I’m working on impact markets – markets to trade nonexcludable goods. (My profile.)
I have a conversation menu and a Calendly for you to pick from!
If you’re also interested in less directly optimific things – such as climbing around and on top of boulders or amateurish musings on psychology – then you may enjoy some of the posts I don’t cross-post from my blog, Impartial Priorities.
Pronouns: Ideally she or they. I also still go by Denis and Telofy in various venues.
Same… Anna Riedl recommended working for something that is at least clearly net positive, a product that solves some important problem like scaling Ethereum or whatever. Emotionally, the exact order of magnitude of the impact probably doesn’t make a proportional difference so that the motivation will be there, and the actual impact can flow from the donations. Haven’t tried it yet, but I will if I go back to ETG.
Yeah, ETG seems really strong to me at the moment! What do you think is a good threshold for the average EA in terms of annual USD donations that they can make at which they should seriously consider ETG?
(E.g., using markets on Manifold. :-D)
AI Safety Events is one of the projects where we expanded the time window because they were on a hiatus in earlier 2023. The events that got evaluated were from 2022. Otherwise yes. (But just to be clear, this is about the retroactive evaluation results mentioned at the bottom of the post.)
First it could make sense not to focus too much on the credits. The ranking has to bottom out somewhere, and that’s where the credits come into it, to establish a track record for our donors. The ranking itself is better thought of as the level of endorsement of a project weighed by the track record of the endorsing donors.
We’re still thinking about how we want to introduce funding goals and thus some approximation of short-term marginal utility. At the moment all projects discount donations at the same rate. Ideally we’d be able to use something like the S-Process to generate marginal utility curves that discount the score “payout” that donors can get. I’ve experimented with funding goals around $100k per project, and 10x sharper discounts afterwards, but it hadn’t made enough of a difference that would’ve legitimized the increased complexity and assumptions. Maybe we’ll revive that feature as a configurable funding goal at some point. But there is also the fundamental problem that we don’t have access to complete lists of donations, so less popular, less well-maintained projects would seemingly have higher marginal utility just because their donation records are more incomplete. That would be an annoying incentive to introduce. Those problems paired with the minor, unconvincing results of my experiments have caused me not to prioritize this yet.
But when it comes to the credits, the instructions to the evaluators are probably a good guide:
“Imagine that you’re given a budget of 1,000 impact credits to allocate across projects (not artifacts). (a) Please allocate them to the projects in proportion to how impactful you think they were. (b) There’s no transaction cost and no change in marginal utility (the first credit is worth the same as the last to a project). (c) We’ll … average your scores, multiply the averages with the number of evaluators, normalize, and then allocate that product to minimize the impact of recusals.”
And further down: “Note that this is a purely retroactive evaluation. (a) You can ignore the tractability of producing an output since they’ve all been produced. (b) Likewise please ignore the cost at which the output was produced. (c) Do consider neglectedness, though, and consider how likely some equivalent output would’ve been produced anyway had it not been for the given project. (d) Consider the ex ante expected utility. A bullshit project mustn’t get a high score because it somehow got unpredictably lucky. (Fictional examples.)”
So like everything in our evaluation, the credits are retroactive too, so they are not about the current margin. One reason to ignore costs is that we don’t have the data, though we might request or estimate it next time around. But the other reason is that the donors to overly expensive projects have already gotten “punish” for their nonoptimal investment through the opportunity cost that they’ve paid. Intuitively it seems to me like it would be double-counting to also reduce the credits that they receive.
Oh true! I was only thinking of financial support for struggling projects and project developers, but those kinds of support are also super valuable!
Rethink Wellbeing is definitely on board with mental health for EAs being an important cause area.
I don’t think personal identity makes too much sense, so preventing the extinction of EA-related values (or maybe even some wider set of prosocial, procivilizational values) could be an underexplored cause area. Some sort of EA crisis fund could be a way to achieve that, but also archival of important insights and such.
Yeah… I’ve been part of another community where a few hundred people were scammed out of some $500+ and left stranded in Nevada. (Well, Las Vegas, but Nevada sounds more dramatic.) Hundreds of other people in the community spang into action within hours, donated and coordinated donation efforts, and helped the others at least get back home.
Only Nonlinear attempted something similar in the EA community. (But I condemn exploitative treatment of employees of course!) Open Phil picked up an AI safety prize contest, and I might miss a few cases. I was very disappointed by how little of this sort happened. Then again I could’ve tried to start such an effort myself. I don’t have the network, so I’m pretty sure I would’ve failed. I was also in bed with Covid for the first month.
I suppose it really makes more sense to model EA not as a community but as a scientific discipline. I have a degree in CS, but I wasn’t disappointed that the CS community didn’t support their own after the FTX collapse because I never had the expectation that that is something that could happen. EA seems to me is better understood within that reference class. (Unfortunately – not because there’s something wrong with scientific disciplines but because I would’ve loved to be part of a real community too.)
Awww, thank you! ☺️
Will do, thanks!
I’ll phrase this as a question to not be off-vibe: Would you like to create accounts with AI Safety Impact Markets so that you’ll receive a regular digest of the latest AI safety projects that are fundraising on our platform?
That would save them time since they don’t have to apply to you separately. If their project descriptions left open any questions you have, you can ask them in the Q & A section. You can also post critiques there, which may be helpful for the project developers and other donors.
Conversely, you can also send any rejected projects our way, especially if you think they’re net-positive but just don’t meet your funding bar.
Friends of mine rented a venue for 1200+ people for a weekend in a central location in Germany for some €60k iirc. There was no catering, but we were allowed to buy food at nearby diners (with great vegan options) and bring it into the venue.
Do the practices around forcing catering companies on organizers maybe vary by countries, so that EAGs could move to nearby countries where the venues are cheaper and more chill? Maybe countries with smooth air and rail connections from (e.g.) London?
There’s also the Zuzalu model. Zuzalu itself may be an option. Or something similar closer to the UK, maybe in France, Germany, Poland, or the Czech Republic. In many cases you’ll be able to just interview the EAGx organizers in those regions to figure out what the costs and customs are around catering. That would increase the cost for travel compensations a bit but probably reduce the other costs much more.
I vaguely remember that Eunkyoung Kwon of the ICNK (at the time at least?) was able to organize a great number of tiny conferences in many countries at no extra cost. Not the right scale for EAGs, but if EAGs should get replaced with single-topic conferences at some point, maybe some should ask her for her secrets.
Yeah, agreed! I haven’t thought about impact markets through Linch’s particular lens. (I’m cofounder of AI Safety Impact Markets.)
Distinguishing different meanings of costly: Impact markets make applying for funding more costly in terms of reputation, in the sense that people might write public critiques of proposals. But they make applying less costly in terms of time, in the sense that you can post one standardized application rather than one bespoke one per funder.
But most people I’ve talked to don’t consider costly in terms of reputation to be a cost at all because they’re eager to get feedback on their plans to improve them, and rejections from funders rarely include feedback. (Critical feedback would then reflect badly on the previous draft but not on the latest one.)
Conversely, I’ve also heard of funders reasoning like “This project falls into the purview of the LTFF, so if it hasn’t gotten funded by them, there’s probably something wrong with it, and I shouldn’t fund it either.” Public feedback like, “We decided not to fund this project because we couldn’t find an expert in the field to assess its merits” would actually be “negatively costly” or beneficial in terms of reputation. It could also help with unwarranted yellow flags because impact markets are all about aggregating and amplifying specialized local knowledge. If, for example, the rumor mill claims that someone is a drug addict, a longterm flatmate could make a symbolic donation and clarify that the person in question only microdoses LSD, no hard drugs. That could silence the incorrect rumor. The flatmate could thus become an early donor to the project and reap an outsized (compared to the size of the donation) reward in terms of their score or later Impact Marks for adding this information to the market.
Our scores will be based on evaluations of the outputs, so all the issues that have to do with lacking rigor or not publishing anything in the first place are priced in. The issue with plagiarism, low integrity, and interpersonal harm is more concerning for me. I’ll consider adding a “whistle-blowing” tab to the comment section where users can post anonymously to deter low-integrity actors from using the platform. We (GoodX) can also manually intervene if we become aware of bad actors.
Generally, my “bias” is to keep things public by default. The funding ecosystem can then make exceptions in cases where that is not possible. The current default seems to be secret by default, which seems unnecessarily costly to me in multiple ways (reapplying to multiple funders in different formats, no feedback, bad coordination between funders, few funding gaps for small donors).
FYI: Our (GoodX’s) project AI Safety Impact Markets is a central place where everyone can publish their funding applications, and AI safety funders can subscribe to them and fund them. We have ~$350k in total donation budget (current updated number) from interested donors.
(If you’re a donor interested in supporting early-stage AI safety projects and you’re interested in this crowdsourced charity evaluator, please sign here.)
Amazing, thank you!
I’m considering crossposting this prize, but is it still funded? If you already received the funding, will you be able to pay out even if it’s clawed back? Thank you!
Thanks! Yeah, I’ve included that in the application form in one or two cases in the hope it’ll save time (well, not only time – I find interview processes super stressful, so if I’m going to get rejected or decline, I’d like (emotionally) for that to happen as early as possible) but I suppose that’s too early. I’ll ask about it later like you do. I haven’t gotten so far yet with any impact-focused org.