|
Maksym Zavershynskyi
@
mzavershynskyi
San Francisco, CA
|
|
Scaling decentralized applications.
|
|
|
250
Tweetovi
|
196
Pratim
|
229
Osobe koje vas prate
|
| Tweetovi |
|
Maksym Zavershynskyi
@mzavershynskyi
|
30. sij |
|
If anyone is interested in being paid for contributing to github.com/tokio-rs/loom under guidance of @carllerche please write me. Experience contributing to Rust concurrency projects is desirable. #rustlang
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
Aliaksandr Hudzilin
@AliaksandrH
|
23. sij |
|
Who wants to build community owned service like this on @NEARProtocol ? Happy to help with coaching (both engineering and business) and funding ❤️ twitter.com/ljin18/status/…
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
genderAlgebraist
@porglezomp
|
19. sij |
|
Look at that beautiful implementation 😍 pic.twitter.com/9j7jA6qGbC
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
Evgeny Kuzyakov
@ekuzyakov
|
14. sij |
|
NEAR Protocol transactions
NEAR went through a few iterations of transaction design before settled on the current implementation.
Thread 👇
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
13. sij |
|
The number of people in the US taking Chinese language courses.
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
Aliaksandr Hudzilin
@AliaksandrH
|
12. sij |
|
Given that 7% of internet users are in North America and 50% are in Asia, and we are witnessing the birth of internet native culture as penetration is so deep, we will see outsized influence of Asian culture in the world. Korean dramas and BTS are early signs of the massive trend
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
12. sij |
|
Just be careful to not press the auto-complete button on wrong suggestions otherwise it will learn to show them at the top causing you to accidentally press them even more frequently, which will reinforce its learning even more, ad infinitum until the black hole swallows your pc.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
Actually, some syntactic sugar makes it worse. It creates a false sense of understanding the Rust code... until compiler hits you with a lifetime error which you cannot understand because you never knew lifetimes were there and rarely practiced understanding them.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
* Lack of pure Rust libraries, if you have to link a library in C/C++ that is critical for your project then you might as well write the whole thing in C/C++;
* Lifetimes and trait bounds are a huge learning barrier despite attempts to hide it under syntactic sugar.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
Irregular spikes are not solved through resharding, but through other mechanisms, e.g. infrastructure used by FANG handles irregular spikes by having backup compute resources, rerouting, etc. Regular spikes though would be taken into account by the dynamic resharding algorithm.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
With dynamic resharding that someone is an algorithm.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
2/2 The advantages of DevX when you don't need to worry about collocation greatly outweigh marginal performance improvements. In the future people will think of contract collocation as frequently as they think of thread collocation today.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
10. sij |
|
1/2 People who want to collocate contracts on the same shard in a sharded blockchain (like Eth 2.0) because it makes them slightly faster (not on @NEARProtocol though) forget how rare it is these days to intentionally write code that collocates OS threads on the same CPU core.
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
Evgeny Kuzyakov
@ekuzyakov
|
9. sij |
|
Sharded cross-contract composability at NEAR protocol.
Before I can explain it, I need to give a lot of background on sharding design at NEAR.
Thread 👇 pic.twitter.com/13ZuzriSH5
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
8. sij |
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
8. sij |
|
Full remote development of Rust in CLion would be a killer feature. Is @intellijrust planning to add one anytime soon? Is anyone in the #rustlang community interested in developing this feature as an external contributor? Please contact if you have interest or experience to do it
|
||
|
|
||
| Maksym Zavershynskyi proslijedio/la je tweet | ||
|
Evgeny Kuzyakov
@ekuzyakov
|
7. sij |
|
NEAR Protocol Usability Summary:
- Account Names instead of hashes of public keys.
- There can be multiple Access Keys per account.
- Function call limited access key.
Thread 👇
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
26. pro |
|
I agree, good code does not create recurring revenue.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
26. pro |
|
2/2 The reason why large companies have shitty code is because its quality is largely decided by the ICs, and in large companies ICs have no incentive to keep it clean and simple (the incentive is often the opposite). I think you are mistaking correlation for causality.
|
||
|
|
||
|
Maksym Zavershynskyi
@mzavershynskyi
|
26. pro |
|
1/2 Most companies have top-down decisions so they are very prone to enterprise sales strategies a.k.a golfing which Microsoft and alike are very good at.
|
||
|
|
||