Promising.

to support rigorous safety analysis and clear communication of AI risks.

I believe the AI Futures Project offers excellent research focused on achieving an aligned AI future.

Keep up the great work guys!
Maybe also promote this page a bit more, only found it by chance.

My donation is small, but my support is sincere. To truly raise awareness, I believe the most impactful path is a fictional film or TV show—not a documentary—that powerfully illustrates the risks and human stakes. Storytelling moves people more than facts alone. Wishing you success in spreading the message.

Thanks for AI 2027.

Impressive group devoted to critically important AI safety work

I'm working on building tech infrastructure for making policy research and forecasting more efficient. Would love to have a chat and discuss your use cases. You can contact me at matib275@gmail.com

AI 2027 was great 👍

Big fan of this project! Having this artifact to share and discuss with my friends, family, and colleagues has made it much easier to be concrete about what might be coming with AI advancements.

Looking forward to what the team does next (and to one day try the TTX myself)!

I'm excited for what the team does next!

Personally loved AI 2027, and am exploring ways to contribute positively to the field :)

One thing on my mind:

As a physicist, I've observed that our most innovative solutions often emerge when diverse perspectives converge. I'd love to see more initiatives actively recruiting non-STEM majors into AI safety. Humanities scholars bring unique analytical frameworks and ethical considerations that complement technical approaches, as brilliantly illustrated in Gillian Tett's 'Anthrovision'. Perhaps dedicated fellowships or collaborative projects pairing technical and humanities researchers could catalyse these valuable cross-disciplinary insights?

I'm excited for more thoughtful and informed public discussion of AI futurism.

Approving this proposal. AI Futures Project has been one of the most hyped efforts in my circles for a while now, and for good reason: it brings together some of the most accomplished individuals in the AI safety scene, working under a single banner. For this reason alone, I think the team is worth taking a bet on.

They've also already made their mark, with AI 2027. I'd gotten a sneak peek, and had actually been a bit unimpressed - I had high expectations, but imo the preview suffered from "too many cooks in the kitchen" wrt writing & site design. But by launch, AIFP had upped their game, with a polished product that's been well & widely received. It's certainly shaped how I think about the next few years of AI development. Kudos to the team for being willing to share an early v0, and then iterating to make it better over time!

I think the people working on this are super smart and probably know what they're doing, but I figured I'd throw in my unsolicited 2c:

1. It seems like the core team is already heavy on researchers, so it's unclear to me that hiring more researchers is the right strategic move, vs investing in roles that can produce great content for a wide audience. Right now their plan seems to be to partner with really great folks (eg Scott for writing, Oli for website design, Dwarkesh for podcasts), and it seems to be working so far, but I would guess that having in-house expertise on this could be super valuable, much more so than a marginal researcher.

2. Specifically with the TTX, I haven't played through one myself, but my understanding is it's currently costly to run (requiring an in-person expert facilitator). I'd be pretty excited for ways to automate that, scale it out, and get much wider distribution, eg by shipping an interactive web experience powered by LLMs, or packaging it as a commercial board game.

Anyways, AIFP is one of the most exciting efforts I'm currently tracking; I've made a small personal donation as a show of support. I expect that AIFP will be amply funded by larger parties like OpenPhil and SFF, and as Neel says, is not really in my comparative advantage; but I still think that independent donations are valuable for diversifying funding streams.

I don't believe AI 2027 will be my comparative advantage as a regranter, but I think AI 2027 is great, they're an excellent place to donate money, and I wanted to put my money where my mouth is with a token donation.

I was very impressed with AI 2027. While I don't agree with all of the assumptions, and it's substantially faster than I predict, I found it scarily plausible. It was very productive for concretising my own thoughts and making me more aware of key considerations, for example, whether an AI will try to align its successor. I think this kind of writing is valuable and I would love to see more.

It also seems to have been very well received publicly (much more so than I expected), getting a ton of attention while being much higher quality than the messages that many people may have previously been exposed to. The team has an excellent track record, and I think the success of AI 2027 gives them significant momentum, in addition to Daniel's profile and positive reputation as a whistleblower who tried to walk away from large amounts of money, so it's plausible to me that they can get a good amount of public/policy attention. Though I would be much more optimistic about this if they hired someone with substantial policy/lobbying experience.

While I don't expect to agree with the team on everything, I broadly think they have good judgment, care about the correct things, and will do things that are valuable and helpful. I view this donation and endorsement as a bet on the team and their future judgment, more so than any specific future plan.