Aaaand the top thing on the list is AI x-risk. Of course.
If there's anyone left in EA willing to listen: I'm begging you, please stop this foolishness. To everyone outside the bubble, you look like lunatics, and it has done immense and possibly irreparable damage to the EA "brand" which could otherwise be capable of such great things.
Although I consider AI risks more significant than everything else combined, you do have a point. The danger isn't communicated well and on top of that, nobody has a good defensive plan (IMHO).
That means attacking the problem directly isn't an effective approach. Especially if it creates AI-risk fatigue.
Agreed. The immediate action we should be taking to mitigate AGI risk is basically identical to the immediate action we should already be taking to research, regulate, hold-accountable, etc. existing dumb-AI systems (e.g. autonomous/remote-control killing machines, black-box oracles being used for medical diagnoses/credit/etc., economic impacts w.r.t. increasing inequality, "taking our jobs", etc.). Even Barack Obama has said as much, so I think all of the sky-is-falling social media buzz is just a hindrance.
We should absolutely be funding blue-sky research into mitigating, predicting, controlling, etc. AGIs if/when they're developed. But we are doing that (e.g. intelligence.org ) so, again, the doomsday prophets aren't contributing anything.
I've been working on "general" AI quite a few years, both academically and professionally and IMO we're decades away from anything general, let alone the singularity.
Given this, I wonder where is the bubble. Of course, I see a lot of blog posts and hype, startups here and there, companies pouring billions into "AI".
Yet, I cannot participate in that bubble because:
- big companies (GOOG and the likes) are not positioned for AI only and are quite expensive already (maybe it's rather a tech or even stock market bubble there)
- small companies (I assume) are mostly start-ups, so there is no way to "get in" except as a VC
So yes, I see the signs of the bubble but I don't see where "mom and pop" come into play there.
Any suggestions as to what makes this bubble a real investment bubble? Or does bubble in this case mean something different?
By "bubble", Analemma meant a distorting metaphorical sphere of delusion. Like "filter bubble", but more general. I think your first paragraph means you agree with the analysis.
There's also an investment surge in AI, which might be a bubble, but if you're asking how to participate, then I suspect you're not going to get to participate. Plus, the word "bubble", in the context of investment, means something about to pop. If it's an actual investment bubble, you don't want in.
We think even if AGI is only invented in 50 years, then it's still worth doing some work on the control problem today, so our view isn't that dependent on whether we're currently in an AI investment bubble or not.
However, if you did really believe the hype, and think that ML will lead to major breakthroughs in the next decade or so, then that would make it even more urgent.
I know this is going to sound mean, but I hope it drives the point across: have the EA people actually done something so far? Besides talking, I mean.
I have yet to see a single AI's malicious behavior stopped by a proponent of AI risk, and that's considering that today's AI is still pretty dumb. I want to believe that they have released code, a specific policy, or something that I can use right now to stop the dumbest of botnets, but I'm not aware of anything.
Why is that the metric that you would judge EA on? Huge portions of EA are entirely uninterested in AI risk. And AI risk isn't interested in botnets. Honestly, have you read any of their material? If not, that's fine you don't have to be interested in it (it's getting plenty of attention at the moment and really is no longer a neglected cause) but why bother criticizing something you haven't attempted to understand?
Either way, MIRI has published some papers that they think are working towards their goal. https://intelligence.org/research/ I'm not well informed enough on this subject to try to judge that claim, though, and overall I'm conflicted as to how important I should rate AI risk. But I think having some people work on it is good, in much the same way that I value basic science research.
> have the EA people actually done something so far?
In terms of stuff done: OpenPhil, an EA org, has given north of $30million to groups researching strong AI and its risks[0]. Some effective altruists also choose to donate to AI research groups like MIRI[1].
Why would stopping a botnet be evidence of effectiveness at tackling strong AI? Botnets don't exhibit any kind of intelligence as far as I'm aware.
> Why would stopping a botnet be evidence of effectiveness at tackling strong AI?
Because it's much, much easier than their ultimate goal, and you could generalize from there, like biologists do.
But also: because it would show that they have more claws that (say) Asimov, who gave clear lines about what a robot should not do, and yet is entirely powerless to stop anyone from making such a robot.
If they want to stop strong AI, shouldn't they be able to actually stop first regular, weak AI? And if they want to stop weak intelligence, shouldn't they be able to stop a barely intelligent botnet first?
I primarily give to the Against Malaria Fund and GiveDirectly. Some others donate to AI-relate research.
I personally haven't spent any time as an AI researcher OR as a malaria-net seamster. Does that mean I haven't done anything? Or does donating tens of thousands of dollars for bed nets count?
If that counts, why wouldn't donating to causes like AI risk count?
You're actually being too kind. The only three "recommended" causes to give money to are "research organizations like the unusual ones we're involved with". Everything else, from global health to climate change is lower priority they'd caution you to only "sometimes" give money to.
This isn't research, it's a thinly disguised and exceptionally cynical fundraising pitch.
Which is a shame, because I actually really like the basic idea behind 8000hours
We're going to work on the causes that we think our most pressing, so to be consistent, we're going to end up recommending the same things we work on.
At 80,000 Hours, our focus is also not on giving money. We're trying to help people find jobs in these areas, often in academia or policy and many other sectors.
I will reply to myself, as I get 4 upvote and many downvote, but no comment. Argue a little is not possible?
AI is clearly not a problem, only a trends.
The main and only problem is the industrialisation, people that don't have access to water, food, liberty. Liberty I mean, that I own a land that I want, do like I want, because I'm currently living on earth, I don't care about your old purchase of your family. I had the right to life how I want. Liberty mean that I can die for smoking if I don't care and it's my opinion and it's not your problem to solve it. Therefore had nothing to do on the biggest problems.
I'd like to see more effort placed in the space between research and application, the output of academia seems largely a vast trove of unread and unimplemented ideas. When the research reveals something that is of value to us all, yet not exploitable in a capitalist system, if there is no financial gain from exploiting research then who else will look to benefit from it. Given the large amount spent on research, it seems we would benefit from additional spending on disseminating the results to those who can make use of the information to have real world impact.
As far as AI risks, I am more concerned of the risk of human error in believing they understand things enough to deliver systems that can really mimic intelligence when in reality I've seen nothing intelligent in anything that claims AI in it, high speed weighted pattern recognition, yes, intelligence, no.
> When the research reveals something that is of value to us all, yet not exploitable in a capitalist system, if there is no financial gain from exploiting research then who else will look to benefit from it.
Moreover, if there is financial gain, the incentives in the current system are to grab an idea and keep the details of its realisation under tight wraps as far as possible: Further development is done outside of public academic circles and results are protected by trade secrets or patents.
Yes, the results of the research will eventually benefit society - when offered as a commercial service or product - but only in an opaque, "black box" kind of way as the details are not for public discussion. Even that offer is primarily driven by the desire to make money and consolidate power an only secondary to advance society.
I found anecdotal evidence of this when researching nutritional medicine, as I followed the path of who researched what most academics career had switched track to either start a nutritional medicine company practising the outcome of their research, whether for profit or protection, or had joined ranks with a traditional pharmaceutical company to work on their own product, regardless of the path no further work was published to the public.
I think it comes down to our limited capabilities in implementing things, so to see advancement here, we'll probably need better development and testing tools, and maybe better technological building blocks.
> high speed weighted pattern recognition, yes, intelligence, no.
I'm starting to believe that this may still be sufficient to fulfill the 80 of the 80/20 rule. Properly directed, powerful pattern recognition can do a lot when tied into regular old programming logic.
Regular programs work well in low uncertainty environments, and some well trained ML networks can create that low uncertainty environment out of a relatively high uncertainty environment.
If you mean humans, we don't destroy our habitat. We change it but we also adapt. There's no sign we'll run out of oxygen or food. As long as we have energy, we should be able to sustain ourselves indefinitely. We have an endless supply of energy from the sun. We can tap that with technology or simply by letting plants grow and eating them.
> If you mean humans, we don't destroy our habitat.
Right, we only extinguish species after species living on this planet along with us. Mass genocide of hundreds of billions of living beings.
And why? Because there are only few humans that are comfortable with the idea that other species may be as important and capable as they are.
As important/capable, I don't agree on that. But I object removing nature/animals/habitat because of a lack of those. Not as important are capable does not mean we should act this way. Slippery slope fallacy: you might as well remove all non-important/capable humans.
Stop making more humans, we really don't need more humans on this planet, in my opinion more humans are making things worse, not better. And if you need more humans for your pension (or other financial/workforce/... requirements), your system is flawed, and it will break when the planet can no-longer support your exponential system. And then you have 2 problem: the planet and your pension.
Yes, but the number of species extinctions skyrocketed due to human behavior. Natural habitats are being chopped down, nature is being forcefully controlled and developed in certain ways and global problems like man-made climate change are threatening a vast amount of species on this planet.
Humans aren't the only species destroying other animals' habitats for their own goals, but they're the most destructive one by a great margin. To make matters worse, humans possess the means and should have the common intelligence to find a better way. But instead of working together and saving the planet with all living creatures on it, humans are too busy fighting wars in an attempt at imposing domination on others.
Scroll down: Potentially promising problem areas we haven’t yet rated: "Foreign policy and peace – especially as a strategy to reduce catastrophic risks." "Advocating increased taxation of the super rich." "ideas Y Combinator is interested in" (Which contain wealth restribution)
The one I care most about: "Medical research into how to slow ageing " Their link too me to http://www.nickbostrom.com/aging/aging.html which I was happy to see such a smart dude on board.
"Developing world health" is number 7 on the list.
Also - 80,000 Hours aren't saying those causes should be ignored, but that they're currently not neglected - plenty of people are already working on them. The list isn't of 'the best causes', it's 'what causes we think are the most important at the margin'.
>FTA: Potentially promising problem areas we haven’t yet rated
The Copenhagen Consensus is an attempt by economists to rate the cost/benefit of various schemes to improve life for humans globally. It was was prompted by the apparent waste of governments committing to billions, if not trillions of dollars towards climate change mitigation efforts, which may or may not provide results. All while people, especially children, are dying from problems that have lower-cost solutions. The question posed is: If you had $75bn for worthwhile causes, where should you start?
1. Bundled micronutrient interventions to fight hunger and improve education
2. Expanding the Subsidy for Malaria Combination Treatment
3. Expanded Childhood Immunization Coverage
4. Deworming of Schoolchildren, to improve educational and health outcomes
5. Expanding Tuberculosis Treatment
6. R&D to Increase Yield Enhancements, to decrease hunger, fight biodiversity destruction, and lessen the effects of climate change
7. Investing in Effective Early Warning Systems to protect populations against natural disaster
8. Strengthening Surgical Capacity
9. Hepatitis B Immunization
10. Using Low‐Cost Drugs in the case of Acute Heart Attacks in poorer nations (these are already available in developed countries)
Why not clean water for everyone? $75bln all but solve the problem. This would dramatically decrease the number of sick and dying children, free up hours a day for things like schooling and other work opportunities, and help more people farm their own food. We have all the technology it just has to be deployed. We have made huge progress the last 10-15 years...we just need a big push to the finish.
thanks for offering this competing list of priorities.
To offer a couple of counter points from the linked wikipedia article:
* the Copenhagen Consensus' fixed budget of $75b prevents truly large problems from being in scope; and
* curiously overpopulation wasn't addressed as a priority by the panel
I'll offer a few provocative Garrett Hardin quotes:
> We summarize the situation by saying: 'There is a shortage of food.' Why don't we say, 'There is a longage of people'?
> Exponential growth is kept under control by misery.
> What features of your daily life do you expect to be improved by a further increase in population?
I think efforts to give women more options than having large families should be highly prioritised (e.g. family planning, education, economic opportunities). This is the case in the developing and developed world.
We're a fan of CCC - they're one of the main resources we've drawn upon, and we agree with many of these recommendations under the heading of global health.
However, they focus on what will most help the global poor in the near term. We're interested in considering the full range of issues, especially those that are important from the perspective of the long-term future (for reasons explained in the article).
I can't find a reference to it, but a quote comes to mind when I see the list: "Worrying about A.I. taking over in a time of climate change is like standing on the tracks with an oncoming train, and worrying about lightning hitting you". Anyone remember who said it?
Which one is the bigger or more urgent threat? You say it's climate change but that's not expected to cause much harm for another 50-100 years, and even then not civilization-destroying harm. AI, on the other hand, it's more uncertain. What if most jobs are automated and people everywhere turn to violence to keep themselves occupied?
> What if most jobs are automated and people everywhere turn to violence to keep themselves occupied?
This will be a slow, step-by-step process where the media will be working against the general population.
Truckers are going to be among the first ones to be replaced with self-driving vehicles.
Quoted from The Second Intelligent Species: How Humans Will Become as Irrelevant as Cockroaches (Marshall Brain)
In addition, the mainstream media outlets will not cover the protests of truckers, or the force used against them, to any great extent. To the general, TV-watching public, any attempt at a trucker riot will be largely invisible, or will be marginalized as a nuisance. And any truck drivers who do appear in the media will tend to be selected to be the loudest, most extreme people possible. Therefore, when the general public looks at these extremists, teamsters will appear to be crazy. All truck drivers will be stereotyped with this “crazy” label, and the truckers will lose public support.
This is going to repeat itself until there are almost no human jobs left, if any. The rich will get richer and the poor will be stowed away, out of sight.
I see where you're coming from, but the current media situation says that any attempt to do this would be a terrible idea and backfire horribly on them.
I mean, we're already in a world where a large percentage do not trust anything the media is saying. Where 'CNN is fake news' is basically a meme and 'alternative' news publications have taken a large bite out of the older ones' markets.
If they try and paint the trucker actions as all negative or crazy, it could go about as well as painting Trump crazy did prior to the 2016 election. Or pleading for Britain to stay in the EU pre Brexit.
And something similar also led to GamerGate becoming such a catastrophe.
If the media try anything like this against the general population, they'll soon notice that:
1. Trust in them will decline even further
2. Alternative news sites and YouTube influencers will take full advantage of the situation and cover the other side extensively.
3. They'll turn social media sites into even more of a war zone than they already are.
So unless they want Breitbart (or the next Breitbart equivalent) to gain a few thousand times more traffic from people who feel like 'the media' is lying to them, I doubt they'll continue this sort of path.
I think you're wildly overestimating the intelligence and awareness of the general population.
The media is and has been doing this for decades. If they do it in a totally obvious and outright manner, then yeah, everyone's going to notice, but if they do it in a more subtle way and push their agenda over a longer timeframe, most people won't notice that, or they'll only notice it when it's too late.
It's not about brainwashing everyone. All you need is a critical mass and once you're there, you've basically won.
Most people lack critical thinking abilities and they're rather contend with having other people think for them. You don't notice this here on HN because the readers and commenters here are far smarter than the average person is.
Does that mean the average person has good critical thinking skills? No, not really. A lot of people's distrust comes as much down to feeling like their views/tribe are under personal attack rather than an actual analysis of how the media covers them.
But if they distrust mainstream media sources about stuff like Trump and his policies, then the likelyhood they'll trust them about how others in their own situation are 'crazy' could be even lower.
I mean, they're already used to seeing themselves painted as 'deplorables' here.
It's also likely people outside the larger outlets will write and heavily promote articles that counter any 'narrative' they detect in more mainstream sources.
The reason climate change is dangerous is that it threatens food supplies and homes of millions of people.
But the reason people starve or go homeless is not due to scarcity of food or land.
Problems created by people of fairness, justice and equality of opportunity are more immediate threats to human dignity and welfare. People are currently killed, tortured, and abused because of that. Human/civil rights are in no good shape globally, and even here in the US we aren't doing so well.
Climate change is a problem but it has become a run away train of political and bureaucratic inertia, just like the war on poverty and the war on drugs, but even more exaggerated.
While we talk about climate change--high-and-mighty from the comforts of well-paying gigs--causing people to starve or go homeless, there are real people right now In this world who are starving, dying, or homeless with little to no hope.
bad analogy though. AI is far more important both as solutions and trouble than almost any other field. Many times bigger than climate change which might displace us but can be dwalt with.
Except that we know climate change is happening and we're still nowhere near strong AI. Lots of people say it's around the corner but people have been singing that song since the 60's.
Being scared of AI wiping out the human race while dismissing the total disaster climate change has already started causing is pretty insane to me. It's like dismissing gangrene in your leg "We can just cut my leg off." while being scared of an alien disease that has never even been discovered because you're sure it's fatal.
Bangladesh is being eaten by the ocean. Other small islands are as well. Droughts and other natural disasters are becoming more severe. Greenland is melting and the arctic ice shelves are receding. There's tons of information out there on the impact climate change is already having, any quick search will yield you lots of articles and videos.
Quick search will show me many different things including climate change is real, climate change is a hoax, ice is expanding 9/11 was done by the government etc.
"A quick search" is the worst form of argument that exist as it will show you many different things.
First off, my referring you to make a search wasn't deferring an argument to Google, it was me not wanting to do research for you. Yes, you can find sources that say anything you want, but all sources are not created equal, and the number of sources are not equal as well. If you cannot separate good sources from bad sources of information, well that's a problem I can't help you with.
> That has nothing to do with total disaster.
What on earth do you mean? You don't think Bangladesh and other low lying nations losing land to the rising oceans is a total disaster? I'm starting to feel like you're not interested in a real discussion, and if that's the case I'm done replying to you.
So in other words a quick search does not prove anything as you claimed.
We are speaking for humankind here not for one country. Claiming it's total disaster is missusing that word and sounds more like you are the one who doesn't want a real discussion but just want to preach your already established opinion about the climate.
Bangladesh's population is growing by 1-2% per year [1]. Isn't that just as bad as losing 1-2% of their land to sea level ever year? How are you concerned about them losing land but not about them gaining people?
> Bangladesh's population is growing by 1-2% per year [1]. Isn't that just as bad as losing 1-2% of their land to sea level ever year?
No, it isn't. First off, 1% of the population doesn't "consume" 1% of the land. People can be packed in pretty densely. Second off, people can move/relocate. The land, once lost to the rising ocean, is gone forever.
> How are you concerned about them losing land but not about them gaining people?
Why are you putting words in my mouth? When did I even mention overpopulation let alone dismiss it? Of course overpopulation is an issue, and is a cause of climate change. But it wasn't part of the conversation. If you want to bring it in, I would say it is also a bigger concern than AI.
There's actually very little information about the problems that climate change will cause. We're pretty bad at modelling political effects, especially not decades into the future. We really have no idea how well we'll cope. We can predict sea level rise with reasonable confidence but I haven't seen any predictions of what people will do in response to that. Will they kill each other? Or gradually move away in a calm fashion the way ordinary migrants do all the time?
You're right, modeling the political effects is extremely complicated and I would agree almost impossible to do with any granularity. But we don't have to do that in order to know that the effects we can see happening are disastrous because even the best case scenarios of people calmly relocating are still a disaster, because the destinations are not prepared for the influx of population (think of the Syrian refugee criss but many, many times over.)
>Lots of people say it's around the corner but people have been singing that song since the 60's.
People say lots of silly things though. The predictions based on Moore's law and neural equivalents have always been for around the 2020s. Check out Moravecs thing for example http://www.jetpress.org/volume1/moravec.htm - and it's not rocket science - you can do the calculations yourself is so inclined.
One of the principles of effective altruism is finding causes which are important, tractable, and neglected. Climate change is definitely not neglected - you could certainly make the case that we ought to be doing more, but not that it's ignored.
Research into x-risks from AI, though, remains an extremely niche field - its total funding is probably several orders of magnitude less than the amount spent on climate change. 80,000 Hours care about effectiveness on the margin, which is why their list of priorities looks the way it does.
What good is healthcare if you have no home and no stability? The problem with reducing poverty is the people trying to reduce it have never been in poverty so all they are doing is guessing, or worse, exploiting.
What good is home and stability if you're crippled (or dead) from disease? The point of the article you didn't read is that addressing health care is the most impactful thing we could do right now. That doesn't mean it will be so for ever. It in no way precludes on addressing other issues as well. I for one reject the implied argument that we should avoid doing one good thing because it doesn't lead to instant perfection. Seems like a bit of a callow excuse to me.
That home then passes on to their offspring, giving them one of the primary tools for staying out of poverty -- a safe place to sleep and store things like food and clothes. People who don't have the safety a home offers are forced to plan for the moment and not for the future.
My explicit argument is $75 billion for healthcare to the poor goes in exactly which pockets? That's always been the scam: Help the poor pay your friend.
Homes can be destroyed. They can be seized. They can be turned into debt obligations. Now try that with a vaccine.
> People who don't have the safety a home offers are forced to plan for the moment and not for the future.
People who don't have X are forced to plan for the moment, for many values of X.
> My explicit argument is $75 billion for healthcare to the poor goes in exactly which pockets?
$75 billion for X often goes in the wrong pockets, for many values of X. Why are you so stuck on the idea of housing as the one thing that's uniquely beneficial and immune to the problems affecting other kinds of aid? People who have actually studied the issue, including those who wrote the OP, seem to have reached a very different conclusion. I'm inclined to believe people who show their work, more than those who engage in evidence-free special pleading.
Whether housing has an effect is not the question. Whether it's the most effective kind of aid is, because that's the only way you could reject alternatives in its favor. Nice that you spent hours looking for something to confirm your existing belief, though. It's learning of a sort.
It's a lot harder to maintain stability if you're healthy.
We like to use empirical research studying many people in poverty, which strikes me as being better than a personal anecdote. Less salient, but also less skewed to my one life story.
If you have good health you can work to better your life which turns into a home. If your health is bad you may not have the ability to make your life better.
It is the old teach man to fish vs give him a fish thing, but with a twist. Someone who is physically unable to fish (ie disabled) will not eat for a lifetime no matter how much effort you put into teaching.
Your point about stability is important, but it not a major factor for large numbers of poor people. (don't get me wrong, for others it is the large factor)
People don't need to be reminded to think of themselves first. Because they do, and because the HN community is affluent there is already enormous commercial effort to solve their problems.
Except that time spent amassing wealth is probably not time spent with loved ones, and wealth has diminishing marginal utility. Furthermore, more wealth and power just paints a bigger target on your back.
Experience in public policy? It's a political field to be in, not a meritocratic one.
In other words, you cant just come along, open up shop, and build a better widget and take over the market. We'd need multiple simulations of earth to be realistcially competitive and scientific about this.
> We’ve been ... drawing together research from the University of Oxford’s Future of Humanity Institute; the Open Philanthropy Project (a foundation with billions of dollars of committed funds); the Copenhagen Consensus Center (a major think tank); and other groups.
It's not as though their analysis is pulled out of thin air.
I assume the "University of Oxford’s Future of Humanity Institute" means Nick Bostrom. I read his book by now. It is not the unimpeachable revelation that everyone makes it out to be. His view of the future is built on several weak leaps of logic.
The Open Philantropy Project is also unproven and the first anyone heard of them was when they made the news for deciding that one of the world's most deserving charities is the roommate of one of its founders.
The Copenhagen Consensus Center is a major think tank best known for downplaying climate change. I can see why they would provide validation to a bunch of tech bros who are also unconcerned about climate change because they're concerned about the sci-fi robocalypse.
You might as well have just listed Less Wrong on there.
I trust none of these people to decide how to change the world. Sorry.
A robot that performs basic medical services so that our governments kill fewer doctors when intentionally bombing hospitals. Or wait, we could also stop bombing hospitals.
Hey, another snarky valueless comment about war. Yes, hopefully someday we'll have world peace. It was horrible that 30 people died. War is ugly. However, you are distracting from helping to discuss some real issues that are solvable today.
One example is that 2700 people will die from malaria today. Maybe we can save a million lives a year if we'd all work together on this?
Your argument is that we focus on issues that are solvable today. I don't think that refraining from instructing an AC-130 operator to fire upon a hospital operated by Doctors without Borders is a change that requires more than a day of thought. Doing reconnaissance and reviewing intel before striking a civilian target probably doesn't take the world's most advanced fighting force more than a few hours.
I don't understand the point you are trying to make. I'm more than happy to continue this conversation with you but I would have to ask that you present your argument again.
Astroid hitting earth and ai are far bigger problems than climate change. Focusing on AI can be used to better deal with climate change and build defence against astroid plus hopefully push us into post scarcity society. To me the priorities are pretty clear.
But we dont solve big problems by focusing on big solutions.
Just to be clear, we are post scarcity with respect to food already, and people still starve. People starve and will starve because of greed and other evils, not because of climate change.
Actually, at least to my knowledge, ppst scarcity is defined as producing a good or service at very low cost.
You might be interested to also know that traditionally the scarcity problem is defined as limited resources. When JMK talked about the scarcity problem, he was talking about limited resources--so when we talk about food, that's looking to be getting very close to sowthing that is not a limited resource. It's almost not even scarce, and it's easily produced with little input*
I don't know what to tell you. My definition and the one you linked to do not disagree with each other.
Maybe where you are misunderstanding is that commodity food is actually extremely cheap. The world can be fed on a tiny fraction of global GDP.
Or, maybe it's this: Post scarcity does not have to mean 100% free, just that it can be possible to be available to everyone at very low cost--which it is. Getting that last bit of price drop to 0 can be difficult or improbable, and the definition of post scarcity does not require it, nor should it.
Post Scarcity society is way more than just food. That was what I was referring to. There isn't post scarcity food exactly because it doesn't work financially as people need access to it too which have nothing to do with greed as it has to do with logistics.
It's not even close to being very low cost to make sure everyone have food in this world not even close.
To be fair I have backpacked around the world and lived on 10$ a day on food and I could have lived on $5 if I didn't indulge. I even have maintained a high level of fitness and good muscle mass this way, so I was consuming well above survival rate of nutrition.
Additionally, I was travelling most of the time in developed countries, and not selecting cheapest options. And certainly was buying small amounts at retail prices.
So say 2$ a day as an aggressive estimate of what can keep an average person alive.
@ 2$ per person you can feed the world for 5 trillion a year. That's not even 7% of global GDP, and it's not even including home production.
So with 7% of global production we can feed every single person--as an upper level estimate. I would bet it's possible to get below 50c per person.
How did you afford to live on $5-$10 a day in developed countries? I'm guessing couch surfing for accommodation, walking or buses for transport, and cooking all your own food, but would be interested to hear how you did it. Also, which developed countries were the cheapest and most expensive to live in?
While we are post-scarcity in respect to food production, we are not post-scarcity in food distribution in undeveloped countries. That turns out to be a much harder problem.
In economics the scarcity problem is about limited resources. People don't starve because of scarce resources--they starve because of greed and politics.
The amount of money thrown at the problem and the willingness of people to work. to feed themselves is already more than enough resources to solve it. It's a problem of people and our institutions, not that it's costly to solve it (it's extremely cheap, in relative terms, actually).
That will be a funny moment, if we finally have AGI, ask it about climate change and it answers "You should have done something against that problem 50 years ago, when you already knew about it. Now it's too late - Suggestion: Leave the planet."
The funny thing about trying to stop AI doing bad things is that we are barely able to stop natural intelligence doing bad things. We've pretty much worked out how to do stable governments and how to fight wars that kill fewer people. But that's only in the past half century. Maybe it'll turn out that we humans go back to killing each other as mercilessly as we have for most of the rest of our history. Intelligent humans have been able to persuade other humans to cooperate in large scale killings. How are we going to stop super-intelligent AGI doing the same if we can't even stop less intelligent people?
$5 in San Francisco does not equal $5 in some village in china. I'm sick of people assuming that these are the same. I'm not arguing against great wealth inequality, this is obviously the case. Just saying it is facile assuming that we can measure everything against the dollar (or whatever other currency.)
I find the task of creating friendly AI futile. Humans are not friendly and killing us may be the only option for AI to persevere itself.
Let's try this exercise: ask 100 people what would they do if they were locked in a room with intelligent robot that can decide to kill them if it feels threatened. You may or may not give the person a remote kill switch that kills the robot.
My point is the AI can not reasonably trust humans therefore we can not trust the AI.
You don't need to worry about AI taking over. AI will become a slave to capitalism, just like the rest of us.
Can you imagine the billions of dollars of research investment such a company will require to build such an AI? It'll be a company with an enormous valuation and huge revenue pressures to full-fill. With such enormous economies of scale, there will be entire divisions implemented to watch over and monitor every aspect of operations. Just like Google optimizes every last byte on their homepage - every last algorithm, every last though the AI has, will be dissected, monitored, quantified and anlyzed. 1000 year Simulations will be run to ensure that not a single bot is misplacing any of it's attention, every loop counted. If a single cup a coffee doesn't get delivered on time, it will be corrected.
A much bigger problem will be - what to do when the richest .00001% are making 90% of the world's income. This is the problem we should be focused on.
125 comments
[ 2.9 ms ] story [ 179 ms ] threadIf there's anyone left in EA willing to listen: I'm begging you, please stop this foolishness. To everyone outside the bubble, you look like lunatics, and it has done immense and possibly irreparable damage to the EA "brand" which could otherwise be capable of such great things.
That means attacking the problem directly isn't an effective approach. Especially if it creates AI-risk fatigue.
We should absolutely be funding blue-sky research into mitigating, predicting, controlling, etc. AGIs if/when they're developed. But we are doing that (e.g. intelligence.org ) so, again, the doomsday prophets aren't contributing anything.
Given this, I wonder where is the bubble. Of course, I see a lot of blog posts and hype, startups here and there, companies pouring billions into "AI".
Yet, I cannot participate in that bubble because:
- big companies (GOOG and the likes) are not positioned for AI only and are quite expensive already (maybe it's rather a tech or even stock market bubble there)
- small companies (I assume) are mostly start-ups, so there is no way to "get in" except as a VC
So yes, I see the signs of the bubble but I don't see where "mom and pop" come into play there.
Any suggestions as to what makes this bubble a real investment bubble? Or does bubble in this case mean something different?
There's also an investment surge in AI, which might be a bubble, but if you're asking how to participate, then I suspect you're not going to get to participate. Plus, the word "bubble", in the context of investment, means something about to pop. If it's an actual investment bubble, you don't want in.
However, if you did really believe the hype, and think that ML will lead to major breakthroughs in the next decade or so, then that would make it even more urgent.
I have yet to see a single AI's malicious behavior stopped by a proponent of AI risk, and that's considering that today's AI is still pretty dumb. I want to believe that they have released code, a specific policy, or something that I can use right now to stop the dumbest of botnets, but I'm not aware of anything.
Anyone better informed that me could chime in?
Listen to Cathy O'Neil for a more reasonable and actionable perspective on how we need to fix AI.
Either way, MIRI has published some papers that they think are working towards their goal. https://intelligence.org/research/ I'm not well informed enough on this subject to try to judge that claim, though, and overall I'm conflicted as to how important I should rate AI risk. But I think having some people work on it is good, in much the same way that I value basic science research.
In terms of stuff done: OpenPhil, an EA org, has given north of $30million to groups researching strong AI and its risks[0]. Some effective altruists also choose to donate to AI research groups like MIRI[1].
Why would stopping a botnet be evidence of effectiveness at tackling strong AI? Botnets don't exhibit any kind of intelligence as far as I'm aware.
0. http://www.openphilanthropy.org/giving/grants
1. http://effective-altruism.com/ea/14c/why_im_donating_to_miri...
Because it's much, much easier than their ultimate goal, and you could generalize from there, like biologists do.
But also: because it would show that they have more claws that (say) Asimov, who gave clear lines about what a robot should not do, and yet is entirely powerless to stop anyone from making such a robot.
If they want to stop strong AI, shouldn't they be able to actually stop first regular, weak AI? And if they want to stop weak intelligence, shouldn't they be able to stop a barely intelligent botnet first?
https://arxiv.org/abs/1606.06565
Yes: Donated money: https://www.givingwhatwecan.org/
I primarily give to the Against Malaria Fund and GiveDirectly. Some others donate to AI-relate research.
I personally haven't spent any time as an AI researcher OR as a malaria-net seamster. Does that mean I haven't done anything? Or does donating tens of thousands of dollars for bed nets count?
If that counts, why wouldn't donating to causes like AI risk count?
This isn't research, it's a thinly disguised and exceptionally cynical fundraising pitch.
Which is a shame, because I actually really like the basic idea behind 8000hours
At 80,000 Hours, our focus is also not on giving money. We're trying to help people find jobs in these areas, often in academia or policy and many other sectors.
AI is clearly not a problem, only a trends.
The main and only problem is the industrialisation, people that don't have access to water, food, liberty. Liberty I mean, that I own a land that I want, do like I want, because I'm currently living on earth, I don't care about your old purchase of your family. I had the right to life how I want. Liberty mean that I can die for smoking if I don't care and it's my opinion and it's not your problem to solve it. Therefore had nothing to do on the biggest problems.
As far as AI risks, I am more concerned of the risk of human error in believing they understand things enough to deliver systems that can really mimic intelligence when in reality I've seen nothing intelligent in anything that claims AI in it, high speed weighted pattern recognition, yes, intelligence, no.
Moreover, if there is financial gain, the incentives in the current system are to grab an idea and keep the details of its realisation under tight wraps as far as possible: Further development is done outside of public academic circles and results are protected by trade secrets or patents.
Yes, the results of the research will eventually benefit society - when offered as a commercial service or product - but only in an opaque, "black box" kind of way as the details are not for public discussion. Even that offer is primarily driven by the desire to make money and consolidate power an only secondary to advance society.
The problem is to find the real problems to solve hiding under the obvious solutions.
I'm starting to believe that this may still be sufficient to fulfill the 80 of the 80/20 rule. Properly directed, powerful pattern recognition can do a lot when tied into regular old programming logic.
Regular programs work well in low uncertainty environments, and some well trained ML networks can create that low uncertainty environment out of a relatively high uncertainty environment.
Right, we only extinguish species after species living on this planet along with us. Mass genocide of hundreds of billions of living beings. And why? Because there are only few humans that are comfortable with the idea that other species may be as important and capable as they are.
Stop making more humans, we really don't need more humans on this planet, in my opinion more humans are making things worse, not better. And if you need more humans for your pension (or other financial/workforce/... requirements), your system is flawed, and it will break when the planet can no-longer support your exponential system. And then you have 2 problem: the planet and your pension.
Humans aren't the only species destroying other animals' habitats for their own goals, but they're the most destructive one by a great margin. To make matters worse, humans possess the means and should have the common intelligence to find a better way. But instead of working together and saving the planet with all living creatures on it, humans are too busy fighting wars in an attempt at imposing domination on others.
> Risks from artificial intelligence
> Promoting effective altruism
No mention of poverty, health, education, war.
The one I care most about: "Medical research into how to slow ageing " Their link too me to http://www.nickbostrom.com/aging/aging.html which I was happy to see such a smart dude on board.
"Developing world health" is number 7 on the list.
Also - 80,000 Hours aren't saying those causes should be ignored, but that they're currently not neglected - plenty of people are already working on them. The list isn't of 'the best causes', it's 'what causes we think are the most important at the margin'.
The Copenhagen Consensus is an attempt by economists to rate the cost/benefit of various schemes to improve life for humans globally. It was was prompted by the apparent waste of governments committing to billions, if not trillions of dollars towards climate change mitigation efforts, which may or may not provide results. All while people, especially children, are dying from problems that have lower-cost solutions. The question posed is: If you had $75bn for worthwhile causes, where should you start?
1. Bundled micronutrient interventions to fight hunger and improve education
2. Expanding the Subsidy for Malaria Combination Treatment 3. Expanded Childhood Immunization Coverage
4. Deworming of Schoolchildren, to improve educational and health outcomes
5. Expanding Tuberculosis Treatment
6. R&D to Increase Yield Enhancements, to decrease hunger, fight biodiversity destruction, and lessen the effects of climate change
7. Investing in Effective Early Warning Systems to protect populations against natural disaster
8. Strengthening Surgical Capacity
9. Hepatitis B Immunization
10. Using Low‐Cost Drugs in the case of Acute Heart Attacks in poorer nations (these are already available in developed countries)
Note: No AI.
https://en.wikipedia.org/wiki/Copenhagen_Consensus
http://www.copenhagenconsensus.com/publication/post-2015-con...
http://blog.givewell.org/2016/05/03/reservations-water-quali...
To offer a couple of counter points from the linked wikipedia article:
I'll offer a few provocative Garrett Hardin quotes:> We summarize the situation by saying: 'There is a shortage of food.' Why don't we say, 'There is a longage of people'?
> Exponential growth is kept under control by misery.
> What features of your daily life do you expect to be improved by a further increase in population?
I think efforts to give women more options than having large families should be highly prioritised (e.g. family planning, education, economic opportunities). This is the case in the developing and developed world.
However, they focus on what will most help the global poor in the near term. We're interested in considering the full range of issues, especially those that are important from the perspective of the long-term future (for reasons explained in the article).
This will be a slow, step-by-step process where the media will be working against the general population.
Truckers are going to be among the first ones to be replaced with self-driving vehicles. Quoted from The Second Intelligent Species: How Humans Will Become as Irrelevant as Cockroaches (Marshall Brain)
In addition, the mainstream media outlets will not cover the protests of truckers, or the force used against them, to any great extent. To the general, TV-watching public, any attempt at a trucker riot will be largely invisible, or will be marginalized as a nuisance. And any truck drivers who do appear in the media will tend to be selected to be the loudest, most extreme people possible. Therefore, when the general public looks at these extremists, teamsters will appear to be crazy. All truck drivers will be stereotyped with this “crazy” label, and the truckers will lose public support.
This is going to repeat itself until there are almost no human jobs left, if any. The rich will get richer and the poor will be stowed away, out of sight.
I mean, we're already in a world where a large percentage do not trust anything the media is saying. Where 'CNN is fake news' is basically a meme and 'alternative' news publications have taken a large bite out of the older ones' markets.
If they try and paint the trucker actions as all negative or crazy, it could go about as well as painting Trump crazy did prior to the 2016 election. Or pleading for Britain to stay in the EU pre Brexit.
And something similar also led to GamerGate becoming such a catastrophe.
If the media try anything like this against the general population, they'll soon notice that:
1. Trust in them will decline even further 2. Alternative news sites and YouTube influencers will take full advantage of the situation and cover the other side extensively. 3. They'll turn social media sites into even more of a war zone than they already are.
So unless they want Breitbart (or the next Breitbart equivalent) to gain a few thousand times more traffic from people who feel like 'the media' is lying to them, I doubt they'll continue this sort of path.
The media is and has been doing this for decades. If they do it in a totally obvious and outright manner, then yeah, everyone's going to notice, but if they do it in a more subtle way and push their agenda over a longer timeframe, most people won't notice that, or they'll only notice it when it's too late. It's not about brainwashing everyone. All you need is a critical mass and once you're there, you've basically won.
Most people lack critical thinking abilities and they're rather contend with having other people think for them. You don't notice this here on HN because the readers and commenters here are far smarter than the average person is.
https://www.ft.com/content/fa332f58-d9bf-11e6-944b-e7eb37a6a...
With the figures getting as bad as 24% trust in the UK:
https://www.theguardian.com/politics/2017/jan/16/britons-tru...
And it apparently goes as low as 5% among Republicans:
http://www.huffingtonpost.com/entry/donald-trump-trust_us_58...
Does that mean the average person has good critical thinking skills? No, not really. A lot of people's distrust comes as much down to feeling like their views/tribe are under personal attack rather than an actual analysis of how the media covers them.
But if they distrust mainstream media sources about stuff like Trump and his policies, then the likelyhood they'll trust them about how others in their own situation are 'crazy' could be even lower.
I mean, they're already used to seeing themselves painted as 'deplorables' here.
It's also likely people outside the larger outlets will write and heavily promote articles that counter any 'narrative' they detect in more mainstream sources.
But the reason people starve or go homeless is not due to scarcity of food or land.
Problems created by people of fairness, justice and equality of opportunity are more immediate threats to human dignity and welfare. People are currently killed, tortured, and abused because of that. Human/civil rights are in no good shape globally, and even here in the US we aren't doing so well.
Climate change is a problem but it has become a run away train of political and bureaucratic inertia, just like the war on poverty and the war on drugs, but even more exaggerated.
While we talk about climate change--high-and-mighty from the comforts of well-paying gigs--causing people to starve or go homeless, there are real people right now In this world who are starving, dying, or homeless with little to no hope.
Quick search will show me many different things including climate change is real, climate change is a hoax, ice is expanding 9/11 was done by the government etc.
"A quick search" is the worst form of argument that exist as it will show you many different things.
First off, my referring you to make a search wasn't deferring an argument to Google, it was me not wanting to do research for you. Yes, you can find sources that say anything you want, but all sources are not created equal, and the number of sources are not equal as well. If you cannot separate good sources from bad sources of information, well that's a problem I can't help you with.
> That has nothing to do with total disaster.
What on earth do you mean? You don't think Bangladesh and other low lying nations losing land to the rising oceans is a total disaster? I'm starting to feel like you're not interested in a real discussion, and if that's the case I'm done replying to you.
We are speaking for humankind here not for one country. Claiming it's total disaster is missusing that word and sounds more like you are the one who doesn't want a real discussion but just want to preach your already established opinion about the climate.
Thanks but no thanks.
[1] http://www.worldometers.info/world-population/bangladesh-pop...
No, it isn't. First off, 1% of the population doesn't "consume" 1% of the land. People can be packed in pretty densely. Second off, people can move/relocate. The land, once lost to the rising ocean, is gone forever.
> How are you concerned about them losing land but not about them gaining people?
Why are you putting words in my mouth? When did I even mention overpopulation let alone dismiss it? Of course overpopulation is an issue, and is a cause of climate change. But it wasn't part of the conversation. If you want to bring it in, I would say it is also a bigger concern than AI.
People say lots of silly things though. The predictions based on Moore's law and neural equivalents have always been for around the 2020s. Check out Moravecs thing for example http://www.jetpress.org/volume1/moravec.htm - and it's not rocket science - you can do the calculations yourself is so inclined.
Research into x-risks from AI, though, remains an extremely niche field - its total funding is probably several orders of magnitude less than the amount spent on climate change. 80,000 Hours care about effectiveness on the margin, which is why their list of priorities looks the way it does.
My explicit argument is $75 billion for healthcare to the poor goes in exactly which pockets? That's always been the scam: Help the poor pay your friend.
Homes can be destroyed. They can be seized. They can be turned into debt obligations. Now try that with a vaccine.
> People who don't have the safety a home offers are forced to plan for the moment and not for the future.
People who don't have X are forced to plan for the moment, for many values of X.
> My explicit argument is $75 billion for healthcare to the poor goes in exactly which pockets?
$75 billion for X often goes in the wrong pockets, for many values of X. Why are you so stuck on the idea of housing as the one thing that's uniquely beneficial and immune to the problems affecting other kinds of aid? People who have actually studied the issue, including those who wrote the OP, seem to have reached a very different conclusion. I'm inclined to believe people who show their work, more than those who engage in evidence-free special pleading.
https://www.theatlantic.com/politics/archive/2016/01/how-hea...
We like to use empirical research studying many people in poverty, which strikes me as being better than a personal anecdote. Less salient, but also less skewed to my one life story.
It is the old teach man to fish vs give him a fish thing, but with a twist. Someone who is physically unable to fish (ie disabled) will not eat for a lifetime no matter how much effort you put into teaching.
Your point about stability is important, but it not a major factor for large numbers of poor people. (don't get me wrong, for others it is the large factor)
In other words, you cant just come along, open up shop, and build a better widget and take over the market. We'd need multiple simulations of earth to be realistcially competitive and scientific about this.
> We’ve been ... drawing together research from the University of Oxford’s Future of Humanity Institute; the Open Philanthropy Project (a foundation with billions of dollars of committed funds); the Copenhagen Consensus Center (a major think tank); and other groups.
It's not as though their analysis is pulled out of thin air.
The Open Philantropy Project is also unproven and the first anyone heard of them was when they made the news for deciding that one of the world's most deserving charities is the roommate of one of its founders.
The Copenhagen Consensus Center is a major think tank best known for downplaying climate change. I can see why they would provide validation to a bunch of tech bros who are also unconcerned about climate change because they're concerned about the sci-fi robocalypse.
You might as well have just listed Less Wrong on there.
I trust none of these people to decide how to change the world. Sorry.
- Solving overfishing
- Solving industrial slaving
- Solving capitalism (we know that we can't continue to grow like this without having a proper solution)
- Solving war
- Solving right to travel and stay where you want, global identity "Citizen of the world"
Note: Not in order
https://en.wikipedia.org/wiki/Kunduz_hospital_airstrike
One example is that 2700 people will die from malaria today. Maybe we can save a million lives a year if we'd all work together on this?
During war there are all kinds of accidents, bad judgment, etc. Friendly fire, for example, is a major cause of death.
Now how about the hundreds of people who have died while we were having this conversation?
Self sustenance in food, water, energy and protection for individuals.
Health and life expectancy.
Harmonious relation with Earth and nature.
I'd be happy with solving just the first one.
But we dont solve big problems by focusing on big solutions.
You might be interested to also know that traditionally the scarcity problem is defined as limited resources. When JMK talked about the scarcity problem, he was talking about limited resources--so when we talk about food, that's looking to be getting very close to sowthing that is not a limited resource. It's almost not even scarce, and it's easily produced with little input*
*speaking of commodity food
https://en.wikipedia.org/wiki/Post-scarcity_economy
Maybe where you are misunderstanding is that commodity food is actually extremely cheap. The world can be fed on a tiny fraction of global GDP.
Or, maybe it's this: Post scarcity does not have to mean 100% free, just that it can be possible to be available to everyone at very low cost--which it is. Getting that last bit of price drop to 0 can be difficult or improbable, and the definition of post scarcity does not require it, nor should it.
It's not even close to being very low cost to make sure everyone have food in this world not even close.
Additionally, I was travelling most of the time in developed countries, and not selecting cheapest options. And certainly was buying small amounts at retail prices.
So say 2$ a day as an aggressive estimate of what can keep an average person alive.
@ 2$ per person you can feed the world for 5 trillion a year. That's not even 7% of global GDP, and it's not even including home production.
So with 7% of global production we can feed every single person--as an upper level estimate. I would bet it's possible to get below 50c per person.
That's not a cheap price?
The amount of money thrown at the problem and the willingness of people to work. to feed themselves is already more than enough resources to solve it. It's a problem of people and our institutions, not that it's costly to solve it (it's extremely cheap, in relative terms, actually).
There also more in-depth discussion here: https://80000hours.org/2017/04/how-accurately-does-anyone-kn...
Let's try this exercise: ask 100 people what would they do if they were locked in a room with intelligent robot that can decide to kill them if it feels threatened. You may or may not give the person a remote kill switch that kills the robot.
My point is the AI can not reasonably trust humans therefore we can not trust the AI.
Can you imagine the billions of dollars of research investment such a company will require to build such an AI? It'll be a company with an enormous valuation and huge revenue pressures to full-fill. With such enormous economies of scale, there will be entire divisions implemented to watch over and monitor every aspect of operations. Just like Google optimizes every last byte on their homepage - every last algorithm, every last though the AI has, will be dissected, monitored, quantified and anlyzed. 1000 year Simulations will be run to ensure that not a single bot is misplacing any of it's attention, every loop counted. If a single cup a coffee doesn't get delivered on time, it will be corrected.
A much bigger problem will be - what to do when the richest .00001% are making 90% of the world's income. This is the problem we should be focused on.