Skip to content

Comment on Memorizing Transformersparent

Comments

We are educated, paid and motivated to push the boundaries of research, not to think about all potential fallout

That’s the whole problem that led to the introduction of these sections.

That's debatable, would an "ethics" section on the original deepfake paper have changed anything?

ML research isn't as inaccessible as genetics research, if there's something idiotic that people can do with DL, they will eventually do it. Acting as if having people add a paragraph to their paper where they "reflect" on the consequences will change anything is only showing how disconnected you are with reality.

Research is research, there shouldn't be any "forbidden knowledge", we have laws for a reason.

AboutSource Built by g1lg1l

Hackerly is an independent reader for Hacker News, built on the public HN API. Not affiliated with Y Combinator.