trishume

26 ms·

211.
by trishume 9y ago
Binary-trees is allocation bottlenecked, and Java has bump pointer allocation where Go has slow allocation. The root cause being that Go's GC is non-moving so it can't do bump pointer allocation without fragmentation, so it has to
212.
by trishume 9y ago
The trick with this kind of thing is separating malicious deliberate exclusion from just trying to finance a project by catering to people who have money to fund it. Sure it would be very nice and charitable of Musk to make sure that poorer
213.
by trishume 10y ago
Qt, as long as you back the view with a model with O(1) lookup, can handle up to INT_MAX rows perfectly fast. I expect other cross-platform GUI toolkits work similarly. Most native table/list views work on an assumption of uniform row
214.
by trishume 10y ago
I don't actually understand your comment enough to know how to respond, but I'll try to respond: > come here to share in our Ideology. That Ideology is the reason for the economy ... If you're here for the economic incenti
215.
by trishume 10y ago
BuckleScript is by Bloomberg, specifically https://github.com/bobzhang , not Jane Street. The developer of js_of_ocaml, the other JS to OCaml compiler, works at Jane Street though.
216.
Things I've Learned Doing Internships(thume.ca)
4 points by trishume 10y ago|1 comments
217.
by trishume 10y ago
Exactly. I predicted every benchmark result except for the trie one, which surprised me a ton (wut??!! but pointer chasing costs!), apparently though that was just bad benchmarking putting the only path ever tested in cache.
218.
by trishume 10y ago
So basically none of the results are remotely relevant to real use. Except maybe that unordered_map is faster on large data sets than map. Another thing I noticed is it starts with the example of a type column on a database, then proceeds t
219.
by trishume 10y ago
Shameless advertising: since the developer of Vintageous stopped maintaining it, @gerardroche and I have started a fork that we will review PRs for and occasionally work on. We've already merged a bunch of outstanding PRs, added some n
220.
by trishume 10y ago
Hypothesis: Deja Vu happens when the long term memory isn't suppressed properly during the first few days. I often find when I experience Deja Vu I think I remember doing a thing I did the day before a long time before that. It feels l
221.
by trishume 10y ago
Easy link to the relevant table from the paper: http://imgur.com/a/YRFTh
222.
by trishume 10y ago
The claim you mention about getting to know the person better in random interviews as well as "People can’t help seeing signals, even in noise." is misleading and unsurprising. People ask questions in interviews that they want to
223.
by trishume 10y ago
Can someone who actually has access to the story answer the following questions for me. Is this before or after controlling for: - field of work. Is this the whole company or just devs? - years at the company. - age, as a crude proxy for to
224.
by trishume 10y ago
Just because psychology doesn't tell us about how the brain functions doesn't mean it isn't useful. Learning about how humans behave in practice is arguably more useful. And yes tons of psych studies are flawed but the reason
225.
by trishume 10y ago
Why do you think neuroscience is a better way to learn about human behaviour (which is what psychology deals with)? Neuroscience is incredibly difficult and the brain is super complex. We have learned far far less about human behaviour thro
226.
by trishume 10y ago
This sounds an awful lot like Google's 20% time and the "hack days" done by many other companies. Very few guidelines and no consequences for doing something insufficiently useful, but it ends up with a lot of valuable ideas
227.
by trishume 10y ago
It's really easy to make punchy-sounding arguments about why a certain field isn't as important to teach, so we should be rather reluctant to go along with them. For example your argument about psychology could equally applied to
228.
by trishume 10y ago
I was talking purely about humans finding messages they don't have the schema for during the course of debugging. Computers do equally badly with JSON data and Cap'n Proto data they don't have the schema for. A fake example s
229.
by trishume 10y ago
I don't think this is as big a problem as the article suggests. There are a substantial number of tradeoffs in serialization protocols, and each application/ecosystem can choose their protocol to get the best of these tradeoffs. A
230.
by trishume 10y ago
Situation: there are now n+1 competing serialization protocols https://xkcd.com/927/
231.
by trishume 10y ago
There's a lot of things in this presentation that it would be awesome if the JVM had. But unfortunately it seems it might take a long time before even some of these features are released.
232.
by trishume 10y ago
What I want is a way to flag comments, and threads on the homepage with comments, by users I have marked as interesting. This may be an okay proxy, but I have no idea how noisy it is since I normally don't see karma. There's certa
233.
by trishume 10y ago
I think I agree with you way more than you think I do. I'm a huge fan of Vim-style-editing. I use it all the time. Every editor I've used since has been with some sort of Vim emulation. I don't think, or at least I didn'
234.
by trishume 10y ago
I'm definitely aware of this, it's why I mentioned "varying levels of caching". The Conrod imgui that I mentioned basically uses retained mode GUI data structures behind an immediate mode API through diffing for performa
235.
by trishume 10y ago
It seems that this article only really asserts that "meritocracy is a myth", without much evidence. Most of the article is just saying "these people we don't like support meritocracy, and these people we do like don&#x27
236.
by trishume 10y ago
This argument looks like you and pcwalton are arguing about different definitions of "immediate mode API". I think both of you agree with each other on object-level propositions. pcwalton seems to be presuming that part of the con
237.
by trishume 10y ago
Author here: I can tell you why I didn't. Yes it would have been faster and more stable, however then I would just complain about the lack of a bunch of features from Sublime that I like, and the terrible keybindings. I actually have m
238.
by trishume 10y ago
As the author, this is the kind of condescension of Sublime users that annoys me. I think I know vim quite well, I've been using vim bindings for 5 years now across varying editors. I know almost all Vim bindings. How about a test? Sup
239.
by trishume 10y ago
Note that you don't want to maximize customers, you want to maximize revenue. So even if hugely undercutting your competition gets you the most customers, you may still want to raise prices. If doubling your prices doesn't produce
240.
by trishume 10y ago
We even know California's population is large enough to support its own single-payer system since Canada has a slightly smaller population and a single payer system. The trick is whether the transition to a single payer system is feasi

More ›