Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
kohlerm
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
11 ms
·
61.
▲
by
kohlerm
4y ago
Maybe chicken meat is just easier to replicate with plants? I was really surprised about how good https://www.eatplanted.com/ is (I am not a vegan).
62.
▲
by
kohlerm
4y ago
I inclined to believe this is all non-sense. Did ever someone hear a difference or compare hashes and got sometimes an error and sometimes not?
63.
▲
by
kohlerm
4y ago
pretty standard in german households
64.
▲
by
kohlerm
4y ago
That probably very much depends how often you open it. If you open the vertical fridge often then the difference might be larger. Vertical fridges in supermarkets are a crime IMHO.
65.
▲
by
kohlerm
4y ago
A 1% slowdown is a ridiculous low number IMHO. I would argue that almost no program is within 1% of the optimal speed (however you define it). Therefore practically it does not matter. If you think it matters then the consequence would be t
66.
▲
by
kohlerm
4y ago
Not sure I agree with this . A dependency is a dependency and you should always think about whether you want that dependency. Whether it is a framework or not is not really the important point. The important point is how much does this dep
67.
▲
by
kohlerm
4y ago
They claim that a "Join" of APIs is possible. I wonder how that would work in the general case and what the performance characteristic are? Any experiences?
68.
▲
by
kohlerm
4y ago
Or the zippers slider breaking into 2 parts on your relatively expensive Ski Jacket in a remote location at the beginning of the vacation. Happened to me, luckily I had a (cheap) fleece jacket with me with the same zipper size and I was abl
69.
▲
by
kohlerm
4y ago
NO
70.
▲
by
kohlerm
4y ago
Chrome devtools comes with memory analysis tools. https://developer.chrome.com/docs/devtools/memory-problems/a... Some of these (heap dumps) can also be used with nodejs. That being said, from my experience m
71.
▲
by
kohlerm
4y ago
Thank you Trump, you idi..
72.
▲
by
kohlerm
4y ago
This article starts with the wrong assumption that programs/computers need to be deterministic. They are clearly not. Monte Carlo tree search is no t really deterministic, nor is the training if large neural networks. To me conscious h
73.
▲
by
kohlerm
4y ago
I use a Canon 600d with a 50mm 1.4 lens.people say it looks professional
74.
▲
by
kohlerm
4y ago
Do I miss something,but isn't the real solution for this to elect using proportional representation?
75.
▲
by
kohlerm
5y ago
for a 10kg foldable bike it is certainly affordable. Most of them (if not all) are more expensive
76.
▲
by
kohlerm
5y ago
Sure Brompton has the most experience with these kind of bikes, but a steel frame on a bike that is supposed to be carried around? If I would spend that amount of money I would probably go for https://www.r-m.de/de/bike
77.
▲
by
kohlerm
5y ago
There are keyboards where you have (programmable) layers. Moving your hand is very inefficient compared to just pressing a layer key with then one hand and then using keys in the home row (or WASD style) as cursor keys. Similar for numpad
78.
▲
by
kohlerm
5y ago
It is the normal way to go. First mechanical, then columnar staggered (split) with at least 60% of the "normal" keys. Then towards 40% split boards :-)
79.
▲
by
kohlerm
5y ago
I am German and I use the US (Colemak) layout,because the German layout is very annoying for programming. I rarely use my laptops keyboard. I would rather carry a smaller external keyboard around (Atreus or similar) if I would need to use L
80.
▲
by
kohlerm
5y ago
There is no point running something like Hadoop or Spark or Flink if your problem is not bottlenecked by the resources of of a single machine. E.g. how would this solution work if the data is so large or the problem requires more context (
81.
▲
by
kohlerm
5y ago
Apache Flink is an alternative with "real" (not micro batching) streaming support.
82.
▲
by
kohlerm
5y ago
Haha nice. Hard to play on my QMK based keyboard, probably because I mapped space to be ctrl if hold and that causes some latency
83.
▲
by
kohlerm
5y ago
It is a shame this is based on X1 cores not X2 cores. Feels like they are trying to rip off people.
84.
▲
by
kohlerm
5y ago
Of fine running alone might not help you to lose weight. Also there might be a feedback loop to not eat too much crap for a serious runner. E.g. running with less weight is just more fun :-) But does that mean other activities where you gai
85.
▲
by
kohlerm
5y ago
Not sure I understand what the message of the article is. Sure not everything companies with big scalability requirements do, makes sense for smaller projects. Nevertheless some principles still make sense. E.g. a mono-repo can make sense
86.
▲
by
kohlerm
5y ago
They are potentially. If your "problem" is to decide between a Ferrari and a Porsche, you might go for Porsche because of reliability.
87.
▲
by
kohlerm
5y ago
The irony is that people used to complain that JVM applications are memory hogs, but there's still the trend to use JS Web based Frameworks(Electron etc) for Desktop applications. For the JVM memory leaks and high memory usage are a pr
88.
▲
by
kohlerm
5y ago
Same issue in Java at least for older JDKs. We used to have special (SAP) internal commands to detect this issue in the Eclipse Memory Analyzer.
89.
▲
by
kohlerm
5y ago
Great video, saw it yesterday, need to see the second one ASAP
90.
▲
by
kohlerm
5y ago
monorepo culture, everyone needs to use the latest and "greatest"
More ›