Hacker News Reader

Comment by godelski | Hacker News Reader

It seems like you agree with me yet simply do not easy to. I can't do anything with that. You seek evidence that is incredibly difficult to present. History is recorded differently now than in the past yet things are the same. You're ignoring the biases of how it's recorded. Any example I give you'll say it's not enough and point out another. It's a fruitless game

  > but when I said "do homework" I didn't mean podcasts

Maybe you should judge the information on its merit rather than its medium

View on HN

Possibly. I don't think either extreme opinion is correct— i.e., both can be partially correct, hence we may agree on that. Merit is often arbitrarily attributed (and sometimes, perhaps often enough, wrongly). I don't think that's enough to say that that's "always" or "most often" the case. That was my main point. You argued that no one does something strictly alone or from scratch. I think the "from scratch" part is irrelevant and think that a single person, a very small group, or a few independently (even if at different points in time) still warrants the same individual recognition.

Neither really fits what Jürgen Schmidhuber argues for. He clearly argues against intentional non-attribution and/or ignorance. Both are evidently the case deeply in current AI research (probably also outside, but this is what OP is about). At the same time, claiming that his points are worthless because he's just pointing to other individuals that suffer from the same misattribution problem is generally a wrong take. Work concentration and results are simply non-linear and generally clustered around individuals and groups. Failing to attribute, either from ignorance or malice, is equally wrong.

Reply on news.ycombinator.com