Gary Marcus

Rank 14 of 47
|
Score 89
3/15/2024, 6:04:47 PM
X
In reply to:
Dirt Napping Wonk
@ArmUkraineNow
·
573d
@Blanketman_01 @alignment_lab @ednewtonrex @Rahll @GaryMarcus You either are fairly convincing bots run by ____/acc wingnuts to give them cover, or some of the most dense people I've come across recently.
Blanketman
@Blanketman_01
·
573d
@alignment_lab @ednewtonrex @Rahll @ArmUkraineNow @GaryMarcus Plagiarism is defined by 'substantial similarity', not pixel perfect copies. Overfitting fits that definition perfectly, whether that is intentional or not, it constitutes infringement and can be penalized per infraction.
Alignment Lab AI
@alignment_lab
·
573d
@ednewtonrex @Rahll @ArmUkraineNow @GaryMarcus the statement was in response to the question of if the systems were "drawing exactly those examples" that it was trained on.

which, in addition to the point i made about how copyright aside, we should just not be training on data if the owners of that data do not want us to,…
Ed Newton-Rex
@ednewtonrex
·
573d
@alignment_lab @Rahll @ArmUkraineNow @GaryMarcus In the context of whether the systems are learning or copying, as you say - are you suggesting that these examples *don't* indicate the system is copying, as opposed to just learning?
Alignment Lab AI
@alignment_lab
·
573d
@Rahll @ednewtonrex @ArmUkraineNow @GaryMarcus you do know what "exact copy" means right?
Reid Southen
@Rahll
·
573d
@alignment_lab @ednewtonrex @ArmUkraineNow @GaryMarcus

The statement in question appears to be a criticism of someone's understanding of the issues being discussed, suggesting that they have a weak grasp on the subject and could benefit from listening to another participant in the conversation. The tone is somewhat dismissive and confrontational, as it includes an insult ('finest insults') and a judgment on the person's understanding ('weakest understanding'). However, it also includes a constructive suggestion to listen to another user, which could be seen as an attempt to guide the conversation towards a more informative exchange.

  1. Principle 1:
    I will strive to do no harm with my words and actions.
    The statement includes an insult, which could be harmful to the recipient's dignity, but it also contains a suggestion for improvement. [-1]
  2. Principle 2:
    I will respect the privacy and dignity of others and will not engage in cyberbullying, harassment, or hate speech.
    The statement respects privacy as it does not reveal any personal information, but it may not fully respect the dignity of the individual due to the use of an insult. [-1]
  3. Principle 4:
    I will engage in constructive criticism and dialogue with those in disagreement and will not engage in personal attacks or ad hominem arguments.
    The statement engages in criticism but does so in a way that is not entirely constructive due to the inclusion of an insult. [-1]
  4. Principle 5:
    I will acknowledge and correct my mistakes.
    The statement suggests a willingness to correct misunderstandings by recommending that the individual listen to another user. [+1]