@ESYudkowsky
MIRI is the organization that I founded in 2000. In the old days it did most of the deep theory on alignment research, if you filter out the wacky optimistic hopium. MIRI is approximately the only organized group on the planet where you can ask "So what goes wrong if you just try training an AI to do nice things and then make it be superintelligent?", and get back a long list of technical issues that distinguishes deep problems from shallow problems; as opposed to getting back "Nothing will go wrong and it'll work great!" or "Well, you've got to probe the activations to see if it's plotting against you," or "Why, who could possibly predict *that* in advance, we'll just have to try it and see if it goes great!" MIRI is what supported me and Nate Soares through writing the NYT bestselling "If Anyone Builds It, Everyone Dies". MIRI continues to support Nate and others as they run around Washington DC trying to talk to representatives and senators and staff about the importance of the United States not dying -- the way we'd all die if anyone, anywhere on Earth, inside or outside the United States, tried to build something vastly smarter than themselves and of course lost control of it. There are mostly not other people doing exactly this job, because it is not an easy road to funding or political power, compared to flogging "what if China gets it first", or regional problems like datacenters consuming electricity. This means we rely on YOU -- among others, yes, but also YOU - to fund this work continuing. We are not fundraising powerhouses; in 2021 we surprisingly and unusually got a couple of crypto donations totaling $21M, and we used this to focus on doing our jobs and executing some sharp difficult pivots instead of fundraising over the next few years. Our end-of-2025 fundraiser is the first one we've run since then. As with any other 501c3, federal law says that MIRI's status as a public charity depends on us being supported by "the public" rather than single large donors. This means that your small donations to us have an outsized effect *even though* they will in fact be dwarfed by the donations of a few people who can afford to donate seven figures. The rules and the laws are set up to make your smaller donations count. We are trying to fill a matching grant from the Survival and Flourishing Fund which has $0.95M of $1.6M to go, and runs until the end of the year. For perspective, that's less than 0.1% of what Mark Zuckerberg is willing to spend on one researcher who'll help destroy the world. Destroying the world is easy and cheap and pays off big money in the short term -- before of course we all die, in what's increasingly looking like it might also possibly be a fairly short term. We are trying to figure out how to spend *more* money. Even if you look at our full fundraiser rather than the short-term SFF matching grant, it is an embarrassingly small amount of money to be trying to raise while the world is ending and banks with it. Alas, it is not that easy to spend money in useful ways if you demand that it actually be useful; and that has always been the distinguishing feature of MIRI as an organization, among an increasingly crowded field of imitators. So! Please donate an additional $1M before the end of the year. While we do not promise that this will prevent machine superintelligence from eating Earth and its surrounding galaxies, we will try to only do things which make this less likely rather than more likely; and in the meanwhile and regardless, say things that are true rather than false. https://t.co/bh3Ye48Dfr
