Results 1 to 15 of 15

Thread: woman with fake "woke" quota Google job, "finding discrimination in AI code, gets too woke, Google shit cans black female engineer

  1. #1
    Diamond dwai's Avatar
    Reputation
    1653
    Join Date
    Jan 2013
    Posts
    7,855
    Load Metric
    65704726

    woman with fake "woke" quota Google job, "finding discrimination in AI code, gets too woke, Google shit cans black female engineer


  2. #2
    Plutonium sonatine's Avatar
    Reputation
    7369
    Join Date
    Mar 2012
    Posts
    33,372
    Load Metric
    65704726
    shes one of the leading specialists in artificial intelligence / machine learning.


    and you are... what.. exactly?

     
    Comments
      
      dwai: some guy?
    "Birds born in a cage think flying is an illness." - Alejandro Jodorowsky

    "America is not so much a nightmare as a non-dream. The American non-dream is precisely a move to wipe the dream out of existence. The dream is a spontaneous happening and therefore dangerous to a control system set up by the non-dreamers." -- William S. Burroughs

  3. #3
    Canadrunk limitles's Avatar
    Reputation
    1650
    Join Date
    Mar 2012
    Location
    In Todd's head
    Posts
    17,648
    Blog Entries
    1
    Load Metric
    65704726
    Quote Originally Posted by dwai View Post
    Regardless of the story and what you may think of it you cannot survive on crap headlines like this. It's barely intelligible and is only interesting to me because of the degree of fail.

    I'm sure there is no story to begin with but I will not look into

  4. #4
    Plutonium sonatine's Avatar
    Reputation
    7369
    Join Date
    Mar 2012
    Posts
    33,372
    Load Metric
    65704726
    google is claiming she essentially fired herself by refusing to back down from some ultimatum and if i had to guess id say thats probably exactly what happened.

    the guy heading the division she worked under has an absolutely spotless reputation btw.

    also the issue of racial bias in deep linguistic code is well documented so its unclear why she would choose this hill to die on tbh.
    "Birds born in a cage think flying is an illness." - Alejandro Jodorowsky

    "America is not so much a nightmare as a non-dream. The American non-dream is precisely a move to wipe the dream out of existence. The dream is a spontaneous happening and therefore dangerous to a control system set up by the non-dreamers." -- William S. Burroughs

  5. #5
    Owner Dan Druff's Avatar
    Reputation
    10110
    Join Date
    Mar 2012
    Posts
    54,627
    Blog Entries
    2
    Load Metric
    65704726
    It looks like she wanted to prove that AI is racist and therefore black people are oppressed by AI. She tried to publish a paper claiming that. The review process at Google rejected the paper, citing that it had various flaws, including failing to disclose that steps were already being taken to fix the AI racial bias in areas like facial recognition.

    She got all butthurt and said that she wanted the name of every single person involved in cross-researching her research, or she was quitting. She was told that it's not done that way, and that they don't want to create a witch hunt by revealing the names of every person they asked for information regarding that review. Obviously the fear was that she would go apeshit on social media and call everyone they spoke to racist, perhaps publicly. When they denied that request, she quit.

    Sounds like she let her identity politics get in the way of her work. This is actually a problem with woke researchers these days. They set out wanting to prove their social worldview, rather than attempt to research the facts. They attempt to ignore counter points found while researching, and then compose papers which purposely ignore or avoid these counter points.

    Good for Google for not playing this game. One of the rare instances of responsible behavior from them.

     
    Comments
      
      dwai: great summary

  6. #6
    Canadrunk limitles's Avatar
    Reputation
    1650
    Join Date
    Mar 2012
    Location
    In Todd's head
    Posts
    17,648
    Blog Entries
    1
    Load Metric
    65704726
    Quote Originally Posted by Dan Druff View Post
    Sounds like she let her identity politics get in the way of her work. This is actually a problem with woke researchers these days. They set out wanting to prove their social worldview, rather than attempt to research the facts.
    First of all you can't say there is a problem with "woke researchers these days"
    That is hilarious. How would you know?

    "Timnit Gebru had been working on a research paper that she hoped to publish, but ran into resistance from her superiors at Google.


    Gebru, an alumni of the Stanford Artificial Intelligence Laboratory, is one of the leading voices in the ethical use of artificial intelligence. She is well-known for her work on a landmark study in 2018 that showed how facial recognition software misidentified dark-skinned women as much as 35% of the time, whereas the technology worked with near precision on white men."

    Granted Gebru was an employee of a private organization and they are not obligated to publish anything from anybody.

    You said this about Abigail Shrier's book on transgender-ism being censored.

    "Should Target take this book off the shelves? Should Amazon refuse to run ads for it? Should legacy media refuse to allow journalists to review it? Should Gofundme have stopped the fundraiser for the book's promotion?"

    All of the above are privately run enterprises. Why are you not similarly outraged at what appears to be censorship?
    Last edited by limitles; 12-06-2020 at 09:32 PM.

  7. #7
    Puts His Dick in the Mashed Potatoes
    Reputation
    487
    Join Date
    Nov 2013
    Posts
    2,212
    Load Metric
    65704726
    Druff, I've never seen anybody as worried about wokeness/identity politics as you are Why are you so concerned about it?

  8. #8
    Owner Dan Druff's Avatar
    Reputation
    10110
    Join Date
    Mar 2012
    Posts
    54,627
    Blog Entries
    2
    Load Metric
    65704726
    Quote Originally Posted by rum dick View Post
    Druff, I've never seen anybody as worried about wokeness/identity politics as you are Why are you so concerned about it?
    Because it's leading to Democrats advocating for insane policies which will really hurt this country.

    Wokeness/identity politics has also been the basis for a recent assault on free speech -- a cause I've felt strongly about for over 30 years, even when it wasn't all the rage to be a free speech conservative.

  9. #9
    Owner Dan Druff's Avatar
    Reputation
    10110
    Join Date
    Mar 2012
    Posts
    54,627
    Blog Entries
    2
    Load Metric
    65704726
    Quote Originally Posted by limitles View Post
    Quote Originally Posted by Dan Druff View Post
    Sounds like she let her identity politics get in the way of her work. This is actually a problem with woke researchers these days. They set out wanting to prove their social worldview, rather than attempt to research the facts.
    First of all you can't say there is a problem with "woke researchers these days"
    That is hilarious. How would you know?

    "Timnit Gebru had been working on a research paper that she hoped to publish, but ran into resistance from her superiors at Google.


    Gebru, an alumni of the Stanford Artificial Intelligence Laboratory, is one of the leading voices in the ethical use of artificial intelligence. She is well-known for her work on a landmark study in 2018 that showed how facial recognition software misidentified dark-skinned women as much as 35% of the time, whereas the technology worked with near precision on white men."

    Granted Gebru was an employee of a private organization and they are not obligated to publish anything from anybody.

    You said this about Abigail Shrier's book on transgender-ism being censored.

    "Should Target take this book off the shelves? Should Amazon refuse to run ads for it? Should legacy media refuse to allow journalists to review it? Should Gofundme have stopped the fundraiser for the book's promotion?"

    All of the above are privately run enterprises. Why are you not similarly outraged at what appears to be censorship?
    She wanted to publish this through her workplace. They weren't objecting to her publishing a personal study on her own website, or whatever.

    She wanted this to be published through Google as a peer-reviewed paper. The peers reviewed it and said, "Nope, this sucks. Go back, change it, and stop omitting important counterpoints."

    She first demanded all the names of people who helped the peers review her work (so she could harass/blame them, which is totally against peer reviewing policy there), and then said she'd quit if they didn't agree to break this protocol.

    This is far different than a large retailer banning a book because woke people on Twitter pressured them (a decision which was reversed, by the way).

    If Timmit Gebru wants to publish the paper outside of Google, I think she should have every right to do so.

     
    Comments
      
      dwai: best summary I've read

  10. #10
    Owner Dan Druff's Avatar
    Reputation
    10110
    Join Date
    Mar 2012
    Posts
    54,627
    Blog Entries
    2
    Load Metric
    65704726
    In general, there's a big problem when someone does a study on a controversial topic, where they have already have a strong opinion on either side of that topic.

    In such cases, it tends to become an exercise of proving already-held biases, rather than doing quality academic research.

    For example, let's say someone does research on whether or not the BLM protests of 2020 caused a COVID spike. Let's say the person behind this study was a huge supporter of the BLM protests, and has already repeatedly stated that these protests did not contribute to worsening COVID.

    Would you trust a scientific research paper on the matter from this person? If you would, you're a fool.

    Why?

    Clearly there's zero point zero chance they'd ever release a paper with a conclusion counter to their strongly held prior beliefs. Thus, if the research reveals that the protests DID cause a COVID increase, they will either abandon the paper entirely or take a different approach in order to publish something which "proves" their previous belief. If you already know the person doing the research will never publish a paper counter to their beliefs prior to the study, then the research is automatically worthless.

    Of course, this is also true on the right. I would not trust a study on abortion done by right-wing Christians. I might take the information they present and attempt to independently verify it, but I would never trust their conclusions on their own.

    Unfortunately, in recent years many academics have let bias take precedence over science, to where it's more important to affirm their political beliefs than get to the actual truth.

  11. #11
    Banned
    Reputation
    489
    Join Date
    Jul 2015
    Posts
    699
    Load Metric
    65704726
    While it may be true that some academics have an axe to grind, this is why we have peer review. The scientific method works. She had sloppy methodology and it was easily sussed out.

    She is right about facial recognition though. It's a pervasive issue and well understood. This isn't because AI is "racist", but rather because these are developed by neural networks, which are trained from datasets of photographs, which typically aren't going to have a ton of black people in them.

    There are similar issues in voice recognition technology. My Google Assistant VR works perfectly, but sometimes when my phone isn't connected I have to use my car's voice recognition and it's horrible. It took me a while to figure this out, but I found that it's way more reliable if I use an approximation of a thick Japanese accent. This is probably because it was developed in Japan and trained to recognize the commands of Japanese people speaking English.

  12. #12
    100% Organic MumblesBadly's Avatar
    Reputation
    94
    Join Date
    Jun 2015
    Location
    In the many threads of this forum
    Posts
    9,408
    Load Metric
    65704726
    Quote Originally Posted by Rick Sanchez View Post
    While it may be true that some academics have an axe to grind, this is why we have peer review. The scientific method works. She had sloppy methodology and it was easily sussed out.

    She is right about facial recognition though. It's a pervasive issue and well understood. This isn't because AI is "racist", but rather because these are developed by neural networks, which are trained from datasets of photographs, which typically aren't going to have a ton of black people in them.

    There are similar issues in voice recognition technology. My Google Assistant VR works perfectly, but sometimes when my phone isn't connected I have to use my car's voice recognition and it's horrible. It took me a while to figure this out, but I found that it's way more reliable if I use an approximation of a thick Japanese accent. This is probably because it was developed in Japan and trained to recognize the commands of Japanese people speaking English.
    Much of what you say here is true, and that paper on the likely-unintential racial bias of facial recognition software may well have been validly peer reviewed, but not all academic publishing is subject to rigorous scientific peer review, and in particular the more social-jucticy fields of study. Here is a video featuring a brutal takedown of the academic publishing standards. Here’s a hilarious send-up of an intentional hoax prank of the publishing standards of some of the major grievance studies journals.



    P.S. The academics who conducted this purpose-driven hoax (to expose bias-driven corruption in the publishing practices of some of those leading grievance studies journals) self-identify as liberals.
    Last edited by MumblesBadly; 12-07-2020 at 11:00 AM.
    _____________________________________________
    Quote Originally Posted by Dan Druff View Post
    I actually hope this [second impeachment] succeeds, because I want Trump put down politically like a sick, 14-year-old dog. ... I don't want him complicating the 2024 primary season. I just want him done.
    Quote Originally Posted by Dan Druff View Post
    Were Republicans cowardly or unethical not to go along with [convicting Trump in the second impeachment Senate trial]? No. The smart move was to reject it.

  13. #13
    Banned
    Reputation
    489
    Join Date
    Jul 2015
    Posts
    699
    Load Metric
    65704726
    Quote Originally Posted by MumblesBadly View Post
    but not all academic publishing is subject to rigorous scientific peer review, and in particular the more social-jucticy fields of study.
    Yeah the social "sciences" have all sorts of issues. Fortunately that's not as much of an issue in fields like AI research and mathematics. Like you can't just fake solve the Reimann Hypothesis and claim your $1 million lol

  14. #14
    Plutonium sonatine's Avatar
    Reputation
    7369
    Join Date
    Mar 2012
    Posts
    33,372
    Load Metric
    65704726
    https://www.theverge.com/2020/12/7/2...tatement-fired


    some interesting data/clarification therein.

    people get fucked up about publicizing their papers, especially when those papers are retracted. factor in the aspie genius element and its easy to see how this sailed off the fucking rails.
    "Birds born in a cage think flying is an illness." - Alejandro Jodorowsky

    "America is not so much a nightmare as a non-dream. The American non-dream is precisely a move to wipe the dream out of existence. The dream is a spontaneous happening and therefore dangerous to a control system set up by the non-dreamers." -- William S. Burroughs

  15. #15
    Owner Dan Druff's Avatar
    Reputation
    10110
    Join Date
    Mar 2012
    Posts
    54,627
    Blog Entries
    2
    Load Metric
    65704726
    I don't doubt that there is an unintentional racial bias in the technology. Of course that should be addressed. However, supposedly there are efforts underway to fix this, so that's really all the industry can do.

    Not everything has an evil motive to it. Too often you have woke leftists trying to equate inequity to an intentionally biased system.

Thread Information

Users Browsing this Thread

There are currently 1 users browsing this thread. (0 members and 1 guests)

Similar Threads

  1. The next revolution in home internet? Google rolls out "Gigabit"
    By Dan Druff in forum The Hacker's Delight
    Replies: 4
    Last Post: 09-01-2022, 04:37 PM
  2. Replies: 1
    Last Post: 11-28-2020, 06:05 AM
  3. Replies: 4
    Last Post: 03-01-2020, 10:11 AM
  4. Replies: 3
    Last Post: 05-20-2015, 03:37 PM
  5. Replies: 2
    Last Post: 05-16-2014, 01:26 PM