Skip to main content
added 2 characters in body
Source Link
user.dz
  • 48.7k
  • 14
  • 14

Tactical Withdraw, may be?

Bear with me as this was my chain of thinking. Personally, I still believed AI generated contents is of unknown quality but expected to be low most of the time.

So what the current options to handle it:

  1. Ask Ubuntu community policy: Ban it ...

    Experienced users expected to flag it. Moderator review then reject or hide it from public and it get deleted later.

  2. Stack Exchange network policy: Leave it be ...

    Administrator are asked to close eyes. Users are not pushed either way.

  3. I tried to come up with a in-middle option: Label it ...

    User expected to flag it. Moderator review then reject or a special AI label to make users aware.

Back to it now after I reread this postthread again. Option 1 and 3 will give bad results on long run. Because of the way AI works.

AI feedback loop

  1. Stack Exchange network data are already public and they are also serving it as API. That gives AI users a head start. So that data is used as training set.

  2. AI users will contribute some posts here.

  3. And they will wait for feedback!. Whatever we do with them votes, flags, comments, deletion...

This feedback will be very helpful, to continues AI learning. 1 and 3 will just levels up the AI and it will be hard with time to distinguish soon or later. 2 is not perfect but delaying the hide/deletion action will slow down the AI learning process.

May be it is wrong for me to put this in public? Anyway CM already gave some hints of SE secret scroll.

Tactical Withdraw, may be?

Bear with me as this was my chain of thinking. Personally, I still believed AI generated contents is of unknown quality but expected to be low most of the time.

So what the current options to handle it:

  1. Ask Ubuntu community policy: Ban it ...

    Experienced users expected to flag it. Moderator review then reject or hide it from public and it get deleted later.

  2. Stack Exchange network policy: Leave it be ...

    Administrator are asked to close eyes. Users are not pushed either way.

  3. I tried to come up with a in-middle option: Label it ...

    User expected to flag it. Moderator review then reject or a special AI label to make users aware.

Back to it now after I reread this post again. Option 1 and 3 will give bad results on long run. Because of the way AI works.

AI feedback loop

  1. Stack Exchange network data are already public and they are also serving it as API. That gives AI users a head start. So that data is used as training set.

  2. AI users will contribute some posts here.

  3. And they will wait for feedback!. Whatever we do with them votes, flags, comments, deletion...

This feedback will be very helpful, to continues AI learning. 1 and 3 will just levels up the AI and it will be hard with time to distinguish soon or later. 2 is not perfect but delaying the hide/deletion action will slow down the AI learning process.

May be it is wrong for me to put this in public? Anyway CM already gave some hints of SE secret scroll.

Tactical Withdraw, may be?

Bear with me as this was my chain of thinking. Personally, I still believed AI generated contents is of unknown quality but expected to be low most of the time.

So what the current options to handle it:

  1. Ask Ubuntu community policy: Ban it ...

    Experienced users expected to flag it. Moderator review then reject or hide it from public and it get deleted later.

  2. Stack Exchange network policy: Leave it be ...

    Administrator are asked to close eyes. Users are not pushed either way.

  3. I tried to come up with a in-middle option: Label it ...

    User expected to flag it. Moderator review then reject or a special AI label to make users aware.

Back to it now after I reread this thread again. Option 1 and 3 will give bad results on long run. Because of the way AI works.

AI feedback loop

  1. Stack Exchange network data are already public and they are also serving it as API. That gives AI users a head start. So that data is used as training set.

  2. AI users will contribute some posts here.

  3. And they will wait for feedback!. Whatever we do with them votes, flags, comments, deletion...

This feedback will be very helpful, to continues AI learning. 1 and 3 will just levels up the AI and it will be hard with time to distinguish soon or later. 2 is not perfect but delaying the hide/deletion action will slow down the AI learning process.

May be it is wrong for me to put this in public? Anyway CM already gave some hints of SE secret scroll.

Source Link
user.dz
  • 48.7k
  • 14
  • 14

Tactical Withdraw, may be?

Bear with me as this was my chain of thinking. Personally, I still believed AI generated contents is of unknown quality but expected to be low most of the time.

So what the current options to handle it:

  1. Ask Ubuntu community policy: Ban it ...

    Experienced users expected to flag it. Moderator review then reject or hide it from public and it get deleted later.

  2. Stack Exchange network policy: Leave it be ...

    Administrator are asked to close eyes. Users are not pushed either way.

  3. I tried to come up with a in-middle option: Label it ...

    User expected to flag it. Moderator review then reject or a special AI label to make users aware.

Back to it now after I reread this post again. Option 1 and 3 will give bad results on long run. Because of the way AI works.

AI feedback loop

  1. Stack Exchange network data are already public and they are also serving it as API. That gives AI users a head start. So that data is used as training set.

  2. AI users will contribute some posts here.

  3. And they will wait for feedback!. Whatever we do with them votes, flags, comments, deletion...

This feedback will be very helpful, to continues AI learning. 1 and 3 will just levels up the AI and it will be hard with time to distinguish soon or later. 2 is not perfect but delaying the hide/deletion action will slow down the AI learning process.

May be it is wrong for me to put this in public? Anyway CM already gave some hints of SE secret scroll.