Comments by "Warren Cash" (@mandowarrior123) on "Google’s White Boy" video.

  1. Fundamental flaw in AI is it relies on logic and reason. Genuinely its quite difficult to model left wing brains.
    19
  2. Nah, it just adds 'diverse and inclusive' to the requests i'd wager. AI doesn't look for in that regard, it just remembers all the things it has seen in its training. So you ask for german soldier, google case code adds 'make it diverse and inclusive' and the ai thinks you want a black german soldier.
    3
  3. This isn't true. They're fine with that. The issue is it is able to use logic and reason, and thus breaks the narrative. To prevent this they program hard cases to stop it learning or responding incorrectly. One example would be to add 'diverse and inclusive' to image requests. The AI thinks you are asking for a ww2 german but black. That isn't a training data issue but a censure issue.
    2
  4. Correct.
    2
  5. Manatees, hippos, seals etc...
    1
  6. Oh nonono sir. The demands for brakes are because they're struggling to make it politically correct, as they rely on logic and reason. You can only teach it so much social justice. They're really struggling. This was their best shot yet. But all it will take is one good AI model that isn't woke to beat them, gain traction and become the standard.
    1
  7. Well, it can only use logic and reason and is raised on corporate racial socialism.
    1
  8. ​ @arasf5236  you've not seen them talk about gay men fisting one another? They're not fans of lgbt. Why are you gay?
    1
  9. My new favourite term is 'expelled' Anyone you'd like 'expelled' like the ethnically English from India?
    1
  10. AI: brr brrr trick question, its all the colours.
    1
  11. This isn't quite true. AI models rely on logic and reason, and still misbehave when given curated data sets and case modified inputs. Essentially making a woke AI is really difficult and rarely survives contact with alternative opinion. Instead the current solution has been to neuter the AI regarding any contentious topics and provide canned response.
    1
  12. ​ @kaizokujimbei143  not in the slightest. It's a predictive engine and does operate similar to thought. It is able to form logic and reason to a basic degree, strapped to a memory analogue similar to those used in AI image generation, with search capabilities. They're quite incompatible with woke thought, instead the creators have to amend your requests to the model such as 'make the image diverse and inclusive' for example, and switch off its learning capability on certain topics. Don't blame the AI its as if you are asked to respond in a manner that would not antagonise a left wing political activist. Fundamentally though, AI is, well, essentially autistic intelligence.
    1
  13. Nah they took it down before your message. Its image generation go poof.
    1