nhaliday + automation + war + arms   5

The Hanson-Yudkowsky AI-Foom Debate - Machine Intelligence Research Institute
How Deviant Recent AI Progress Lumpiness?: http://www.overcomingbias.com/2018/03/how-deviant-recent-ai-progress-lumpiness.html
I seem to disagree with most people working on artificial intelligence (AI) risk. While with them I expect rapid change once AI is powerful enough to replace most all human workers, I expect this change to be spread across the world, not concentrated in one main localized AI system. The efforts of AI risk folks to design AI systems whose values won’t drift might stop global AI value drift if there is just one main AI system. But doing so in a world of many AI systems at similar abilities levels requires strong global governance of AI systems, which is a tall order anytime soon. Their continued focus on preventing single system drift suggests that they expect a single main AI system.

The main reason that I understand to expect relatively local AI progress is if AI progress is unusually lumpy, i.e., arriving in unusually fewer larger packages rather than in the usual many smaller packages. If one AI team finds a big lump, it might jump way ahead of the other teams.

However, we have a vast literature on the lumpiness of research and innovation more generally, which clearly says that usually most of the value in innovation is found in many small innovations. We have also so far seen this in computer science (CS) and AI. Even if there have been historical examples where much value was found in particular big innovations, such as nuclear weapons or the origin of humans.

Apparently many people associated with AI risk, including the star machine learning (ML) researchers that they often idolize, find it intuitively plausible that AI and ML progress is exceptionally lumpy. Such researchers often say, “My project is ‘huge’, and will soon do it all!” A decade ago my ex-co-blogger Eliezer Yudkowsky and I argued here on this blog about our differing estimates of AI progress lumpiness. He recently offered Alpha Go Zero as evidence of AI lumpiness:

...

In this post, let me give another example (beyond two big lumps in a row) of what could change my mind. I offer a clear observable indicator, for which data should have available now: deviant citation lumpiness in recent ML research. One standard measure of research impact is citations; bigger lumpier developments gain more citations that smaller ones. And it turns out that the lumpiness of citations is remarkably constant across research fields! See this March 3 paper in Science:

I Still Don’t Get Foom: http://www.overcomingbias.com/2014/07/30855.html
All of which makes it look like I’m the one with the problem; everyone else gets it. Even so, I’m gonna try to explain my problem again, in the hope that someone can explain where I’m going wrong. Here goes.

“Intelligence” just means an ability to do mental/calculation tasks, averaged over many tasks. I’ve always found it plausible that machines will continue to do more kinds of mental tasks better, and eventually be better at pretty much all of them. But what I’ve found it hard to accept is a “local explosion.” This is where a single machine, built by a single project using only a tiny fraction of world resources, goes in a short time (e.g., weeks) from being so weak that it is usually beat by a single human with the usual tools, to so powerful that it easily takes over the entire world. Yes, smarter machines may greatly increase overall economic growth rates, and yes such growth may be uneven. But this degree of unevenness seems implausibly extreme. Let me explain.

If we count by economic value, humans now do most of the mental tasks worth doing. Evolution has given us a brain chock-full of useful well-honed modules. And the fact that most mental tasks require the use of many modules is enough to explain why some of us are smarter than others. (There’d be a common “g” factor in task performance even with independent module variation.) Our modules aren’t that different from those of other primates, but because ours are different enough to allow lots of cultural transmission of innovation, we’ve out-competed other primates handily.

We’ve had computers for over seventy years, and have slowly build up libraries of software modules for them. Like brains, computers do mental tasks by combining modules. An important mental task is software innovation: improving these modules, adding new ones, and finding new ways to combine them. Ideas for new modules are sometimes inspired by the modules we see in our brains. When an innovation team finds an improvement, they usually sell access to it, which gives them resources for new projects, and lets others take advantage of their innovation.

...

In Bostrom’s graph above the line for an initially small project and system has a much higher slope, which means that it becomes in a short time vastly better at software innovation. Better than the entire rest of the world put together. And my key question is: how could it plausibly do that? Since the rest of the world is already trying the best it can to usefully innovate, and to abstract to promote such innovation, what exactly gives one small project such a huge advantage to let it innovate so much faster?

...

In fact, most software innovation seems to be driven by hardware advances, instead of innovator creativity. Apparently, good ideas are available but must usually wait until hardware is cheap enough to support them.

Yes, sometimes architectural choices have wider impacts. But I was an artificial intelligence researcher for nine years, ending twenty years ago, and I never saw an architecture choice make a huge difference, relative to other reasonable architecture choices. For most big systems, overall architecture matters a lot less than getting lots of detail right. Researchers have long wandered the space of architectures, mostly rediscovering variations on what others found before.

Some hope that a small project could be much better at innovation because it specializes in that topic, and much better understands new theoretical insights into the basic nature of innovation or intelligence. But I don’t think those are actually topics where one can usefully specialize much, or where we’ll find much useful new theory. To be much better at learning, the project would instead have to be much better at hundreds of specific kinds of learning. Which is very hard to do in a small project.

What does Bostrom say? Alas, not much. He distinguishes several advantages of digital over human minds, but all software shares those advantages. Bostrom also distinguishes five paths: better software, brain emulation (i.e., ems), biological enhancement of humans, brain-computer interfaces, and better human organizations. He doesn’t think interfaces would work, and sees organizations and better biology as only playing supporting roles.

...

Similarly, while you might imagine someday standing in awe in front of a super intelligence that embodies all the power of a new age, superintelligence just isn’t the sort of thing that one project could invent. As “intelligence” is just the name we give to being better at many mental tasks by using many good mental modules, there’s no one place to improve it. So I can’t see a plausible way one project could increase its intelligence vastly faster than could the rest of the world.

Takeoff speeds: https://sideways-view.com/2018/02/24/takeoff-speeds/
Futurists have argued for years about whether the development of AGI will look more like a breakthrough within a small group (“fast takeoff”), or a continuous acceleration distributed across the broader economy or a large firm (“slow takeoff”).

I currently think a slow takeoff is significantly more likely. This post explains some of my reasoning and why I think it matters. Mostly the post lists arguments I often hear for a fast takeoff and explains why I don’t find them compelling.

(Note: this is not a post about whether an intelligence explosion will occur. That seems very likely to me. Quantitatively I expect it to go along these lines. So e.g. while I disagree with many of the claims and assumptions in Intelligence Explosion Microeconomics, I don’t disagree with the central thesis or with most of the arguments.)
ratty  lesswrong  subculture  miri-cfar  ai  risk  ai-control  futurism  books  debate  hanson  big-yud  prediction  contrarianism  singularity  local-global  speed  speedometer  time  frontier  distribution  smoothness  shift  pdf  economics  track-record  abstraction  analogy  links  wiki  list  evolution  mutation  selection  optimization  search  iteration-recursion  intelligence  metameta  chart  analysis  number  ems  coordination  cooperate-defect  death  values  formal-values  flux-stasis  philosophy  farmers-and-foragers  malthus  scale  studying  innovation  insight  conceptual-vocab  growth-econ  egalitarianism-hierarchy  inequality  authoritarianism  wealth  near-far  rationality  epistemic  biases  cycles  competition  arms  zero-positive-sum  deterrence  war  peace-violence  winner-take-all  technology  moloch  multi  plots  research  science  publishing  humanity  labor  marginal  urban-rural  structure  composition-decomposition  complex-systems  gregory-clark  decentralized  heavy-industry  magnitude  multiplicative  endogenous-exogenous  models  uncertainty  decision-theory  time-prefer 
april 2018 by nhaliday
Information Processing: US Needs a National AI Strategy: A Sputnik Moment?
FT podcasts on US-China competition and AI: http://infoproc.blogspot.com/2018/05/ft-podcasts-on-us-china-competition-and.html

A new recommended career path for effective altruists: China specialist: https://80000hours.org/articles/china-careers/
Our rough guess is that it would be useful for there to be at least ten people in the community with good knowledge in this area within the next few years.

By “good knowledge” we mean they’ve spent at least 3 years studying these topics and/or living in China.

We chose ten because that would be enough for several people to cover each of the major areas listed (e.g. 4 within AI, 2 within biorisk, 2 within foreign relations, 1 in another area).

AI Policy and Governance Internship: https://www.fhi.ox.ac.uk/ai-policy-governance-internship/

https://www.fhi.ox.ac.uk/deciphering-chinas-ai-dream/
https://www.fhi.ox.ac.uk/wp-content/uploads/Deciphering_Chinas_AI-Dream.pdf
Deciphering China’s AI Dream
The context, components, capabilities, and consequences of
China’s strategy to lead the world in AI

Europe’s AI delusion: https://www.politico.eu/article/opinion-europes-ai-delusion/
Brussels is failing to grasp threats and opportunities of artificial intelligence.
By BRUNO MAÇÃES

When the computer program AlphaGo beat the Chinese professional Go player Ke Jie in a three-part match, it didn’t take long for Beijing to realize the implications.

If algorithms can already surpass the abilities of a master Go player, it can’t be long before they will be similarly supreme in the activity to which the classic board game has always been compared: war.

As I’ve written before, the great conflict of our time is about who can control the next wave of technological development: the widespread application of artificial intelligence in the economic and military spheres.

...

If China’s ambitions sound plausible, that’s because the country’s achievements in deep learning are so impressive already. After Microsoft announced that its speech recognition software surpassed human-level language recognition in October 2016, Andrew Ng, then head of research at Baidu, tweeted: “We had surpassed human-level Chinese recognition in 2015; happy to see Microsoft also get there for English less than a year later.”

...

One obvious advantage China enjoys is access to almost unlimited pools of data. The machine-learning technologies boosting the current wave of AI expansion are as good as the amount of data they can use. That could be the number of people driving cars, photos labeled on the internet or voice samples for translation apps. With 700 or 800 million Chinese internet users and fewer data protection rules, China is as rich in data as the Gulf States are in oil.

How can Europe and the United States compete? They will have to be commensurately better in developing algorithms and computer power. Sadly, Europe is falling behind in these areas as well.

...

Chinese commentators have embraced the idea of a coming singularity: the moment when AI surpasses human ability. At that point a number of interesting things happen. First, future AI development will be conducted by AI itself, creating exponential feedback loops. Second, humans will become useless for waging war. At that point, the human mind will be unable to keep pace with robotized warfare. With advanced image recognition, data analytics, prediction systems, military brain science and unmanned systems, devastating wars might be waged and won in a matter of minutes.

...

The argument in the new strategy is fully defensive. It first considers how AI raises new threats and then goes on to discuss the opportunities. The EU and Chinese strategies follow opposite logics. Already on its second page, the text frets about the legal and ethical problems raised by AI and discusses the “legitimate concerns” the technology generates.

The EU’s strategy is organized around three concerns: the need to boost Europe’s AI capacity, ethical issues and social challenges. Unfortunately, even the first dimension quickly turns out to be about “European values” and the need to place “the human” at the center of AI — forgetting that the first word in AI is not “human” but “artificial.”

https://twitter.com/mr_scientism/status/983057591298351104
https://archive.is/m3Njh
US military: "LOL, China thinks it's going to be a major player in AI, but we've got all the top AI researchers. You guys will help us develop weapons, right?"

US AI researchers: "No."

US military: "But... maybe just a computer vision app."

US AI researchers: "NO."

https://www.theverge.com/2018/4/4/17196818/ai-boycot-killer-robots-kaist-university-hanwha
https://www.nytimes.com/2018/04/04/technology/google-letter-ceo-pentagon-project.html
https://twitter.com/mr_scientism/status/981685030417326080
https://archive.is/3wbHm
AI-risk was a mistake.
hsu  scitariat  commentary  video  presentation  comparison  usa  china  asia  sinosphere  frontier  technology  science  ai  speedometer  innovation  google  barons  deepgoog  stories  white-paper  strategy  migration  iran  human-capital  corporation  creative  alien-character  military  human-ml  nationalism-globalism  security  investing  government  games  deterrence  defense  nuclear  arms  competition  risk  ai-control  musk  optimism  multi  news  org:mag  europe  EU  80000-hours  effective-altruism  proposal  article  realness  offense-defense  war  biotech  altruism  language  foreign-lang  philosophy  the-great-west-whale  enhancement  foreign-policy  geopolitics  anglo  jobs  career  planning  hmm  travel  charity  tech  intel  media  teaching  tutoring  russia  india  miri-cfar  pdf  automation  class  labor  polisci  society  trust  n-factor  corruption  leviathan  ethics  authoritarianism  individualism-collectivism  revolution  economics  inequality  civic  law  regulation  data  scale  pro-rata  capital  zero-positive-sum  cooperate-defect  distribution  time-series  tre 
february 2018 by nhaliday

bundles : dismalityeconembodiedguvnor

related tags

80000-hours  abstraction  acmtariat  ai  ai-control  algorithms  alien-character  alignment  altruism  analogy  analysis  anglo  anglosphere  arms  article  asia  audio  authoritarianism  automation  backup  barons  biases  big-yud  biotech  books  bostrom  britain  capital  career  charity  chart  china  civic  class  clever-rats  cocktail  cog-psych  commentary  comparison  competition  complement-substitute  complex-systems  composition-decomposition  computer-vision  conceptual-vocab  contrarianism  convexity-curvature  cooperate-defect  coordination  corporation  corruption  coupling-cohesion  creative  crux  cycles  data  death  debate  decentralized  decision-theory  deep-materialism  deepgoog  defense  definite-planning  detail-architecture  deterrence  dirty-hands  discussion  distribution  economics  eden  eden-heaven  EEA  effective-altruism  egalitarianism-hierarchy  EGT  ems  endogenous-exogenous  enhancement  epistemic  error  essay  estimate  ethics  EU  europe  evolution  evopsych  existence  farmers-and-foragers  flexibility  flux-stasis  foreign-lang  foreign-policy  formal-values  frontier  futurism  gallic  games  gedanken  generalization  geopolitics  gnon  google  government  gregory-clark  growth-econ  hanson  hardware  heavy-industry  hmm  hsu  human-capital  human-ml  humanity  hypocrisy  iidness  india  individualism-collectivism  inequality  innovation  insight  intel  intelligence  interview  intricacy  intuition  investing  iq  iran  iteration-recursion  japan  jobs  journos-pundits  labor  language  large-factor  law  left-wing  legacy  len:long  lesswrong  letters  leviathan  links  list  local-global  long-short-run  machine-learning  magnitude  malthus  marginal  martial  measurement  media  meta:war  metameta  migration  military  miri-cfar  model-organism  models  moloch  moments  morality  multi  multiplicative  musk  mutation  n-factor  nationalism-globalism  nature  near-far  network-structure  neuro  neuro-nitgrit  news  nitty-gritty  nonlinearity  nuclear  number  offense-defense  optimism  optimization  org:biz  org:foreign  org:lite  org:local  org:mag  org:popup  org:rec  org:sci  parable  parenting  pdf  peace-violence  philosophy  planning  plots  podcast  polisci  popsci  prediction  presentation  privacy  pro-rata  proposal  psychology  psychometrics  publishing  rationality  ratty  realness  recommendations  reference  regularizer  regulation  research  revolution  rhetoric  risk  robotics  russia  scale  science  scitariat  search  security  selection  sex  shift  singularity  sinosphere  sky  smoothness  social  society  software  speculation  speed  speedometer  ssc  stories  strategy  structure  studying  subculture  survey  sv  teaching  tech  technology  the-great-west-whale  thinking  threat-modeling  time  time-preference  time-series  top-n  track-record  trade  travel  trends  trust  tutoring  twitter  uncertainty  universalism-particularism  urban-rural  usa  values  video  war  wealth  white-paper  wiki  winner-take-all  yvain  zero-positive-sum  🐸 

Copy this bookmark:



description:


tags: