Even when people could afford these costs (for instance by funging against donations), asking for money serves another useful purpose: it serves as a sanity check on whether this is a reasonable use of resources.

For people who can afford it and are confident the expense is justified, it makes sense for them just to pay. Having a system they can apply to makes sense both for those who cannot afford it, and those who want a hard-nosed outside opinion.

Here's how I replied:

I think the reasoning is that people are irrationally offput from spending small amounts of money, and in particular offput from getting VAs to finish projects. In addition, there's high and offputting overhead for first getting set up with VAs, which we've had to handle anyway.

Of course, people could afford to pay these small sums themselves, and I'm very sensitive to concerns about donating to the global rich rather than GiveDirectly. Indeed I think a lot of spending in the EA movement broadly conceived is misplaced on these grounds, ultimately gets focused on helping those globally rich people, and does less good than GiveDirectly. But I think this particular .impact spending, which is after all especially small, is worth trying.

Reductio ad absurdum: Why we should give any money to organizations like Givewell, that then turn around and pay that money to their globally rich employees who are doing EA research work?

Paying Givewell employees a solid salary lets them focus on EA stuff full-time while not having to expend a lot of stress & energy making sure their needs get met. It's effectiveness that matters, not overhead. Providing virtual assistants to independent EAs seems like it could do much the same thing--giving them a free virtual assistant exoskeleton lets them spend more of their time on nonreplaceable EA work and less on menial, outsourceable tasks that arise during the course of their EA work.

The counterargument would be that many EAs have full-time earning-to-give jobs, so they could just deduct the funding necessary for their independent EA projects from whatever their normal charitable donations would be. So I think funds like this make sense for EAs who don't have full-time earning-to-give jobs: students who want to do effective activism (say for pandemic risks?) in their spare time, someone who quits their job and lives off savings in order to research EA-related topics full-time, someone who thinks earning to give is oversubscribed and is focusing on an EA project instead of advancing their career, etc.