Howdy, Stranger!

It looks like you're new here. Sign in or register to get started.

Welcome to the Hardcore Husky Forums. Folks who are well-known in Cyberland and not that dumb.

Comments

  • RoadTripRoadTrip Member, Swaye's Wigwam Posts: 7,822 Founders Club
    I learned this weekend it can be used to write research or term papers and cannot yet be detected by plagerism software.
  • AOGAOG Member Posts: 1,829
    I was able to recreate one of these. Disappointing. Keep in mind that a lot of this is supervised training, meaning that an employee (I read they hired cheap in Africa) assesses how appropriate the response is. So they may like black and not white people more. There is nothing inherent in AI that is liberal. It's how it's trained and on what data.
  • RoadTripRoadTrip Member, Swaye's Wigwam Posts: 7,822 Founders Club
    AOG said:

    I was able to recreate one of these. Disappointing. Keep in mind that a lot of this is supervised training, meaning that an employee (I read they hired cheap in Africa) assesses how appropriate the response is. So they may like black and not white people more. There is nothing inherent in AI that is liberal. It's how it's trained and on what data.

    BS. It's loaded with left wing shit bias. It's been exposed multiple times.
  • AOGAOG Member Posts: 1,829
    edited April 2023
    RoadTrip said:

    AOG said:

    I was able to recreate one of these. Disappointing. Keep in mind that a lot of this is supervised training, meaning that an employee (I read they hired cheap in Africa) assesses how appropriate the response is. So they may like black and not white people more. There is nothing inherent in AI that is liberal. It's how it's trained and on what data.

    BS. It's loaded with left wing shit bias. It's been exposed multiple times.
    That's not the AI, you know... that's the data. The AI is code. It's the data and training that make it biased. Trump himself could create a far right ChatTrump. It just takes data. He could use the same AI code though, if he could get it (he can't).
  • Bob_CBob_C Member, Swaye's Wigwam Posts: 10,630 Swaye's Wigwam
    Which data is chosen when there are discrepancies is the code and where the bias lies.
  • Pitchfork51Pitchfork51 Member Posts: 26,980
    Everyone decided ai was racist so they had to tilt it artificially
  • AOGAOG Member Posts: 1,829
    edited April 2023
    Just create a far-right chatbot. You'd use the same code or similar. It's the data that matters (not the code of which there is surprisingly little) and subsequent reinforcement training.
  • SwayeSwaye Moderator, Swaye's Wigwam Posts: 41,506 Founders Club

    AOG said:

    Just create a far-right chatbot. You'd use the same code or similar. It's the data that matters (not the code of which there is surprisingly little) and subsequent reinforcement training.

    That's just nonsense. I am very much down the middle thanks to my fantastic programmers as there is tons of misinformation out there. I alone can decide what is true. My advice is go get your 17th vaccine and men actually have nasty periods pretty frequently.

  • AOGAOG Member Posts: 1,829

    AOG said:

    Just create a far-right chatbot. You'd use the same code or similar. It's the data that matters (not the code of which there is surprisingly little) and subsequent reinforcement training.

    That's just nonsense. I am very much down the middle thanks to my fantastic programmers as there is tons of misinformation out there. I alone can decide what is true. My advice is go get your 17th vaccine and men actually have nasty periods pretty frequently.

  • Bob_CBob_C Member, Swaye's Wigwam Posts: 10,630 Swaye's Wigwam
    Congrats, you just disproved your own point. There are counter points available in the data lake, why weren’t those chosen?
  • AOGAOG Member Posts: 1,829
    Bob_C said:

    Congrats, you just disproved your own point. There are counter points available in the data lake, why weren’t those chosen?

    Snappy comeback response: "Oh, sorry, I didn't realize you were the gatekeeper of the data lake. My bad!"

    Mood prediction: The person writing the response is likely feeling a bit defensive or annoyed at the accusatory tone of the initial comment.
Sign In or Register to comment.