nly

6 ms·

91.
by nly 5mo ago
Heterogenous containers are a performance killer for all kinds of other reasons. A vector of variants still wouldn't be optimal. Better to use something like https://www.boost.org/doc/libs/latest/doc&#x2F
92.
by nly 5mo ago
The big problem with virtual calls is it hinders inlining, but if you design your API correctly that can usually be optimised away by design (put loops inside the call, don't call inside a loop)
93.
by nly 5mo ago
> everything is statically known unless you explicitly ask for it to be virtual. This is true in C++ also...
94.
by nly 5mo ago
It's still the most convenient way to implement type erasure, so no.
95.
by nly 5mo ago
Qt made this pretty easy 10 years ago
96.
by nly 5mo ago
Ironically the CGNAT at my ISP is so broken at peak times the only way I can actually use the internet is via a VPN (presumably because I then only occupy one connection tracking slot on the NAT) I'm also stuck in a 2 year ISP contract
97.
by nly 5mo ago
It is of course inconceivable that the NSA do not have the private keys for dozens of browser trusted certificate authorities That nonetheless doesn't help them unless they are doing active MITM. In order to do that they'd have to
98.
by nly 5mo ago
The important thing is to raise the question and have the discussion. By asking the question, you're not precluding the experiment.
99.
by nly 5mo ago
And then those areas become more expensive...
100.
by nly 5mo ago
To do new things no number of humans can do
101.
by nly 5mo ago
Constructors and destructors are also just function calls in C++ And you can't forget to type defer
102.
by nly 5mo ago
That's significantly more challenging for an LLM (and a human)
103.
by nly 5mo ago
"Fucking brilliant" = I'm so sick of this shit, fuck my life "absolutely fantastic" also works
104.
by nly 5mo ago
Thank you for buying their books
105.
by nly 5mo ago
It's like an amplification attack.
106.
by nly 5mo ago
Nothing you describe is recognisable to me. It just seems like you chose to work at bad places.
107.
by nly 5mo ago
The best paid engineers I know seem to be the super fast hackers who write unfathomable amounts of code in short order. Unfortunately thoughtful design and engineering doesn't get recognised
108.
by nly 5mo ago
I find the same. I can blow through the session limit on Pro in <an hour.
109.
by nly 5mo ago
Encourage forks, not PRs
110.
by nly 5mo ago
The major prop shops and market makers are all over crypto, for sure. But they're only there because these markets are poorly regulated and there's a lot of retail juice to squeeze. A friend of mine works for one of the major cryp
111.
by nly 5mo ago
Because real world assets are heavily regulated and regulation has costs. The competition is also stiff with decades of experience and network effects The truth is these crypto shops have a pretty poor reputation in the traditional finance
112.
by nly 5mo ago
Searching the upper cache lines in your binary search tree (sorted vector) for your target is unlikely to yield results. Instead you want to use the extra data in the line to shorten the search, which leads you to a B-Tree or B+tree. For 4
113.
by nly 6mo ago
It's fairly easy to be submarine rich, and fly completely below the radar. Just brush off questions about your work with vagueness. If you're not flashy, nobody will suspect you're rich
114.
by nly 6mo ago
They aren't teaching themselves anything though are they, they are basically getting AI to do their work for them.
115.
by nly 6mo ago
Top of the book? Nobody on the secondary market is investing $30bn
116.
by nly 6mo ago
The win could be only decoding the fields you actually care about, rather than all fields. It's the same for any other high performance decoding of TLV formats (FIX in finance for instance).
117.
by nly 6mo ago
18-35 is a broad range... Survey people in their 30s. That is when the decision really needs to be made, when the issue is forced.
118.
by nly 6mo ago
The main issue with dual counters is that most of the time, in low latency usecases, your consumer is ~1 message behind the producer. This means your consumer isn't getting a lot of benefit from caching the producers position. The queu
119.
by nly 6mo ago
Your statements on queues ignore the state of the art > MPSC (multiple-producer single-consumer) requires a compare-and-swap loop on the head pointer so that two producers can each reserve a contiguous slot without overlap. Martin Thomps
120.
by nly 6mo ago
There's no alternative

More ›