Let's do this!

Hi Everyone!

We are now a bit more than six months in and wanted to update:

We managed to get a very pro-EA person to update the texts (copy) on our website to increase SEO for us for the symbolic sum of roughly 100 USD in Swedish currency. This leaves us with still having the large part of the grant left, however it seems too little to get a UX designer to create a visually and in other ways user compelling website for us. We are still looking for someone experienced in UX to upgrade the site fully without the risk of having parts of it badly made just because we run out of funds, so so far we haven't used the large part of the funding, but we are still on the lookout.

I think AI Safety Atlas is awesome! I highly encourage further donations.

@Maria thanks for asking -- if you add money to your balance, that would count as a charitable donation to Manifund (a US-based 501c3 charity). You can then spend that charity balance towards your choice of projects on our site; Manifund vets each project to make sure it is compliant with our charitable mission.

@manuelallgaier apologies, I've now deleted the other comment!

I ❤️ MATS! So many brilliant and motivated researchers.

Really impressed by what CEEALAR does on such a limited budget. It was a source of inspiration for Cavendish Labs. I also appreciate your transparency and detailed budget.

Saw a recording of the preview and it was pretty epic! Hope to see you live in the Bay :)

wishing you many sleepless nights!

I am new to Manifund and am confused by the platform. Does it make any difference to load cash or charity balance before donating (i.e, would this be a charitable donation)? Asking publicly since maybe someone else also wonders.

That aside, I’m very happy to support the musical production. I had the chance to see part of the musical in Berlin, it was superb: thought-provoking, creative, and extremely well executed.

@manuelallgaier This seems to have been posted twice (technical issue), and I can't seem to delete it? Can anyone here delete my comment, or let me know how I can do it myself?

I've seen parts of the musical and heard many people raving about it, and I'm impressed what you've put together already in personal capacity with only €3k ($3.3k) only.

I think this fills an important gap: We need more stories to engage people emotionally as well. Rational arguments alone won't do, for most people, but we barely have any stories yet, especially not in musical or video format. I hope this becomes really big, and maybe at some point Netflix picks up on it and creates a large budget series.

If you care to elaborate, I'd be curious to hear whether you've thought about ticket pricing for this. If you don't receive the full amount by donations only, I could see this working with investments as well, e.g. people donate now to cover the financial risk but get their donations back if ticket sales cover the costs. Tickets could have a low income tier, or even be priced "pay as you feel".

Just sharing this as idea, there are probably considerations I missed and I totally understand if you want to do this differently.

What progress have you made since your last update?

Extensive research on anti-rebate laws and determined states with less compliance barriers to operation. Revised Commissions for a Cause website .

What are your next steps?

Finalize and execute compliance plan. Begin larger ad spend in new states.

Is there anything others could help you with?

If anyone knows anyone in the states looking to buy life insurance, could make referral. Could use help with marketing.

I co-lead the Turing Seminar while the Atlas was still being developped at ENS Paris. It offers a comprehensive survey of different fields of AIS and is clearly written and coherent. The facilitation guide did most of the heavy-lifting for me, by offering intriguing and entertaining questions to the class. Charbel, who I was in contact with, actively seeks feedback from the students and develops the textbook responsively. The fact that it's online and maintained by dedicated authors allows for rapid updates, which is a big plus.

What progress have you made since your last update?

Invisible Animals Campaign: 

We have gathered nearly 35,000 signatures, demonstrating strong public support for these neglected animals.

What are your next steps?

We are now preparing to bring our demands to parliamentarians at both city and national levels. Even partial adoption of these proposals could significantly improve the lives of thousands of animals.

Is there anything others could help you with?

This campaign was launched with limited funding, yet its potential impact is substantial. Any contribution to help backfill the funding gap would be greatly appreciated and would directly support advocacy efforts for these overlooked animals.

I read part of the Atlas (Evaluation Chapter). As an AI safety research engineer, I find it a pretty good in-depth introduction. It covers a lot in a very clear way.

What progress have you made since your last update?

• ran a month-long 30-person pop-up campus for projects relevant to my topic: Portal

• there I started the community archive, a crowdsourced open database and api of 16M tweets from 240 high quality and topic-relevant accounts

• I researched the foundations of wisdom and wrote about my research.

• I built multiple collective intelligence tools, like this keyword trends app, and an app that shows you a Birdseye view of your tweets with clustering and LLM analysis.

• Finally, I ran a 40-person community archive hackathon in NYC, with 16 submissions and raised prize money.
  

What are your next steps?

• Building tools to map how ideas spread. This can be very useful to defend against psyops and misinformation campaigns, which become bigger risks as AI improves to do them automatically.
  

Is there anything others could help you with?

I could use money to hire my best collaborators, although i would probably start another proposal here

I am familiar with Tyler's work and encouraged him to apply. He did great work at Longview and is currently independent and advising some other philanthropists. I think this is very high impact work and want him to be supported to focus on it as long as seems impactful. Some of the details are a bit sensitive which is why the write-up is fairly sparse. I'm giving 20K now, I'll hopefully increase this if I get a budget top-up.

Thank you @Austin! I would be interested to hear how creating an Act I like space with Mox goes for you -- I think the design space is wide and under explored.

Approving this project as compatible with our charitable mission of furthering public scientific research! Tsvi has a track record within the rationalist community, and this agenda seems intriguing; I hope it goes well.

I don't know Scott at all, but looking through his website, he's previously received OpenPhil funding and won an ELK prize: external credentials that his research may be worthwhile. I also like the way he thinks, both here and on the Act I project -- it seems good to me to treat AIs with the consideration one might show a child or friend. (Indeed, one of my longer term visions for Mox includes something Act I-y, by having models be participants in the space alongside humans). For these reasons, I'm happy to support Scott with a small initial grant!

Let's do this!

Hi Everyone!

We are now a bit more than six months in and wanted to update:

We managed to get a very pro-EA person to update the texts (copy) on our website to increase SEO for us for the symbolic sum of roughly 100 USD in Swedish currency. This leaves us with still having the large part of the grant left, however it seems too little to get a UX designer to create a visually and in other ways user compelling website for us. We are still looking for someone experienced in UX to upgrade the site fully without the risk of having parts of it badly made just because we run out of funds, so so far we haven't used the large part of the funding, but we are still on the lookout.

I think AI Safety Atlas is awesome! I highly encourage further donations.

@Maria thanks for asking -- if you add money to your balance, that would count as a charitable donation to Manifund (a US-based 501c3 charity). You can then spend that charity balance towards your choice of projects on our site; Manifund vets each project to make sure it is compliant with our charitable mission.

@manuelallgaier apologies, I've now deleted the other comment!

I ❤️ MATS! So many brilliant and motivated researchers.

Really impressed by what CEEALAR does on such a limited budget. It was a source of inspiration for Cavendish Labs. I also appreciate your transparency and detailed budget.

Saw a recording of the preview and it was pretty epic! Hope to see you live in the Bay :)

wishing you many sleepless nights!

I am new to Manifund and am confused by the platform. Does it make any difference to load cash or charity balance before donating (i.e, would this be a charitable donation)? Asking publicly since maybe someone else also wonders.

That aside, I’m very happy to support the musical production. I had the chance to see part of the musical in Berlin, it was superb: thought-provoking, creative, and extremely well executed.

@manuelallgaier This seems to have been posted twice (technical issue), and I can't seem to delete it? Can anyone here delete my comment, or let me know how I can do it myself?

I've seen parts of the musical and heard many people raving about it, and I'm impressed what you've put together already in personal capacity with only €3k ($3.3k) only.

I think this fills an important gap: We need more stories to engage people emotionally as well. Rational arguments alone won't do, for most people, but we barely have any stories yet, especially not in musical or video format. I hope this becomes really big, and maybe at some point Netflix picks up on it and creates a large budget series.

If you care to elaborate, I'd be curious to hear whether you've thought about ticket pricing for this. If you don't receive the full amount by donations only, I could see this working with investments as well, e.g. people donate now to cover the financial risk but get their donations back if ticket sales cover the costs. Tickets could have a low income tier, or even be priced "pay as you feel".

Just sharing this as idea, there are probably considerations I missed and I totally understand if you want to do this differently.

What progress have you made since your last update?

Extensive research on anti-rebate laws and determined states with less compliance barriers to operation. Revised Commissions for a Cause website .

What are your next steps?

Finalize and execute compliance plan. Begin larger ad spend in new states.

Is there anything others could help you with?

If anyone knows anyone in the states looking to buy life insurance, could make referral. Could use help with marketing.

I co-lead the Turing Seminar while the Atlas was still being developped at ENS Paris. It offers a comprehensive survey of different fields of AIS and is clearly written and coherent. The facilitation guide did most of the heavy-lifting for me, by offering intriguing and entertaining questions to the class. Charbel, who I was in contact with, actively seeks feedback from the students and develops the textbook responsively. The fact that it's online and maintained by dedicated authors allows for rapid updates, which is a big plus.

What progress have you made since your last update?

Invisible Animals Campaign: 

We have gathered nearly 35,000 signatures, demonstrating strong public support for these neglected animals.

What are your next steps?

We are now preparing to bring our demands to parliamentarians at both city and national levels. Even partial adoption of these proposals could significantly improve the lives of thousands of animals.

Is there anything others could help you with?

This campaign was launched with limited funding, yet its potential impact is substantial. Any contribution to help backfill the funding gap would be greatly appreciated and would directly support advocacy efforts for these overlooked animals.

I read part of the Atlas (Evaluation Chapter). As an AI safety research engineer, I find it a pretty good in-depth introduction. It covers a lot in a very clear way.

What progress have you made since your last update?

• ran a month-long 30-person pop-up campus for projects relevant to my topic: Portal

• there I started the community archive, a crowdsourced open database and api of 16M tweets from 240 high quality and topic-relevant accounts

• I researched the foundations of wisdom and wrote about my research.

• I built multiple collective intelligence tools, like this keyword trends app, and an app that shows you a Birdseye view of your tweets with clustering and LLM analysis.

• Finally, I ran a 40-person community archive hackathon in NYC, with 16 submissions and raised prize money.
  

What are your next steps?

• Building tools to map how ideas spread. This can be very useful to defend against psyops and misinformation campaigns, which become bigger risks as AI improves to do them automatically.
  

Is there anything others could help you with?

I could use money to hire my best collaborators, although i would probably start another proposal here

I am familiar with Tyler's work and encouraged him to apply. He did great work at Longview and is currently independent and advising some other philanthropists. I think this is very high impact work and want him to be supported to focus on it as long as seems impactful. Some of the details are a bit sensitive which is why the write-up is fairly sparse. I'm giving 20K now, I'll hopefully increase this if I get a budget top-up.

Thank you @Austin! I would be interested to hear how creating an Act I like space with Mox goes for you -- I think the design space is wide and under explored.

Approving this project as compatible with our charitable mission of furthering public scientific research! Tsvi has a track record within the rationalist community, and this agenda seems intriguing; I hope it goes well.

I don't know Scott at all, but looking through his website, he's previously received OpenPhil funding and won an ELK prize: external credentials that his research may be worthwhile. I also like the way he thinks, both here and on the Act I project -- it seems good to me to treat AIs with the consideration one might show a child or friend. (Indeed, one of my longer term visions for Mox includes something Act I-y, by having models be participants in the space alongside humans). For these reasons, I'm happy to support Scott with a small initial grant!