Thanks for this link. This made me laugh:

“Everyone Should be a Mapmaker and Fear that Using the Map to Actually Do Something Could Make Them a Worse Mapmaker” would be a much less rousing title, but this is how many EAs and rationalists have chosen to interpret the book.

Thanks for the post @Luke Freeman 🔸 .

Maybe this is a mildly spicy take, but I think as AI risks become more concrete and nearer term, our community should transition a portion of our time from “scouting” to “soldiering”. 

Scout mindset is great for finding new problems and new solutions. But once we have an idea of the problem and ideas for specific solutions, we need to execute the grind to make things happen in the messy real world of people and politics.

This is a great example of how a small bit of “soldiering” by our community helps achieve practical impact. 

If you're interested in this event, but not in Brisbane - we'll be doing similar events around Australia. Sign up to the Good Ancestors' newsletter to receive and email when the details are available, otherwise keep an eye on your local group.

 Newsletter — The Good Ancestors Project 

Thanks for this comment. I couldn't agree more. 

As you say, the overwhelming perspective in Australia up until a few months ago was in favour of acceleration for economic gain and somewhat interested in near-term ethical concerns. It's probably only a marginal overstatement to say that the political environment was actively hostile to AI safety considerations. 

There was a real risk that Australia was going to 'drag the chain' on any international AI safety work. Anyone who has been involved in international negotiations will know how easily one or two countries can damage consensus and slow progress. 

It's obviously hard to calculate the counterfactual, but it seems certain that the work in the community you refer to has helped normalise AI safety considerations in the minds of senior decision-makers, and so there's a real possibility that that work shaped how Australia acted at the Summit. 
 

Thanks so much for the summary Zan.  The letter has attracted a really good spread of AI expertise in Australia, has given us a vehicle to talk to other experts and government advisors less focused on safety issues. The letter is also attracting a reasonable amount of media attention this morning. 

It's hard to overstate how backwards the Australian government's leadership is on AI safety concerns at this point in time. If things continue as they are, it's essentially certain that the Australian government is going to be a skeptical voice in any multilateral negotiations relating to global agreements and standards setting etc. Given Australia's geopolitical position, it would meaningfully harm global efforts if Australia is pulling in the wrong direction.

I'm really hopeful that this effort will have a meaningful impact in Australia correcting course. This is a great start, but it will require sustained effort. 

Thanks Max - I'm glad this is a hot research topic. 

At Good Ancestors Policy, we have begun advocating for the adoption in Australia of various pandemic prevention and mitigation approaches. The residual uncertainty (specifically that we don't have enough evidence to confidently advise on how that risk-benefit calculus should be assessed in different contexts) makes it very difficult currently to advocate for anything specific relating to far-UVC.

My hunch is that government-directed advocacy for far-UVC is only likely to be successful if we can say "this technology has significant benefits during a pandemic, but provides meaningful ongoing benefit from reducing 'colds and flus' even when there isn't a pandemic". That is, if the pitch is instead "install this technology, turn it on if a certain risk threshold is crossed, and the cost-benefit works out because pandemics are super bad" governments might be unlikely to bite even if that cost-benefit assessment is robust.

Will keep following this closely! 

I think the 'good vibes' that help policy advocacy come (in part) from benefiting from other people's positive externalities. That's to say, I'd like us to be in a position where we can say "we're the movement that achieved X, Y and Z. So when we ask your nation to (put 5% of its aid budget to bed nets) you should take us seriously". 

To the extent that we're more centralised and coordinated, it's easier to say "we're the movement that achieved X, Y and Z". When we intentionally fracture and distance ourselves, we also fracture and distance ourselves from those positive externalities.

Again, I recognise that there is a give and take here - risks and opportunities. I just think that we need to put this potential path to policy impact as an 'opportunity' that we largely pass-up when we choose to distance our work and organisations from other another. 
 

I think I mostly agree with what you're saying. Perhaps the difference at the margins is that if 10 organisations get 10 grants for 10 researchers for 6 months, what they'll each do is find a person for a short-term contract. If someone leaves or gets sick etc, a stream of work might not be completed and an organisation will be in a tight spot. 

If that was instead one organisation, maybe it brings on 5-6 researchers for a year (or ongoing), makes some assumptions about staff turnover and part-time arrangements, and when someone leaves or gets sick it's not a big deal because work can be reassigned between the team and project timelines can be shuffled (surge people onto the more urgent work). Basically bigger numbers (dollars; staff; projects) gives managers more wiggle room to find ways to make things work.

But I agree with you, that if funders were hard over on smaller organisations, there are ways they could ease the employment model concerns in other ways. I just think the ideal situation from an employment and management perspective would be longer term and more centralised. 

I'm not sure that's true for two reasons:

First, there's a lot of niche special interest groups that get their way with government. There's lots of ways to pressuring government for a policy outcome that aren't a simple popular appeal.

Second, I don't think it's impossible to build public support for many things we believe in. I think a message like "it's unacceptable that in current year there are people who need something as simple as a net, but don't have access to it - government could fix this today" could easily have popular appeal. Or at least sufficient popular that an alliance of governments would put a few per cent of their aid budgets on the problem and fix it. I agree that not everything we care about could work this way, but many things could. 

A factor in favour of a more coherent EA that this post misses is the importance of policy advocacy. 

I think in almost all of the spheres we care about governments hold most of the levers. For instance, it would be within the power of many countries to unilaterally solve the global insecticide-treated bed net problem if they were sufficiently motivated. I could make this a very long list. 

As a public servant and ministerial adviser, I've been on the receiving end of well-coordinated campaigns by global not-for-profits. They're extraordinal good at getting their way. The example I usually think of is Save the Children. It's hard to know for sure, but StC seems to have fewer people involved than EA and less money. But they have significant access to the leaders of dozens of countries; the ability to drive multilateral agreements through international decision-making bodies (see the UN Declaration on the Rights of the Child); and genuine geostrategic influence.

The EA movement has all the ingredients (global reach; motivated talent; money) necessary to have influence of that kind, or more. But we chose not to (for many of the reasons outlined above), and I think our impact suffers hugely because of that choice. I think we've made a bad deal. I would much rather we paid the price of coordination, managed the risks that result, and used it to be serious players in global policy.