Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
bumby
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
1694 ms
·
5461.
▲
by
bumby
8y ago
Thanks for the clarification on your stance and for sharing the link. Although I know a lot of people think Dawkins worships on the alter of rationality, I think some of his stances on this issue need to be taken with a grain of salt becaus
5462.
▲
by
bumby
8y ago
Where do you personally draw the line? I'm not trying to be adversarial, just trying to understand how cogent your argument is. I think the race argument is a bit reductio ad adsurdum. Why is that any different than if you decide to on
5463.
▲
by
bumby
8y ago
In your opinion, does mindfulness meditation help in regards to developing attention sufficiently to alleviate the problems mentioned by OP?
5464.
▲
by
bumby
8y ago
I think the recent congressional hearings regarding Facebook frames this argument more properly. At the heart, it's whether these platforms are to be considered content providers or publishers. I believe each has unique responsibilitie
5465.
▲
by
bumby
8y ago
Statistics applied to black-box component mishaps have a couple things going against them. 1) you need a relatively large number of failures to build a good sample of data and 2) even if you have the probabilities in place to quantify risk,
5466.
▲
by
bumby
8y ago
The issue with "unknown unknowns" is that they are uncertainties and not risks. Risks we can more easily deal with because they are quantifiable so they can be hedged against with probabilities.
5467.
▲
by
bumby
8y ago
My guess is that most people feel AI should be held to a higher standard is because we feel the need to be able to audit the system in the case of mishaps. When ML becomes a high-level black box, we may not have the confidence in how to rig
5468.
▲
by
bumby
8y ago
I believe some models do precisely that. Random forest ML as an example tallies "votes" on the outcome. I'm not sure how robustly multiple algorithms have been applied to this voting technique, but it would be an interesting
5469.
▲
by
bumby
8y ago
You illustrate the point well and maybe I'm just not well versed enough in the topic. You imply that a charity warps the free market by undercutting profit-motivated competitors but in the last statement you give the vendors the freedo
5470.
▲
by
bumby
8y ago
Your wording is interesting here. Do you mean humans need less data to feel confident or less data to make accurate predictions compared to a ML model? Sorry if I'm parsing your words too much here, just genuinely curious
5471.
▲
by
bumby
8y ago
Is this similar to the concept of pleiotropy?
5472.
▲
by
bumby
8y ago
Doesn't that imply they compound profits by either 1) raising prices and thus reducing demand for their product in the presence of non-profit maximizing suppliers or 2) creating efficiencies to increase profit margin at an equal-or-low
5473.
▲
by
bumby
8y ago
I see your point if the term profit is defined in a general sense, but with the exception of "corporations as people" caveat there's plenty of everyday examples where people are trying to maximize outcomes other than capital
5474.
▲
by
bumby
8y ago
Does it have an end goal of maximizing profits, though? Couldn't the goal of price discovery be to find the point where supply = demand, without necessarily maximizing profit? Granted, this ignores the human trait to increase consumpti
5475.
▲
by
bumby
8y ago
While I understand your point, I don't know if it could be distilled that succinctly. I have friends who have worked in European countries with universal healthcare that was sub-standard by a U.S. perspective. What I'm curious abo
5476.
▲
by
bumby
8y ago
NASA calls out preparing for this specifically in their software engineering requirements. My understanding is that this was added specifically to address bit flips from radiation effects. See section 3.7.2.f https://nodis3.gsfc.
5477.
▲
by
bumby
8y ago
I think that's because HR tends toward avoiding false positives rather than false negatives, as the latter tends to be more expensive. Therefore, they would rather miss an ideal candidate in order to avoid the risk of hiring a sub-stan