Kranar

72 ms·

871.
by Kranar 5y ago
Your statement is too strong and comes with important caveats. Only certain aligned memory accesses to non-floating point data types are atomic. Floating point values, unaligned accesses and even integer operations via SIMD instructions are
872.
by Kranar 5y ago
You have misread what C++/Wikipedia says and are hence incorrect. The Wikipedia article says that actions are potentially concurrent if (they run from two different threads) OR (they are unsequenced AND one operation is performed by a
873.
by Kranar 5y ago
There need not be any parallelism or simultaneity for a data race to occur. What matters is that writes occur concurrently with another read/write, not whether they occur in parallel. For example it's possible on a single core CPU
874.
by Kranar 5y ago
The precise definition of a concurrent access is with respect to a programming language's memory model. Now Python does not provide formal semantics for its memory model (neither does Rust for that matter or C++ prior to C++11), instea
875.
by Kranar 5y ago
If you have a thread decrease the value of x then of course it's possible for x to decrease in value. The point is if you have a bunch of threads where every thread only ever increases the value of x like in the article then in Python
876.
by Kranar 5y ago
The GIL helps users protect against data races, it does not protect against race conditions. In Python, due to the GIL, it is not possible to concurrently access a memory location and hence there is no concurrent access to x. There are conc
877.
by Kranar 5y ago
The definition I gave is exactly the one given by Rust [1], by C++ [2], by Java [3] and the one found on Wikipedia [4]. In safe Rust data races are not possible and yet it is still possible for a Rust program to exhibit the behavior in your
878.
by Kranar 5y ago
Your link doesn't say anything about data races, it's a link to race conditions which are not the same thing as data races.
879.
by Kranar 5y ago
The article mixes up data races and race conditions, which is understandable since it's a common mistake. Race conditions are a fairly general class of bugs involving some non-deterministic sequence of operations, but a data race is a
880.
by Kranar 5y ago
I don't see your point. I'm sure it would have been very interesting back in 15th century to investigate the impact of continued literacy among the populace and its effect on various power structures: religious, economic, and poli
881.
by Kranar 5y ago
>Unless you are counting AdSense as sponsors. Yes, I am going by Google's own definition of AdSense which describes them as "sponsored links". If revenue from sponsored links does not count as sponsorship revenue then to e
882.
by Kranar 5y ago
The interrogators are pretty slick at this, and this is clearly not their first rodeo. The way you stop someone from walking out the door is you pretend that if they cooperate with you, you will help them out and contain this issue; if they
883.
by Kranar 5y ago
From the tweet linked below [1], which sparked a lot of controversy about the definition of piracy (which I entirely disagree with Linus on), about 65% of their revenue comes directly from sponsors and 32% comes from merchandise. [1] https
884.
by Kranar 5y ago
Online accounts are not required for Windows Enterprise.
885.
by Kranar 5y ago
>Apple and Google do it, so MS will do it too. None of iOS, macOS or Android require an online account.
886.
by Kranar 5y ago
It's fine to feel that way, but it's not a particularly effective approach in my opinion and in fact quite the opposite of what I look for in a manager. A good manager manages people; they understand people's strengths, weakn
887.
by Kranar 5y ago
Can you please list which firms you are aware of that got caught front running?
888.
by Kranar 5y ago
Most people don't compare D to Lisp or other languages with excellent metaprogramming. D is almost exclusively compared to C++ and C++ went through a painful metaprogramming phase of its own in the 2000s. While C++ is not by any means
889.
by Kranar 5y ago
We do have a lot more Einstein level discoveries than most people may realize. If anything, we have so many discoveries that it's hard to single one out for special consideration. The big difference between now and then I'd argue
890.
by Kranar 5y ago
This isn't true unfortunately and there are absolutely horror stories of this happening to people. For example at some FANG companies you sign what you think is your regular employment contract and then move on down to the Valley excit
891.
by Kranar 5y ago
If the company you're interviewing for is just some boss rambling about his wife, then sure, don't bother training for it or doing everything you can to increase your chances of success. If the company you're interviewing for
892.
by Kranar 5y ago
Your friend is morally repugnant.
893.
by Kranar 5y ago
>just watch CNN, CNBC, etc Why? CNN and CNBC are viewed by around 1-2 million people during prime time. There are literally Youtubers who do nothing but eat junk food who get more views than CNN [1]. Just because something is a big deal
894.
by Kranar 5y ago
The only people who say this are those who don't work at the major tech companies. I worked at Google and Microsoft and know plenty of people who work at Facebook, AirBnb, Uber. Neither myself or anyone I know is overworked, or is unab
895.
by Kranar 5y ago
Plenty of software developers would kill to work for Amazon, sure the work conditions and prestige is not as great as the other big companies but Amazon still pays pretty darn well. The big issue is the developers Amazon wants to hire are n
896.
by Kranar 5y ago
My speculation on this is that such a failure is catastrophic enough to destroy an entire blockchain. Solana has made billionaires out of a small group of people and they'd much rather fork over the 300 million dollars to plug this hol
897.
by Kranar 5y ago
Don't see the issue, looks like they saved on taxes due to depreciation, the use of stock options to compensate employees, and an R&D tax credit. None of this is remotely sophisticated or fishy.
898.
by Kranar 5y ago
Feel free to do a deep dive on the meaning and nature of earnings estimates, and you'll find it's mostly just a service that exists because it gives the appearance of legitimacy to various financial institutions. Earnings estimate
899.
by Kranar 5y ago
While mixing 250 million dollars is near impossible in the short run, you could easily mix a million dollars a month every month for the next 10 years. Mixers aren't a full proof way to obscure funds but the FBI has outright admitted t
900.
by Kranar 5y ago
DAI is great in principle and is built on fascinating technology, but the incentives for issuing DAI are not worth the risks. DAI is only created as a means of leverage, you take your collateral, lock it into a contract and receive 1/3

More ›