Iv

19 ms·

391.
by Iv 6y ago
It is hard to find a more popular actor of dissemination of knowledge than wikipedia. China is trying hard to remind the world on every occasion how committed it is to censorship.
392.
by Iv 6y ago
The invasion was not enough to create ISIS. To do that, US had to name an idiot as Iraq's governor. Enters Paul Bremer. A single decision of his made ISIS possible. He got to manage a country that just got invaded, that used to have a
393.
by Iv 6y ago
> What happens next? Better ask yourself what happened before. If you get a number, it means you chose a metric and an algorithm. That's what you get. What you do out of that metric depends on which it is. > Objectivity is inter-
394.
by Iv 6y ago
1. There is no opposition between non-deterministic self-organization and quantification. You can measure metrics of any chaotic system. 2. Morality is about the future? You must have skipped a step there. Are you assuming consequentialism?
395.
by Iv 6y ago
> It isn’t sufficient to say “X is a metric of health/ethics, if I think more of X is healthy/ethical and you think less of X is healthy/ethical. The starting point of this conversation is to say that it is neither you nor
396.
by Iv 6y ago
I just don't understand what you are trying to do. > Lets represent "health" as the identity function in Lambda calculus. Health is a function now? Why an identity function? > This is equivalent to the English sentence
397.
by Iv 6y ago
You are on a tech forum. Reductionism is pretty much the norm here. Any programmer that has toyed with chaotic system theory can imagine holist views as explained by emergent behaviors of simpler systems. > Did you purposefully ignore my
398.
by Iv 6y ago
You are correct with my positions. About a) we are not necessarily in disagreement, we just probably talk about different aspects of morality. I am personally not a believer of objective morality but find that the closest we can get to is t
399.
by Iv 6y ago
> but I am skeptical that it yields much insight It allowed economist for instance to measure that the price US society is willing to pay to avoid the loss of one random American life is about 8 million USD. I think this is a valuable in
400.
by Iv 6y ago
The atoms that compose your body are going in random directions right now, at pretty high speed, that's what create temperature. Many of your atoms go in totally opposite directions at any given time. Even drawing the line in the set o
401.
by Iv 6y ago
Philosophers: arguing that morality can be objective while failing to agree on the meaning of words... So to rephrase your objection, you think it is impossible to have an objective measurement of something in the human society? Are you arg
402.
by Iv 6y ago
I am not sure how? Quite the contrary, stating that "good is what the society decides" is a way to solve contradictions in individual choices. That's not my stance, btw, but that's a morality framework that is coherent.
403.
by Iv 6y ago
The idea is to not look at individual choices but at the aggregate of choices made and measure what mankind considers good.
404.
by Iv 6y ago
You are misconstructing my argument to create an artificial contradiction. Talk about an exercise in futility. If you see a flaw, please point it clearly.
405.
by Iv 6y ago
Why do every "philosophical" debate ends up with "let's play semantic and never address the question being actually debated?" Philosophy is about answering real, deep questions, not about trying to find smart ways t
406.
by Iv 6y ago
Not really. I am proposing a theoretical solution to a philosophical problem. You are the one jumping to imagining how it could be implemented in a biased way.
407.
by Iv 6y ago
In this context, "that can't be tricked the way you propose". If you accurately measure the value human society puts on life (that is, avoiding the bias of, e.g. considering only a part of the population that may be privilege
408.
by Iv 6y ago
Last week, I did some philosophy: I dived into the discussions of Debian-legal of what constitutes open source and licenses apply or not to deep learning model. It was some real philosophy, applied philosophy, on complicated matters, touchi
409.
by Iv 6y ago
Goodhart's law/McNamara fallacy: when you put value in a measure, you create incentives to trick it. Here is aa equivalent statement, as terse and clearer, than the comment you made above.
410.
by Iv 6y ago
I don't think so, if it is done properly. For instance you want a human life to be worth more but the only way for that to happen is to actually spend more to save human lives, which makes the value of life increase.
411.
by Iv 6y ago
I like the idea that morality is something that is built by mankind and that needs to be extracted by studies. For instance, economists have found the value at which US society estimates an American's life by seeing at which point risk
412.
by Iv 6y ago
It was back when France was still "communist" and managed to make a profit out of it :-)
413.
by Iv 6y ago
There are lots of talks about that aspect on the debian MLs as well. They argue on whether they should have machines to redo the training of models that are considered open-source, if that should be part of the "build" process. I
414.
by Iv 6y ago
Thanks :-)
415.
by Iv 6y ago
"could" but does it, actually? The breaking changes they did from 2 to 3 caused enough headaches that I think no one wants to go through that again.
416.
by Iv 6y ago
Thanks! It really gave some good insights!
417.
by Iv 6y ago
Thanks for the links below, reading these opinions took me two more hours of my time but helped me grind some thoughts! First a quick answer to your two last questions. Programs and binaries are widely recognized as copyrightable. What I am
418.
by Iv 6y ago
Finally a good news!
419.
by Iv 6y ago
Also, another back and forth in DebConf 2012 that introduces the problem and presents some real world implications: http://penta.debconf.org/dc12_schedule/events/888.en.html
420.
by Iv 6y ago
Well, I'd be happy to have someone to chat with and exchange ideas about it. I am currently digging that rabbit hole that seems to be basically uncharted waters. I would like to find a way to make true open source deep learning models.

More ›