@wata_orz We can probably send your prize in USDt instead of TON if it works better for you.

RT @leanprover: Paperproof: A new proof interface for Lean 4 "Paperproof will inspect how the hypotheses and goals were changing throughou…

@thought_sync Сходи в Jane Street на интервью, там как раз политика, что можно использовать любой язык программирования. И об этом еще перед каждым интервью напоминают.

@shersh_ru @_bravit Ты точно не бот?

@FakePsyho How many such 24h game contests are still alive? :(

@shersh_ru Зачем нужна работа, на которой существуют трехчасовые звонки (даже если на ней можно писать на Haskell)?

@igrekde Написал программу, которая собирает большой абсолютно белый пазл по фотке с телефона.

@rationaleks ¯\_(ツ)_/¯ *place your ad here*

@grishka11 @nikitonsky To test. But users will test, so no need, right?

@dsabdrashitov Пиши на Rust!

Converging on Simulated Annealing

Solving Jigsaw Puzzle with bare Rust. Part 2.

@FakePsyho Ah, yeah, with pow it makes sense.

@FakePsyho Wow, that was unexpected. I thought reading from memory (even from cache) should be slower than simple operation like xor.

@FakePsyho Whenever I try to use SA, it works worse than simple HC (unless it stops in local optima very often). Any secret tips how to choose SA params / make it better? Any examples of tasks where SA works much better than HC?

@lperovskaya Надеюсь количество взятых дней отпуска в году?

@lisyarus You can do "while let Some(p) = queue.pop_back() {" to avoid unwrap.

@lisyarus Then you need to "prove" to the compiler that two members are not the same element (otherwise it breaks the main contract of mut). Does split_at_mut work for you?

Thanks, @gcodingcomps for sending me a jigsaw puzzle after Google HashCode. I was too lazy to solve it manually, so I decided to write a program, which does it automatically. Check out a full story: