core-rats   78

« earlier    

Why read old philosophy? | Meteuphoric
(This story would suggest that in physics students are maybe missing out on learning the styles of thought that produce progress in physics. My guess is that instead they learn them in grad school when they are doing research themselves, by emulating their supervisors, and that the helpfulness of this might partially explain why Nobel prizewinner advisors beget Nobel prizewinner students.)

The story I hear about philosophy—and I actually don’t know how much it is true—is that as bits of philosophy come to have any methodological tools other than ‘think about it’, they break off and become their own sciences. So this would explain philosophy’s lone status in studying old thinkers rather than impersonal methods—philosophy is the lone ur-discipline without impersonal methods but thinking.

This suggests a research project: try summarizing what Aristotle is doing rather than Aristotle’s views. Then write a nice short textbook about it.
ratty  learning  reading  studying  prioritizing  history  letters  philosophy  science  comparison  the-classics  canon  speculation  reflection  big-peeps  iron-age  mediterranean  roots  lens  core-rats  thinking  methodology  grad-school  academia  physics  giants  problem-solving  meta:research  scholar  the-trenches  explanans  crux  metameta  duplication  sociality  innovation  quixotic 
june 2018 by nhaliday
trees are harlequins, words are harlequins — bayes: a kinda-sorta masterpost
lol, gwern:
> What sort of person thinks “oh yeah, my beliefs about these coefficients correspond to a Gaussian with variance 2.5″? And what if I do cross-validation, like I always do, and find that variance 200 works better for the problem? Was the other person wrong? But how could they have known?
> ...Even ignoring the mode vs. mean issue, I have never met anyone who could tell whether their beliefs were normally distributed vs. Laplace distributed. Have you?
I must have spent too much time in Bayesland because both those strike me as very easy and I often think them! My beliefs usually are Laplace distributed when it comes to things like genetics (it makes me very sad to see GWASes with flat priors), and my Gaussian coefficients are actually a variance of 0.70 (assuming standardized variables w.l.o.g.) as is consistent with field-wide meta-analyses indicating that d>1 is pretty rare.
ratty  ssc  core-rats  tumblr  social  explanation  init  philosophy  bayesian  thinking  probability  stats  frequentist  big-yud  lesswrong  synchrony  similarity  critique  intricacy  shalizi  scitariat  selection  mutation  evolution  priors-posteriors  regularization  bias-variance  gwern  reddit  commentary  GWAS  genetics  regression  spock  nitty-gritty  generalization  epistemic  🤖  rationality  poast  multi  best-practices  methodology  data-science 
august 2017 by nhaliday
Superintelligence Risk Project Update II
For example, I asked him what he thought of the idea that to we could get AGI with current techniques, primarily deep neural nets and reinforcement learning, without learning anything new about how intelligence works or how to implement it ("Prosaic AGI" [1]). He didn't think this was possible, and believes there are deep conceptual issues we still need to get a handle on. He's also less impressed with deep learning than he was before he started working in it: in his experience it's a much more brittle technology than he had been expecting. Specifically, when trying to replicate results, he's often found that they depend on a bunch of parameters being in just the right range, and without that the systems don't perform nearly as well.

The bottom line, to him, was that since we are still many breakthroughs away from getting to AGI, we can't productively work on reducing superintelligence risk now.

He told me that he worries that the AI risk community is not solving real problems: they're making deductions and inferences that are self-consistent but not being tested or verified in the world. Since we can't tell if that's progress, it probably isn't. I asked if he was referring to MIRI's work here, and he said their work was an example of the kind of approach he's skeptical about, though he wasn't trying to single them out. [2]
Earlier this week I had a conversation with an AI researcher [1] at one of the main industry labs as part of my project of assessing superintelligence risk. Here's what I got from them:

They see progress in ML as almost entirely constrained by hardware and data, to the point that if today's hardware and data had existed in the mid 1950s researchers would have gotten to approximately our current state within ten to twenty years. They gave the example of backprop: we saw how to train multi-layer neural nets decades before we had the computing power to actually train these nets to do useful things.

Similarly, people talk about AlphaGo as a big jump, where Go went from being "ten years away" to "done" within a couple years, but they said it wasn't like that. If Go work had stayed in academia, with academia-level budgets and resources, it probably would have taken nearly that long. What changed was a company seeing promising results, realizing what could be done, and putting way more engineers and hardware on the project than anyone had previously done. AlphaGo couldn't have happened earlier because the hardware wasn't there yet, and was only able to be brought forward by massive application of resources.
Summary: I'm not convinced that AI risk should be highly prioritized, but I'm also not convinced that it shouldn't. Highly qualified researchers in a position to have a good sense the field have massively different views on core questions like how capable ML systems are now, how capable they will be soon, and how we can influence their development. I do think these questions are possible to get a better handle on, but I think this would require much deeper ML knowledge than I have.
ratty  core-rats  ai  risk  ai-control  prediction  expert  machine-learning  deep-learning  speedometer  links  research  research-program  frontier  multi  interview  deepgoog  games  hardware  performance  roots  impetus  chart  big-picture  state-of-art  reinforcement  futurism  🤖  🖥  expert-experience  singularity  miri-cfar  empirical  evidence-based  speculation  volo-avolo  clever-rats  acmtariat  robust  ideas  crux  atoms  detail-architecture  software  gradient-descent 
july 2017 by nhaliday
In Defense of Individualist Culture | Otium
The salient feature of an individualist environment is that nobody directly tries to make you do anything.


I see a lot of writers these days raising problems with modern individualist culture, and it may be an especially timely topic. The Internet is a novel superstimulus, and it changes more rapidly, and affords people more options, than ever before. We need to think about the actual consequences of a world where many people are in practice being left alone to do what they want, and clearly not all the consequences are positive.

But I do want to suggest some considerations in favor of individualist culture — that often-derided “atomized modern world” that most of us live in.

We Aren’t Clay


Patriarchy is the Problem:
ratty  core-rats  rhetoric  values  social-norms  society  anthropology  individualism-collectivism  higher-ed  labor  incentives  habit  internet  regularizer  behavioral-gen  biodet  ego-depletion  psychology  social-psych  thinking  rationality  tradition  egalitarianism-hierarchy  murray  putnam-like  coming-apart  cohesion  modernity  migration  essay  n-factor  multi  tumblr  social  yvain  ssc  critique  commentary  debate  moloch  community  civil-liberty  truth  cooperate-defect  enlightenment-renaissance-restoration-reformation  markets  open-closed  gender  farmers-and-foragers  religion  christianity  judaism  theos  social-structure  authoritarianism 
june 2017 by nhaliday

« earlier    

related tags

2016-election  2016  academia  accuracy  acm  acmtariat  additive  advice  age-of-discovery  aggregator  ai-control  ai  akrasia  algebra  alignment  altruism  analysis  analytical-holistic  anglo  anthropology  aphorism  arbitrage  arms  arrows  article  asia  atoms  audio  authoritarianism  automation  aversion  bayesian  beeminder  behavioral-gen  being-right  best-practices  bias-variance  biases  big-peeps  big-picture  big-yud  biodet  biohacking  biotech  blog  books  bounded-cognition  brain-scan  broad-econ  california  cancer  canon  career  chapman  charity  chart  checklists  china  christianity  civil-liberty  civilization  clarity  clever-rats  coalitions  cocktail  cohesion  coming-apart  commentary  community  comparison  compensation  computer-vision  confluence  confusion  contrarianism  cooperate-defect  cost-benefit  creative  crime  criminal-justice  criminology  critique  crux  culture  curiosity  current-events  cynicism-idealism  data-science  data  database  debate  decision-making  decision-theory  deep-learning  deepgoog  definite-planning  demographics  density  detail-architecture  developmental  diaspora  discovery  discussion  divergence  drugs  duplication  early-modern  economics  econotariat  eea  effective-altruism  efficiency  egalitarianism-hierarchy  ego-depletion  embodied-cognition  embodied-pack  embodied-street-fighting  embodied  emotion  empirical  energy-resources  enhancement  enlightenment-renaissance-restoration-reformation  entertainment  environment  epistemic  essay  ethics  europe  evidence-based  evolution  exocortex  experiment  expert-experience  expert  explanans  explanation  exposition  faq  farmers-and-foragers  fda  fiction  fighting  flexibility  foreign-lang  formal-values  frequentist  frisson  frontier  futurism  games  gedanken  gender  generalization  genetic-load  genetics  geography  germanic  giants  google  gotchas  grad-school  gradient-descent  gray-econ  great-powers  ground-up  growth-econ  growth  gtd  gwas  gwern  habit  hacker  haidt  hanson  hardware  health  healthcare  heavy-industry  heterodox  hi-order-bits  hidden-motives  high-variance  higher-ed  history  hmm  howto  hsu  human-bean  ideas  ideology  idk  impact  impetus  impro  incentives  individualism-collectivism  industrial-revolution  info-dynamics  info-econ  info-foraging  infographic  infrastructure  init  innovation  insight  integrity  internet  interview  intricacy  ioannidis  iq  iron-age  iteration-recursion  judaism  labor  language  learning  len:long  len:short  lens  lesswrong  let-me-see  letters  leviathan  lifehack  linear-algebra  linear-models  linearity  links  list  logic  lol  long-term  machine-learning  management  maps  marginal  markets  math  mathtariat  measurement  media  medicine  mediterranean  mental-math  meta-analysis  meta:medicine  meta:prediction  meta:research  meta:rhetoric  meta:science  meta:war  metabolic  metabuch  metameta  methodology  michael-nielsen  microsoft  migration  minimalism  miri-cfar  model-class  models  modernity  moloch  moments  money-for-time  money  morality  mostly-modern  motivation  multi  multiplicative  murray  music  mutation  mystic  n-factor  narrative  nature  near-far  neuro-nitgrit  neuro  neurons  nibble  nitty-gritty  nl-and-so-can-you  nlp  noise-structure  nonlinearity  nootropics  novelty  nuclear  nutrition  objektbuch  occam  occident  oly  open-closed  open-problems  optimism  order-disorder  org:bleg  org:ngo  organization  orient  outcome-risk  overflow  p:***  p:whenever  parenting  paying-rent  people  performance  persuasion  phalanges  pharma  phd  philosophy  physics  pic  planning  play  poast  podcast  polarization  policy  politics  population-genetics  population  postmortem  postrat  power  practice  prediction-markets  prediction  presentation  prioritizing  priors-posteriors  probability  problem-solving  procrastination  productivity  proposal  psychology  putnam-like  q-n-a  qtl  quantified-self  questions  quixotic  quotes  randy-ayndy  rat-pack  rationality  ratty  reading  realness  reason  recommendations  reddit  reference  reflection  regression  regularization  regularizer  regulation  reinforcement  religion  replication  research-program  research  retention  retrofit  review  revolution  rhetoric  risk  robust  roots  s-factor  s:*  scaling-up  schelling  scholar  science  scitariat  securities  selection  serene  shalizi  shift  signaling  similarity  singularity  sinosphere  skeleton  skunkworks  sleep  smoothness  social-norms  social-psych  social-science  social-structure  social  sociality  society  sociology  socs-and-mops  soft-question  software  speaking  speculation  speedometer  spock  ssc  startups  state-of-art  stats  status  stories  strategy  stream  street-fighting  stress  study  studying  subculture  success  summary  survey  sv  synchrony  synthesis  systematic-ad-hoc  tactics  tech  technology  techtariat  tetlock  the-classics  the-great-west-whale  the-monster  the-trenches  theos  things  thinking  time-series  time-use  time  tools  top-n  toxoplasmosis  track-record  trade  tradition  travel  trends  tricks  trump  trust  truth  tumblr  uncertainty  unintended-consequences  urban-rural  urban  usa  vague  values  variance-components  video  visualization  vitality  volo-avolo  war  wealth-of-nations  webapp  whiggish-hegelian  wild-ideas  wire-guided  wonkish  workflow  working-stiff  workshop  world-war  worrydream  writing  x-not-about-y  yak-shaving  yvain  zeitgeist  🌞  🎓  🎩  🔬  🖥  🤖  🦀  🦉 

Copy this bookmark: