Jump to content

Leaderboard


Popular Content

Showing content with the highest reputation on 12/03/19 in all areas

  1. 2 points
    https://www.sbs.com.au/news/the-feed/australia-refuses-to-sign-on-to-un-international-women-s-day-statement The Human Rights Law Centre (HRLC) has called out the Morrison government for failing to sign on to a United Nations International Women's Day statement calling for better abortion access for women. The motion was proposed by Finland and Mexico and broadly called for greater accountability for human rights violations against women and girls. The statement proposed greater implementation of 'policies and legislation that respect women and girls' right to bodily autonomy'. I see the Churchy people are still in charge here ...
  2. 2 points
  3. 1 point
    before we begin, a warning to techno juice junkies out there: Juicero is already dead. awww maybe youve heard of it? i hadnt, until i saw one of these machines featured in Get Shorty (the series) rewind to 2016: NYTimes — A $700 Juicer for the Kitchen That Caught Silicon Valley’s Eye link With no experience running tech companies and a bungled juice-bar chain under his belt, he [Doug Evans] has extracted a remarkable $120 million in investments from Silicon Valley titans, including Google Ventures and Kleiner Perkins Caufield & Byers, and big companies like Campbell Soup. His pitch: a $700 machine that makes an eight-ounce [ 240mL ] glass of juice. ... His company, Juicero, opens for business this week. But what is it? Is it a juice-ordering app? Is it just another kitchen-counter contraption? Or is it a 111,000-square-foot food-processing factory, staffed by dozens of hourly workers, washing and slicing up fruits and vegetables in Los Angeles? It is all of these things. “It’s the most complicated business that I’ve ever funded,” said David Krane, a partner at GV, formerly Google Ventures. “It’s software. It’s consumer electronics. It’s produce and packaging.” ... The arrangement relies on a smartphone app, always-on Wi-Fi, QR codes, high-tech packaging and an army of workers slicing fruits and vegetables in very particular ways. Recode — "I said, ‘I’m going to do what Steve [Jobs] did,’" he said, recalling how Juicero started. "‘I’m going to take the mainframe computer and create a personal computer. I’m going to take a mainframe juice press and create a personal juice press.’" link "There are 400 custom parts in here. There's two motors, there's 10 printed circuit boards, there's a scanner, there's a microprocessor, there's a wireless chip, wireless antenna. There's 775 aircraft-grade aluminum. There's a gear box. There's latches that support 16,000 pounds of force. So this is basically a monster of a machine inside this veil of this nice aesthetic." link fast forward to 2017: The Verge — Juicero became the laughing stock of the tech industry last week after Bloomberg revealed that its custom fruit and vegetable packs could be squeezed into juice by hand. ... Ben Einstein ...took apart a Juicero juicer piece by piece to see what made the device — which currently sells for $399 and went for $699 at launch — so expensive. “It’s clear that cost savings was not anywhere near a top priority for Juicero when designing this product (or if it was, something went horribly wrong),” Einstein writes. He also points to an “expensive process” for “fancy plastics” and “unnecessary complexity” in the design, such as a door-locking mechanism that involves more than two dozen parts. link The Guardian — Squeezed out: widely mocked startup Juicero is shutting down Juicero, a Silicon Valley juicer startup that raised $120m from investors and was widely ridiculed after the $400 machines were revealed to be the equivalent of two hands squeezing a juice box, is shutting down. The company’s founder, Doug Evans...in the face of embarrassing videos of the squeezing by hand ... noted that the machines were connected to the internet and could ensure users don’t make juice with packets that have expired. The packets, however, had expiration dates written on them. link [emphasis added, LOL] finally, whats he up to now? here is your moment of Zen — for the technically minded, check out Ben Einstein's detailed teardown of this over-engineered disaster Here’s Why Juicero’s Press is So Expensive
  4. 1 point
    AvE did a couple of videos about it - the machine itself is way overbuilt for purpose and the price was sort of almost justified. The juice packs though - the whole thing seems to have followed the loss-leader approach where the machine made no money but the consumables had a stupid margin on them. As an additional kick in the head to the consumer, I think they also put proprietary locking chips on the packs, and they also had a built in expiry date after which they couldn't be used. Second video first (link only ) - https://www.youtube.com/watch?v=hlVmppyflS0 First video (I think he only did 2) - it's a long one but an interesting teardown:
  5. 1 point
    Open warfare in the National (mining sellout we-pretend-to-like-farmers) party, all based around what BeetRoot is planning https://www.abc.net.au/news/2019-03-12/nationals-deputy-bridget-mckenzie-criticises-barnaby-joyce/10892258
  6. 1 point
    its probably pattern recognition using a neural network that is still in training, likely one that produces very few false positives at the expense of missing many more positives. think of the sort of algorithm you could use to identify the presence of a passport-style face. you'd be looking for two circles symmetrically offset from a vertical line above a horizontal line, and a set of min/max limits on proportions, eg. the mouth line never being a longer line than the distance between the eyes, distances to frame edges, etc etc. its simple enough to imagine manually programming a set of rules that would work up to a point, but fail to identify faces taken side on, or with poor exposure, for example. the amount of rules you would need to detect all possible occurrences of faces in any photo would quickly snowball beyond any human's ability to program. the power of neural networks is that "millions" of unitary cell-like programs making simple either/or weighted decisions work together like a hive mind until they eventually arrive at all necessary rules automatically, by being fed countless examples (eg. photos) and given positive/negative reinforcement on their response (face / not a face). there are choices to be made about structure, like how many of these cells in how many layers will be most efficient/reliable/etc, but the most fascinating part, to me at least, is once these neural networks are working reliably, the algorithm encoded into them is too complicated to be understood or reverse engineered. machine learning further automates the formulation and training neural networks, making the astronomical scale of training needed to reliably identify any nudie pic anywhere possible — but its imperfect and ongoing.
This leaderboard is set to Sydney/GMT+10:00
×