I would love to read an approach to effective political giving. Does anyone know of such a document? If the effective altruism movement had focussed on political action and donations instead of private charity to supply bed nets we might have a much better world (including money for bed nets).
Of course, with bed nets and schistosomiasis treatments, you can perform 1% of the intervention and demonstrate 1% of lives saved (although it might take years to see the full benefits from things like better school attendance) which obviously isn't the case for political interventions.
In my mind any evaluation of charities based on 'expected value' has to take into account not only the error bars on how good the charity is, but also the error bars on how good that evaluation is. In other words, how well the evidence stands up to an evaluation by someone who's sceptical about the whole thing. After all, the whole point of effective altruism is being rigorous in evaluating impact.
For an example, imagine a nuclear disarmament charity. Nobody doubts that mutually assured destruction would be very bad! The question is what's the chance of that happening - and estimates of the chance can vary by two orders of magnitude, putting the charity either at the top or the bottom of the impact league tables, depending on who does the evaluation.
In my mind that's a charity whose impact and effectiveness can't be measured, so it doesn't count as EA.
That's why I give my donations to the Schistosomiasis Control Initiative and Against Malaria Foundation.
Any ranking of problems that puts risks to mankind from artificial intelligence as the most pressing problem can't be taken seriously. I realize that it's become trendy in software to think this is actually a threat, but that's more of a marker of an in-group made possible by the collective ignorance of that in-group of the relevant fields of knowledge.
This comment takes two steps away from a good discussion by HN standards. First, you've apparently picked the weakest element of the thing in order to dismiss it. Second, your dismissal is generic. It doesn't teach us anything other than that you think some people are ignorant.
Instead of doing this, please comment on the stronger or more interesting aspects of a thing, and explain your point with real information. Then we all learn something.
Fair enough. I guess I'll engage with this silliness once. Let's start with the basics. There's this vague, handwaving notion of "intelligence." In the community pontificating about this it seems to be operationalized as a combination of manipulating abstractions and extracting new, higher level entities from lower levels of abstraction. For the latter, the only AI work to date that I am aware of on the latter is Drescher's 'Made-up Minds'. Multi-layered neural nets end up with interior points that map to some sort of features, but there is a big jump from that to a system being able to develop an abstraction and then operate in terms of it. For the former, computers are very good at it because humans built them as tools to extend the human ability to do it, just like we built hammers because our hands are kind of a lousy tool to hit things with. There is an illusion of sudden, dramatic progress on this kind of intelligence because curves with lots of parameters (neural nets, SVMs, etc.) can fit all kinds of things if you can get the fit to converge. The advances in machine learning have been tricks to get fits to converge and ways of mapping data into spaces where we know how to get fits to converge. This isn't a denigration of the field. It's given us enormous capabilities that we didn't have before, but we need to be clear about what it is before we start extrapolating: curve fitting. Machine learning today lets us fit curves in order to engineer task specific algorithms. And even that is already going to get strongly restricted over the next few years as legal requirements get piled on to not discriminate against protected classes, provide an appealable argument that can be subjected to a reasonable person standard in a court of law, and a pile of other stuff.
Plus all of this ignores the wide range of forms of intelligence that psychologists have identified. The sociological reason why that's so are interesting in their own right. The stories the practitioners tell about themselves are about brilliant individuals grappling with hard symbolic problems and conquering themselves, and that is the view they project on intelligence rather than a more nuanced, realistic one.
> There was therefore relatively little time for evolutionary pressure to lead to improvements in human intelligence relative to the intelligence of our hominid ancestors, suggesting that the increases in intelligence may be small on some absolute scale. Yet it seems that these increases in intelligence have meant the difference between mammals with a limited impact on the biosphere and a species that has had massive impact.
This makes very strong assumptions about what drove the shaping of the environment by Homo sapiens. Chimpanzees engage in coordinated problem solving behavior. So do dolphins (and they have engaged humans in their solutions). For most of the history of Homo sapiens we operated at levels not very different. So what led to the accumulation of discoveries that culminated in our modern world? The ability to transmit behavior abstractly was a big one. The ability to engage larger social groups by using language for grooming instead of picking lice off. We know that humans have differences in their mental structure that let us retain vastly larger vocabularies than gorillas or chimpanzees, but that's important for transmission of behaviors rather than social grooming. The accumulation of behaviors that led to our modern world is much more akin to natural selection than it is the model of a lone genius thinking up something.
I have yet to see any indication that there is some superhuman level of various kinds of intelligence in the offing or a link between that and major effect on the world.
