings to the organisations or projects that user considers the most cost-effective.
Hmm, I think there’s some sense to your calculation (and thus I appreciate you doing+showing this calculation), but the $6.17 conclusion—specifically, “engagement time would drop significantly if users had to pay 6.17 $ per hour they spend on the EA Forum, which suggests the marginal cost-effectiveness of running the EA Forum is negative”—strikes me as incorrect.
What matters is by how much engaging with the Forum raises altruistic impact, which, insofar as this impact can be quantified in dollars, is far, far higher than what one would be willing and able to pay out of one’s own pocket to use the Forum. @NunoSempere once estimated the (altruistic) value of the average EA project to be between 10 and 500 million dollars (see cell C4 of this spreadsheet; here’s the accompanying post). That is far higher than the actual dollar cost of running the average project. (Indeed, if one is funded by EA money, then one’s generation of altruistic dollars needs to outpace one’s consumption of actual dollars—and by a large multiplier, if one is to meet the funding bar.)
Going back to Nuño’s spreadsheet: If I make the arrogant assumption that I’m within an order of magnitude of Ben Todd, impact-wise, then that means my lifetime impact is at least 10 million dollars. Assuming linearity (which isn’t a great assumption, but let’s go with it for now) and a career length of 40 years, this means my impact over the past 4 years has been ≥1 million dollars.[1] In that time, I’ve spent maybe 500 hours on the EA Forum.[2] Meanwhile, I’d say that the Forum has contributed greatly to my intellectual development, i.e., added at least 20% to my impact. (The true percentage may in fact be much higher, because of crucial considerations that the Forum has helped me orient toward, but let’s lowball things at 20%, for now.) This would imply that my impact has been amplified by at least $200,000/(500 hours) = $400 per hour spent on the Forum. ([Insert usual caveats about there being large error bars.]) Contrast with your $6.17.
(I did this calculation on myself not because I’m special, but because I know what the numbers are for myself. I’d guess that the per-hour bottom line for other Forum users would be ~similar.)
We can now go one step further, and estimate the Forum’s “altruistic dollar generated per actual dollar spent” multiplier to be at least 400⁄6.17 ≈ 65. Embarrassingly, I don’t know how this compares against today’s funding bar,[3] but seems very plausible to me that it’s above.
(Nonetheless, people may still not pay $6.17/hour to use the Forum because $6.17/hour is a non-trivial cost considering people’s actual incomes. Additionally, people are just used to being able to browse the internet for free, and so I suspect many wouldn’t do the expected value calculations and reach the “rational” conclusion that they should in fact pay.)
Sanity check: 80,000 Hours says that impactful roles generate millions of dollars worth of altruistic impact per year.
That is, 500 hours consuming the Forum’s content. I’ve also spent time writing on the Forum, but if we model the Forum as a two-way market, with writers and consumers, and say that it’s the consumers who benefit from being here, then it doesn’t make sense to include my writing time. (Also—and perhaps more relevantly—I don’t think writing time gets counted by the Forum’s analytics engine as engagement time if it’s spent mostly in a Google doc.)
Further detail: What really matters is what the multiplier is on the margin (i.e., what it is for the last dollar being spent on a project), rather than what it is for the project as a whole.
Thanks for sharing your thoughts, Will! I agree what matters is the additional altruistic impact caused by engaging with the Forum. However, I think my point holds as long as people have accurate views about how to maximise their altruistic impact.
For example, if you believed “factual impact of your marginal hour on the Forum”—“counterfactual impact of this hour” < “impact of donating 100 $[1] to the organisation or project you consider the most cost-effective”, and using the Forum costed 100 $/h, I think you would have a greater altruistic impact by your own lights by spending less time on the Forum, and donating the savings. Do you agree?
Analogously, if the user spending the marginal hour on the Forum believed “factual impact of their marginal hour on the Forum”—“counterfactual impact of this hour” < “impact of donating 6.17 $ to the organisations or projects they consider the most cost-effective”, and using the Forum costed 6.17 $, I think they would have a greater altruistic impact by their own lights by spending less time on the Forum, and donating the savings. In this case, if the marginal user-hour costed 6.17 $ to the Forum team[2], I believe they would also increase altruistic impact in the eyes of the user of the marginal hour by spending less time generating engagement on the Forum, and donating the savings to the organisations or projects that user considers the most cost-effective.
Implying the marginal cost-effectiveness of your time on the Forum is 25 % (= 400⁄100) of the past cost-effectiveness.
I guess it costs more due to increasing user-hours becoming more difficult as user-hours increase.