nhaliday + gig-econ   17

Michael Akilian: Worker-in-the-loop Retrospective
Over the last ten years, many companies have created human-in-the-loop services that combine a mix of humans and algorithms. Now that some time has passed, we can tease out some patterns from their collective successes and failures. As someone who started a company in this space, my hope is that this retrospective can help prospective founders, investors, or companies navigating this space save time and fund more impactful projects.

A service is considered human-in-the-loop if it organizes its workflows with the intent to introduce models or heuristics that learn from the work of the humans executing the workflows. In this post, I will make reference to two common forms of human-in-the-loop:

User-in-the-loop (UITL): The end-user is interacting with suggestions from a software heuristic/ML system.
Worker-in-the-loop (WITL): A worker is paid to monitor suggestions from a software heuristic/ML system developed by the same company that pays the worker, but for the ultimate benefit of an end-user.
techtariat  reflection  business  tech  postmortem  automation  startups  hard-tech  ai  machine-learning  human-ml  cost-benefit  analysis  thinking  business-models  things  dimensionality  exploratory  markets  labor  economics  tech-infrastructure  gig-econ 
6 weeks ago by nhaliday
Federal University | West Hunter
If, as a pilot program, an example, the government set up a new university, mindlessly copying a decent state school from that golden era, like Berkeley or Wisconsin (or maybe from a bit earlier, since we probably want to avoid riots too), I doubt if it would cost a lot more. All those extra administrative personnel? Just don’t hire them. We could manage this by making the project top secret (actually, special access) – that lets you violate a lot of the useless bureaucratic rules, rather like being Uber.

Some things might cost more. If you want a medical school, you have to pay the professors competitive salaries (and MDs make much more than they did back in those days). But then, we could used taped lectures, online courses, etc.

It probably wouldn’t work for long, since politicians would be irresistibly temped to add on useless crap, like preferential admission for Skoptys, or whatever they’re called nowadays.

https://westhunt.wordpress.com/2016/03/30/federal-university/#comment-77371
“Between 1975 and 2005, total spending by American higher educational institutions, stated in constant dollars, tripled, to more than $325 billion per year. Over the same period, the faculty-to-student ratio has remained fairly constant, at approximately fifteen or sixteen students per instructor. One thing that has changed, dramatically, is the administrator-per-student ratio. In 1975, colleges employed one administrator for every eighty-four students and one professional staffer—admissions officers, information technology specialists, and the like—for every fifty students. By 2005, the administrator-to-student ratio had dropped to one administrator for every sixty-eight students while the ratio of professional staffers had dropped to one for every twenty-one students. “

Higher Education In Mass. Enters Full Predatory Mode: http://news.wgbh.org/2016/12/08/local-news/higher-education-mass-enters-full-predatory-mode
academic administrators
https://home.isi.org/somewhere-between-jeremiad-and-eulogy

https://westhunt.wordpress.com/2016/03/30/federal-university/#comment-77423
I would put the kind of knowledge that you acquire in college into four categories. Obviously majors differ in their mix of these four humours. I’m thinking of economic/GDP/health type impacts.

Things that don’t matter. Like neutral genetic variation.
Things that make you better at doing something useful. Ideally, significantly better – at least better at the task than if you’d just spend an hour or two reading the manual.

Things that make you better at inventing techniques in category 2. What Edison, George Green, or Ramanujan learned in college. Overlaps with #2.

Things that ain’t so. Falsehoods. Ones with practical implications. There are obviously some majors that mostly inculcate falsehoods.

Now some of these can be used for signalling, but the content of education matters (in the broad sense – college but also reading Popular Mechanics). If it didn’t we’d all be living in caves and licking mammoth fat off our fingers.

https://westhunt.wordpress.com/2016/03/30/federal-university/#comment-77528
It can also simply be ignored: lots of Silicon Valley companies give pretty explicit IQ tests without ever bothering to get them approved.

https://twitter.com/tcjfs/status/870450589955756032
https://archive.is/7Xm5y
I used to think this, but now I wonder if the degree is used more as a signal of willingness to put up with institutional BS rather than IQ.
--
Yeah, Griggs is terrible, ham-fisted law, shd be overturned. But overrated as a cause of the edu bubble

- thinks its mostly subsidies not ban on IQ testing
- still getting good tests for cognitive ability plus non-cognitive habits, then moving to new equilibrium should be enough right?

Modern Universities Are An Exercise in Insanity: http://scholars-stage.blogspot.com/2018/01/modern-universities-are-exercise-in.html

My alma mater was Brigham Young University-Hawaii. If you are a member of the LDS church attending the school, then in 2017 your tuition was $3,000 a semester. If you are not a member, it was $5,000 for one semester. The school has a special program where you can graduate in three years by taking three semesters each year, and that costs $8,000 and $16,000 a year for LDS and non-member students respectively.

...

The average tenure track professor makes $40 an hour. If you were to employ her as a private tutor at the cost of $60 an hour, and had four hours with her a week, and did that for 14 weeks (that's the length of an average college course folks) that is about $3,400.

Were you to employ three such professor-tutors, that would be about $10,200, or a bit over $20,000 a year. In four years you would have racked up $80,000 in costs. But this is still $30,000 less than the total for the 'cost conscious' universities. It is a quarter of what you would pay for Trinity.
west-hunter  rant  education  higher-ed  institutions  government  proposal  discussion  policy  rent-seeking  scitariat  efficiency  cost-disease  counter-revolution  alt-inst  regulation  ideas  multi  unaffiliated  broad-econ  wonkish  other-xtian  debt  cost-benefit  analysis  money  fertility  intervention  hmm  planning  long-term  parenting  knowledge  signaling  human-capital  truth  realness  poast  pro-rata  gender  sv  tech  recruiting  iq  pinker  trends  critique  news  current-events  vampire-squid  org:ngo  academia  technocracy  gnon  right-wing  twitter  social  speculation  roots  malaise  law  business  industrial-org  psychometrics  race  discrimination  diversity  cycles  impetus  chart  sex  sexuality  judgement  gig-econ 
january 2017 by nhaliday
How Uber Drivers Decide How Long to Work - The New York Times
Over all, there was little evidence that drivers were driving less when they could make more per hour than usual. But that was not true for a large portion of new drivers. Many of these drivers appeared to have an income goal in mind and stopped when they were near it, causing them to knock off sooner when their hourly wage was high and to work longer when their wage was low.
economics  labor  data  news  psychology  micro  behavioral-econ  org:rec  gig-econ 
september 2016 by nhaliday

related tags

academia  aggregator  ai  alt-inst  analogy  analysis  announcement  applications  atoms  auto-learning  automation  barons  behavioral-econ  broad-econ  business  business-models  chart  collaboration  commentary  communication  community  compensation  concept  consulting-freelance  contrarianism  cool  cost-benefit  cost-disease  counter-revolution  critique  current-events  cycles  data  database  dbs  debt  deep-learning  dimensionality  discrimination  discussion  diversity  drama  economics  econotariat  education  efficiency  engineering  ethics  exploratory  fermi  fertility  flexibility  freelance  gender  gig-econ  gnon  google  government  gradient-descent  hard-tech  higher-ed  hmm  hn  human-capital  human-ml  ideas  idk  impetus  industrial-org  institutions  intervention  interview-prep  investing  iq  jobs  judgement  knowledge  labor  law  liner-notes  long-term  machine-learning  malaise  marginal-rev  markets  measurement  media  micro  microbiz  micropayments  minimum-viable  money  multi  musk  natural-experiment  neuro  news  nibble  org:biz  org:bv  org:com  org:lite  org:mag  org:ngo  org:rec  organization  oss  other-xtian  papers  parenting  pinker  planning  poast  policy  postmortem  pragmatic  prediction  pro-rata  programming  project  proposal  psychology  psychometrics  race  rant  realness  recruiting  reflection  regulation  reinforcement  rent-seeking  research  rhetoric  right-wing  roots  scitariat  sex  sexuality  signaling  skunkworks  social  socs-and-mops  software  speculation  startups  street-fighting  supply-demand  sv  system-design  tech  tech-infrastructure  technocracy  techtariat  things  thinking  tools  transportation  trends  truth  twitter  unaffiliated  vampire-squid  venture  west-hunter  wire-guided  wonkish 

Copy this bookmark:



description:


tags: