Quick flag: "The AI Safety Atlas" is already an established educational resource (the textbook/course used by thousands of students a year, and recommended by 80k career advisors). Reusing the exact name will likely cause confusion in the field. Would you consider a different name? "AI Safety Registry," "AI Safety Directory," or similar? The project itself looks valuable; it's just the naming I'd push on.

Congrats on getting the grant/ those numbers!

I wonder if it makes sense to translate those clips into a few languages automatically?

I personally think that this project deserved more. Those numbers are quite impressive, and the Transformer newsletter is a critical epistemic infrastructure for the community.

I think the numbers highlighted are pretty impressive (~$570 per GWWC pledge and ~$1700 per MARS fellow), and I think marketing is important and would benefit many organisations.

From what I see, GWWC estimates the average 10% pledger donates $100,000 over the course of their pledge, and they secured 2 pledges from the EA Netherlands campaign alone—so this is already worth it for this. I would give them money.

Hi, interesting proposal. I've downvoted because I think this awareness strategy is dominated by alternatives like producing short YouTube/TikTok videos explaining the same concepts, which could reach comparable audiences at a fraction of the cost. Here is an analysis of our YouTube strategy for example:  https://manifund.org/projects/scaling-ai-safety-awareness-via-content-creators

Congrats on getting the grant!

I'd be curious which papers or agenda you see as closest to your theoretical framework/broader theory of impact. The project summary mentions preliminary results at AAAI, NeurIPS, CAV—could you link to those?

Congrats on getting the grant. Progress is cool. We need more progress.

I'm sure you have already heard of this, but AI will be the source of most of the progress over the coming years, so curious how you'll approach AI's role in progress, if at all.

Also, it might be interesting to test a strategy for 15 days that involves reaching out to and collaborating with big YouTubers to drive growth. This is what CeSIA did, and we got multiple videos with more than 1M views like this without even being YouTubers ourselves (we just acted as consultants). See a write-up of the strategy here: https://manifund.org/projects/scaling-ai-safety-awareness-via-content-creators

In our first SF batch, we worked with companies like Andon Labs, Lucid Computing, and Workshop Labs, --> What was the counterfactual impact?

Apart has been useful for me to quickly experiment with ideas/improve on quick iteration. I've organised multiple hackathons before knowing apart, and their format is vastly more effective at converting talent/by unit of effort. While I was head of EffiSciences’ AI Safety Unit, this was one of my favorite event formats, and this is one of the format that I encourage alumni of ML4Good to run. Empirically, each apart hackathon that I organized in Paris enabled the long term careers of 0.6 person in AI Safety (see the table). This means that, on average, 0.6 new full-time persons started working on AI safety after each Apart hackathon event in Paris.

We posted an update here: https://manifund.org/projects/ai-safety-atlas

@Markov And the book has a much better name now: "The AI Safety Atlas" !