ot

41 ms·

481.
Multilingual Embeddings – Engineering Blog – Facebook Code(code.facebook.com)
3 points by ot 9y ago|0 comments
482.
by ot 9y ago
> numberBigInt = numberBigInt.minus(1) https://github.com/geonnave/primg/blob/master/index.html#L82 Could go double as fast subtracting 2 instead of 1 (why try the even numbers).
483.
by ot 9y ago
> That is all. Did you read the article?
484.
Work at different management levels(larahogan.me)
2 points by ot 9y ago|0 comments
485.
by ot 9y ago
> We're better served with a society where communications providers are not allowed to deny service based on their customers views for the simple reason that as long as we allow them to censor, chances are someone who is clueless ab
486.
by ot 9y ago
That's quite a false equivalency. Fascism is an ideology that at its core has oppression of free speech and political opponents, white supremacy, authoritarianism, colonialism. Communism as a political theory has none of these things.
487.
Google doesn’t necessarily need innovation(medium.com)
208 points by ot 9y ago|163 comments
488.
by ot 9y ago
Reminds me of the "Evolution of a Haskell programmer": https://www.willamette.edu/~fruehr/haskell/evolution.html Make sure you don't miss the punchline, "Tenured professor". It's the
489.
Fitting larger neural networks into memory(medium.com)
8 points by ot 9y ago|0 comments
490.
by ot 9y ago
If you flip the quadrant you're also flipping the tiles. The flipped tiles may not be in your set.
491.
by ot 9y ago
> like finding larger and larger natural numbers is not the same as having a single number larger than all natural numbers Very nice analogy! > König's lemma implies that for tilings both statements are in fact equivalent. Exactl
492.
by ot 9y ago
That's what I meant by attempting to find a constructive proof. 2 doesn't work because no matter how much you shift down and to the left you'll always have a boundary. > our link proves something stronger No, it proves tha
493.
by ot 9y ago
Another Wang tiles problem I've always been fond of because of its deceptive simplicity is the following: If a finite set of tiles can tile a quadrant of the plane, it can also tile the full plane. One would guess that there is a const
494.
by ot 9y ago
It's not just refcounting, GCs also normally have some sort of bookkeeping that is updated in the background, for example a mark&sweep GC will need to mark the objects. See the comment about the object header in the OP.
495.
Genetic Study Supports Carbohydrate-Insulin Model of Obesity(medium.com)
160 points by ot 9y ago|78 comments
496.
by ot 9y ago
Yes, when you fork all the pages in the address space are referenced, and copied when modified. However if you have a GC it will go around touching objects in the child process, and there's a good chance it will end up touching a few b
497.
by ot 9y ago
DeepMind has produced a few agents playing well classic games. When the authors of this paper are saying that these games are NP-hard, they're taking some "poetic license": complexity classes are defined for problems that hav
498.
by ot 9y ago
Note, this is from 2012. The original paper authors followed up with similar proof for other games: https://arxiv.org/abs/1203.1895 .
499.
Let your code type-hint itself: introducing open source MonkeyType(engineering.instagram.com)
3 points by ot 9y ago|0 comments
500.
Research Blog: Improving End-To-End Models for Speech Recognition(research.googleblog.com)
2 points by ot 9y ago|0 comments
501.
How Affordable Urban Housing Stays Affordable(bloomberg.com)
2 points by ot 9y ago|0 comments
502.
Big company vs. startup work and pay(danluu.com)
3 points by ot 9y ago|0 comments
503.
Waymo’s fully self-driving vehicles are here(medium.com)
1 points by ot 9y ago|0 comments
504.
by ot 9y ago
> You could totally just write the FdWrapper class Which, BTW, is exactly what folly::File does :)
505.
by ot 9y ago
Maybe they got the idea from Abstruse Goose http://www.abstrusegoose.com/527
506.
by ot 9y ago
Sure, but if you look at the implementation in the article, that's a much longer dependency chain: *c = (uint32_t)(s[0] & masks[len]) << 18; *c |= (uint32_t)(s[1] & 0x3f) << 12; *c |= (uint32_t)(s[2]
507.
by ot 9y ago
> My primary — and totally arbitrary — goal was to beat the performance of Bjoern Hoehrmann’s DFA-based decoder. The thing is, that decoder could be already branchless. The core of the algorithm is: *codep = (*state != UTF8_ACCEPT)
508.
Interviewing.io hopes to close the diversity gap with anonymous interviews(techcrunch.com)
12 points by ot 9y ago|0 comments
509.
Zstandard Compression and the Application/zstd Media Type(datatracker.ietf.org)
3 points by ot 9y ago|0 comments
510.
by ot 9y ago
Similar sentiment FTA > “In just the last 48 hours, I’ve spoken to a female tech executive who was grabbed by a male C.E.O. at a large event and another female executive who was asked to interview at a venture fund because they ‘feel lik

More ›