My impression is that CEA's goal is to fund the meta cause area and the main goal of local groups is to organize events. While funding is hard to democratize unless you convince some billionaire, democratizing the organizations that run events is trivial. [Edit: Also, while it makes sense to organize local events directly based on the local community's preferences / demand, I think it makes sense to take a more top-down (principles-oriented) approach when it comes to distributing funding, because the "demand-side" here comprises of every person on the planet who appreciates money.]

But now I do realize that in my head, I equated CEA with OpenPhil's wing for the meta cause area, which might not be accurate. I also feel good about democratizing CEA if I imagine it implemented as an indirect democracy (i.e. with local organizations voting, instead of every EA member). This probably moves me towards the middle of the poll - i.e. I would be in favor of this kind of democracy. Indirect democracy would reduce the problem of uninformed voters, the problem of dealing with problems publicly and the problem of disbalance in the level of reflection between the average member and highly-engaged members.

Thank you! Democratizing local groups sounds clearly good to me and I assumed it was the norm but I didn't find any data on that.

I can see two realistic models for the parallel organization, which I'm not a fan of:

1) A competitor to CEA. Just like CEA, this org would mainly fundraise and fund projects.

I think the problems with selecting members mentioned in this thread are overstated. Any political party faces the same problem. I suspect that in practice, strategically recruiting weakly engaged EAs just isn't a big problem. But it could be either mitigated by requiring members to meet any of the conditions you mentioned (fees, EA org employment, course certificate), or setting a number of votes per regions, e.g. based on similar indicators of the # of engaged members.

Personally, I'm sufficiently satisfied with the general CEA agenda, that I suspect this would be a waste of effort. That's in part because I think highly engaged EAs who dominate these orgs have more philosophically robust views and in part because I don't think this competitor organization would be able to raise more than 10 % of CEA's budget (~80 % of it comes from OpenPhil). So, given the main goal of funding projects, I don't think this org would be sufficiently better to be worth all the costs - and not just costs inherent in the operations, but also the emotional costs of having these debates publicly and the costs of coordinating "who is willing to fund what" which I imagine might already be a nightmare.

2) A union. A soft counter-power to CEA.

If this org's only power were the possibility to strike or produce resolutions, I'm concerned this would artificially inflate unproductive discord. My impression is that unions often produce irrational policies perhaps because they only have quite extreme measures at their disposal, which creates an illusory "us vs. them" aesthetics for relationships that are overall very positive-sum.

However, I have some sympathy for the idea of

3) A community ambassador who would be democratically voted e.g. by all EA Forum members and who's job would be to facilitate the communication between CEA and the community in both directions. I imagine someone at CEA might already effectively hold this job, so perhaps they would be interested in having their choice ratified by the community. Ideally, this community ambassador would collect people's concerns and visit CEA board meetings, in order to be able to integrate both perspectives.

However, I think the cost of this position is non-negligible. Given the power-law distribution of impact among people and given the many rounds of tests, which employees at EA organizations allegedly undergo - a democratic vote would probably yield a much less discerning choice (as most people wouldn't spend more than 30 minutes picking a candidate). I'm not sure to what extent the wisdom of the crowd might apply here.

Because of similar uncertainties and because I wouldn't count this as a "leadership role", I'm voting "moderately disagree".

I think the comparison in energy consumption is misleading because phones use unintuively little energy, as much as 10 Google searches per one charging, (Andy Masley has good articles on AI emissions), using a smartphone for one year costs less than a dollar. I think a good heuristic is "if it's free, it uses so little energy that it's not worth considering".

If you're not paying to generate it, you're also not taking any income away from artists.

The argument that it's bad vibes for artists is a good one.

I don't want to pick any position here but I think one problem of ex-post moral judgements is that the failures themselves can be valuable lessons. Generally, governments seem to underrate the value of experiments - imagine if welfare interventions or tax policies utilized an RCT approach.

Natural experiments that showed any culture is compatible with freedom (e.g. South vs. North Korea; mainland China vs. Taiwan), that healthcare & education can be cheaper even with a government-run system (USA vs. EU) or that central planning seems worse for welfare (West vs. East Germany) - seem like a really important driver of progress.

Newsom's press release and veto message include much more detail and suggest "it's too weak" is not the actual reason.

Reasons mentioned:

1. Discrimination by model size
   1. "SB 1047 only applies to large models, giving us a "false sense of security about controlling this fast-moving technology. Smaller, specialized models may emerge as equally or even more dangerous"
2. "Real risks" are limited to critical decision-making, critical infrastructure etc.
   1. "While well-intentioned, SB 1047 does not take into account whether an Al system is deployed in high-risk environments, involves critical decision-making or the use of sensitive data. Instead, the bill applies stringent standards to even the most basic functions - so long as a large system deploys it. I do not believe this is the best approach to protecting the public from real threats posed by the technology."
   2. Newsom wants to focus on "specific, known" "demonstrable risks to public safety" "rooted in science and fact", like the deepfake laws he signed.

Organizing good EAGx meetups

EAGx conferences often feature meetups for subgroups with a shared interest / identity, such as "animal rights", "academia" or "women". Very easy to set up - yet some of the best events. Four forms I've seen are

a) speed-friending

b) brainstorming topics & discussing them in groups

c) red-teaming projects

d) just a big pile of people talking

If you want to maximize the amount of information transferred, form a) seems optimal purely because 50% of people are talking at any point in time in a personalized fashion. If you want to add some choice, you can start by letting people group themselves / order themselves on some spectrum. Presenting this as "human cluster-analysis" might also make it into a nerdy icebreaker. Works great with 7 minute rounds, at the end of which you're only nudged, rather than required, to shift partners.

I loved form c) for AI safety projects at EAGx Berlin. Format: A few people introduce their projects to everyone, then grab a table and present them in more detail to smaller groups. This form might in general be used to allow interesting people to hold small low-effort interactive lectures & utilizing interested people as focus groups.

Form b) seems to be most common for interest-based meetups. It usually includes 1) group brainstorming of topics 2) voting on the topics 3) splitting up 4) presentations. This makes up for a good low-effort event that's somewhere between a lecture and a 1-on-1 in terms of required energy. However, I see 4 common problems with this format: Firstly, steps 1) and 2) take a lot of time and create unnaturally clustered topics (as brainstorming creates topics "token-by-token", rather than holistically). Secondly, in ad hoc groups with >5 members, it's hard to coordinate who takes the word and in turn, conversations can turn into sequences of separate inputs, i.e. members build less upon themselves. Thirdly, spontaneous conversations are hard to compress into useful takeaways that can be presented on the whole group's behalf.

Therefore, a better way of facilitating form b) may be:

Step 0 - before the event, come up with a natural way to divide the topic into a few clusters.

Step 1 - introduce these clusters, perhaps let attendees develop the sub-topics. Their number should divide the group into subgroups of 3-6 people.

Step 2 - every 15 minutes, offer attendees to change a group

Step 3 - 5 minutes before the end, prompt attendees to exchange contact info

Step 4 - the end.

(I haven't properly tried out this format yet.)

I like the argumentation for possibility & importance. My only nit-pick would be how bad would a realistic bad case scenario actually look like. Hungary seems like a good model - you could get some anti-liberal legislation, more gerrymandering, maybe some politicized audits of media and universities - however, the government is still selected based on the number of votes (Freedom House) and it's not a stereotypical "fall of democracy" accompanied by a collapse of economy that would destroy most EA efforts.

I occasionally ruminate two projects in this area (for 2028):

1) Funding a mock-election with ranked-choice voting. (Now I see wes R proposes something similar). To legitimize it, it would have to 

a) have robust identity checks

b) have a large demographically representative sample

c) be accompanied by a campaign informing people that a consensus candidate X would win if enough people were honest in surveys, cross-voted, cross-registered or switched to a new party.

2) Policies / financial incentives to make the army more representative of the US population.