Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
habitue
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
40 ms
·
481.
▲
by
habitue
5y ago
I rehosted the high-res versions on YouTube: https://www.youtube.com/playlist?list=PLynsFaxiujzqKd7qtSreJ...
482.
▲
Making Generative Art with Rust
(blog.abor.dev)
126 points
by
habitue
5y ago
|
13 comments
483.
▲
by
habitue
5y ago
Mozilla laid off most of the people working on Rust last year, so there is a bit of a diaspora.
484.
▲
by
habitue
5y ago
Rust has been picked up at Amazon, Google, Facebook and Microsoft. I think that gravitational pull is going to keep Rust a going concern for a long time. That being said, it's an investment in the future mostly. There aren't anywh
485.
▲
by
habitue
6y ago
Really didn't expect to be reading about literal containers on HN.
486.
▲
by
habitue
6y ago
Yeah, there is an equivalent article that could be written for Rust as the one language (which is slightly more plausible than Java)
487.
▲
Rust Notebooks with Evcxr
(blog.abor.dev)
3 points
by
habitue
6y ago
|
1 comments
488.
▲
by
habitue
6y ago
Can someone knowledgeable explain why Btrfs isn't like better than ZFS at this point in terms of stability and feature set? Is it just designed poorly, or neglected, or what? It seems like it's been many years and it's still
489.
▲
by
habitue
6y ago
This is just highly dependent on the structure of the company and the ability to objectively determine an individual's input into the final profit of the company. Small companies, sales orgs, hedge funds have an advantage here. If you&
490.
▲
by
habitue
6y ago
I think the page is just saying they are so far from being at the borrow checker stage they aren't going to worry about it initially. But it sounds like eventually they'd implement it.
491.
▲
by
habitue
6y ago
> The project is still in an early phase with the goal to compile the offical Rust test suite. There are no immediate plans for a borrow checker as this is not required to compile rust code and is the last pass in the RustC compiler. Thi
492.
▲
by
habitue
6y ago
kakoune integrates multiple cursors as a first class concept, rather than bolting it on. So editing commands that work on "the current selection" transparently work for multiple selections as well. You can see the changes live, wh
493.
▲
by
habitue
6y ago
Take this with a grain of salt, because I haven't done Kanban personally, but I think the theory is that if you hard cap work in progress, and everyone is sitting around on their hands, it puts a lot of pressure to eliminate bottleneck
494.
▲
by
habitue
6y ago
One of the most important things I think Kanban[1] emphasizes is that once you eliminate unnecessary bottlenecks, you're left with unavoidable bottlenecks, and this naturally means there will be slack. Explicitly expecting slack time i
495.
▲
by
habitue
6y ago
If you materially alter the item, in a way that would surprise the consumer expecting an item with that trademark, it can be a problem. The example legal eagle on youtube gives is a sunglass manufacturer with special protective lenses. A re
496.
▲
by
habitue
6y ago
Is the implication that Microsoft is willing to release this now that the Oracle case is resolved?
497.
▲
by
habitue
6y ago
I'd probably rephrase "trust" to "maintain appropriate confidence intervals on your beliefs" that being said, there's clearly something to trusting yourself beyond rational levels. How far would Kanye have gott
498.
▲
by
habitue
6y ago
it's the only way
499.
▲
by
habitue
6y ago
Agreed, the haskell / rust overloading where there's a loosely enforced[^1] but implied contract is probably the best of both worlds. But even without strong constraints, the language's culture is a huge aspect of whether it
500.
▲
by
habitue
6y ago
> You shouldn't trust anybody other than yourself Oh, definitely don't trust yourself. Everybody is wrong about almost everything, that includes yourself
501.
▲
by
habitue
6y ago
I believe in order to be dependent, a value needs to be able to affect the type somehow. So something like: fn make_array(n: usize) -> [u8; n] would do the trick. I am guessing this isn't actually the feature though, and th
502.
▲
by
habitue
6y ago
My point is that if we take this definition literally, then it's trivially true that no markets are efficient by this standard. It's also true there's nothing to talk about since everyone agrees on this. The interesting versi
503.
▲
by
habitue
6y ago
Just a disclaimer for people reading, `economicprinciples.org`, despite having a very official sounding name, is presenting Ray Dalio's theories, not widely held economic principles. None of it seems to me (although I am not an economi
504.
▲
by
habitue
6y ago
This is also a matter of degree. How long a market takes to fully reflect available information is a range, from "immediately" to "never"
505.
▲
by
habitue
6y ago
If anyone is interested in a rust version (probably the main benefit being you can use it as a rust library) there is a similar tool called skim[1] It's actively developed and seems to have a non-overlapping set of features with fzf ot
506.
▲
by
habitue
6y ago
By "effectively grandfathered", are you referring to their justification of bitcoin and ethereum as not securities under the Howey rule, or are you referring to some other comments? (Genuinely curious)
507.
▲
by
habitue
6y ago
South Korea backed off from just such a ban in 2018[1], the details of which might be informative if anyone is trying to determine how likely a similar ban here might be [1] https://www.pymnts.com/news/bitcoin-tracker&#
508.
▲
by
habitue
6y ago
This doesn't seem to be correct. The article talks about reinforcement learning agents optimizing a communication game to trade off color description accurately vs. effort. The words they use aren't real words, they're partit
509.
▲
by
habitue
6y ago
unsafe isn't like a taint, where things that use unsafe code must propagate their unsafety outward. It's more like saying "there are invariants that this function expects, but does not enforce, and the caller may violate thos
510.
▲
by
habitue
6y ago
I think what you're saying is plausible, but I would expect the different models to diverge at different rates in that case. So, for example, if resnet-real and resnet-ideal stayed within 1% and the other models showed a bigger range,
More ›