Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
bnegreve
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
25 ms
·
211.
▲
by
bnegreve
9y ago
> Why is Portugal's liberal policy on drugs heralded so great when it's not the best of the list? The number of drug induced death has decreased over the past 15 years, and it coincide with the adoption of the liberal policy
212.
▲
by
bnegreve
9y ago
Upvote for the view! I think most people don't realize how important it is to have something to look at.
213.
▲
by
bnegreve
9y ago
Reddit is an us-based media, so this is not surprising that articles are targeted for this audience. If you believe Reddit is an universal and unbiased media, you're wrong. > Why not "including France" ? What does this lit
214.
▲
by
bnegreve
9y ago
> It is simply a syndrome of getting old. I think there is more than just this: any community/social network that grows from thousands to millions of users has to make changes in order to scale (for example, changes in the moderatio
215.
▲
by
bnegreve
9y ago
Is Tensorflow available on Spark? That sounds strange.
216.
▲
by
bnegreve
9y ago
I'd like to know more about this. Check valves [1] do just this, and they work without energy input, what is the difference ? https://en.wikipedia.org/wiki/Check_valve
217.
▲
by
bnegreve
9y ago
That would not be sustainable and would probably lead to some kind of uprising.
218.
▲
by
bnegreve
10y ago
> There are many other cache replacement policies [1] and they can outperform LRU especially The article is about CPU caches so speed and memory usage are very critical. One of the main benefit of LRU is that it only needs one extra bit
219.
▲
by
bnegreve
10y ago
Of course we should be more critical, the methodology here is far from perfect. But right now, people have much faith in interviewing, and this research suggests that this faith may not be justified. As you mentioned, this is not the only r
220.
▲
by
bnegreve
10y ago
His point is that developers can work a lot on their own project, but you can't get them to work that much on someone else's project.
221.
▲
by
bnegreve
10y ago
Usenet[1] has been around for years and is (or was?) a very sophisticated form of forums. [1] https://en.wikipedia.org/wiki/Usenet
222.
▲
by
bnegreve
10y ago
Many concepts inspired from math and used in programming do not correspond anyway. In programming 'functions' can have side effects, 'variables' can be mutated, 'integers' have upper / lower bounds etc. Al
223.
▲
by
bnegreve
10y ago
Machine learning algorithms are actually quite good at solving poorly specified problems. They learn from examples and don't require any definite problem specification. For example, recognizing people in a picture is almost impossible
224.
▲
by
bnegreve
10y ago
Shooting guns in a crowded environment may cause panic.
225.
▲
by
bnegreve
10y ago
It is not meant to scale, it's a way to take down drones without shooting in crowded areas.
226.
▲
by
bnegreve
10y ago
Yes, but not always. When the body of water is big enough, they don't need cooling towers.
227.
▲
by
bnegreve
10y ago
Yes that was my point. If we agree on that you should also agree that "artificially influencing the market" does not mean anything.
228.
▲
by
bnegreve
10y ago
You're assuming that the market is something "natural" and thus fair. GP may not agree with this. It is certainly debatable.
229.
▲
by
bnegreve
10y ago
If I read the blog post I learn that - Latex is widely adopted in hard science. - Latex is not widely used in other disciplines such as sport science . I think this matches with what most latex users think.
230.
▲
by
bnegreve
10y ago
> Or you could envision large carbon sequestration plants powered by on-site nuclear. I'm not sure I understand, you're suuggesting that we 1. Burn oil to produce energy (and CO2 as a byproduct) 2. Use the energy to do X 3
231.
▲
by
bnegreve
10y ago
We burn oil to produce cheap energy and we get CO2 as a byproduct. If you want to reverse the process you need some energy, so what would be the point of burning the oil in the first place?
232.
▲
by
bnegreve
10y ago
Still, I think GP's comment is interesting. I do agree that type name and variable/argument name can be redundant in statically typed languages, such as c++ or java. (I've never used Go) For example, int getAge(Person p
233.
▲
by
bnegreve
10y ago
Oh cool, now we know why it looks so stupid: it's because it's almost impossible to distinguish takeoff from landing if you don't exaggerate the angle.
234.
▲
by
bnegreve
10y ago
> There really is no point to wondering about the reality of time if it doesn't lead to anything we can test/influence or use to advance our understanding and so far there is none. I think this is wrong. There are some (possibl
235.
▲
by
bnegreve
10y ago
Better can mean more than just "faster" or "more responsive". This post argues that Firefox is better in that it does not lock you in in the long term. Since it is not as visible as responsiveness, it is useful to have t
236.
▲
by
bnegreve
10y ago
The number of millionaires is actually quite high in France [1]. [1] https://en.wikipedia.org/wiki/Millionaire#HNWI_population
237.
▲
by
bnegreve
10y ago
> Dictators that prosecute and murder their opponents, Sorry, this is not a discriminative feature of dictators, non-dictators also prosecute and murder their opponents. (e.g. see any government involved in a war) The difference is tha
238.
▲
by
bnegreve
10y ago
> Some of the times it's to make the product better, or better targeted. Some other times it's just for spying on the users. Let's stop complaining These kinds of complains are fine, I don't think we should stop th
239.
▲
by
bnegreve
10y ago
BTW: this reminds me of detexify[1], the tool that can recognize latex symbols from handwritten drawings. Same problem. [1] http://detexify.kirelabs.org/classify.html
240.
▲
by
bnegreve
10y ago
I think GP has raised a fair point. The landing page says: "Can a neural network learn to recognize doodles" but the game doesn't demonstrate that the ANN is able to recognize arbitrary objects. It merely shows that it'
More ›