mratsim

24 ms·

241.
by mratsim 6y ago
I don't want people to create Electron in half the time ;)
242.
by mratsim 6y ago
The main driver are no allocation, thread-safety, easy to generate visualization and code and within performance 1% of hand-optimized code. Ergonomics have been slightly sacrificed in exchange but is still significantly better than writing
243.
by mratsim 6y ago
> no data races for multithreading This is simply not true. Rust borrow checker has no notion of memory models: i.e. sequential consistency, acquire, release or relaxed semantics. The Tokio team needs an additional model checker to ensur
244.
by mratsim 6y ago
You need null pointers for system programming. That said you can already declare `type MyPtr = ptr int not nil` but the compiler is still clunky on proofs and needs a lot of help. it is planned to have a much better prover in the future and
245.
by mratsim 6y ago
According to https://rust-random.github.io/book/guide-rngs.html Chacha8 has a throughput of 2GB/s and xoshiro256++ has a throughput of 8GB/s For Monte-Carlo, the RNG is definitely the bottleneck. For load bal
246.
by mratsim 6y ago
I find in practice that is very hard to keep the table up-to-date with code changes and graphics are easier to inspect than table for complex flows. I have written a state machine generator that both produces the table and the code (no allo
247.
by mratsim 6y ago
Anywhere you want reproducibility - Statistics, - Deep Learning, - Monte-Carlo simulation (finance, reinforcement learning, game AI, raytracing). - Fuzzing - Load balancing - Peer selection (if non adversarial, otherwise use a CSPRNG) Also
248.
by mratsim 6y ago
I've been diving into parallel reproducible PRNG [1] in the end I choose to have a PRNG based on xoshiro and seeding process that could take loop iteration variables [2] for reproducibility: - [1] https://github.com/mra
249.
by mratsim 6y ago
20 jobs at Coinbase, 1 at Facebook, 2 at TrueCar in the front page. Amazing diversity.
250.
by mratsim 6y ago
Working internationally with CSV holding dates is a recipe for disasters.
251.
by mratsim 6y ago
Given the current publish or perish culture, I doubt any student serious about publication can afford to waste time.
252.
by mratsim 6y ago
Unfortunately, without having done the same job, it's hard to evaluate the time, even if you pad. And for lots of complex jobs you have skeleton in the closet, meetings upon meetings, extra management to contact for approval. I'd
253.
by mratsim 6y ago
yes it does, it's called confirmation bias. Not that I endorse their claim as well.
254.
by mratsim 6y ago
Using* any language and writing non-trivial programs.
255.
by mratsim 6y ago
GCC, Clang, ICC do support computed gotos and you can fallback to ``while true`` loop + switch dispatch for others. This is what Nim does: https://nim-lang.org/docs/manual.html#pragmas-computedgoto-p...
256.
by mratsim 6y ago
In Nim the equivalent is using a `static:` block, then everything inside has normal Nim syntax but evaluated at compile-time. You can also do `const a = static(foo(x, y, z))` to force normal function to be evaluated at compile-time and stor
257.
by mratsim 6y ago
Don't forget: - Patricia-Merkle Trees proven in DC++ - DHT proven in eMule Kademlia
258.
by mratsim 6y ago
IPFS has an incentivization problem just like torrents. The main reward for torrent uploaders is fame, would that translate well in IPFS?
259.
by mratsim 6y ago
> Zig has just comptime, through which it achieves all those goals (minus some macro capabilities that it deems harmful anyway) with just one, very simple, cohesive construct. You could argue on whether you like this or not, but you can&
260.
by mratsim 6y ago
The article is both assuming knowledge of how blockchain work but also seems to stem from poor comprehension of how blockchain work. For example at no point are consensus algorithms challenge explained (protect against attacks by having the
261.
by mratsim 6y ago
People are explicitly asked to wear masks too ...
262.
by mratsim 6y ago
Nor were/are Facebook or Google.
263.
by mratsim 6y ago
Indeed, though if the authors are saying > The number field sieve algorithm is still far from being fully understood, in particular for extension fields that are so important for pairing-based cryptography. I won't say I understand
264.
by mratsim 6y ago
You can't, this is called ECDLP (Elliptic Curve Discrete Logarithm Problem) and the cornerstone of elliptic curve cryptography security. Currently the ECDLP problem can only be solved by Pollard Rho which requires exponential time thou
265.
by mratsim 6y ago
It's much better to have experts assessing other experts than laymen (say politics or judges) assessing experts (like what happens with privacy/piracy laws).
266.
by mratsim 6y ago
They are putting their business in the hands of a vendor they pay for a service. When you pay a professional for a service, you should be able to expect that the service is provided.
267.
by mratsim 6y ago
I've written Weave for the Nim programming language: - https://github.com/mratsim/weave - RFC that describes the goals: https://github.com/nim-lang/RFCs/issues/160
268.
by mratsim 6y ago
Mandatory don't trust the folklore: https://hal.inria.fr/hal-01100647/document
269.
by mratsim 6y ago
Similarly Teacher and Trainer should have the most people skills requirement and be all the way to the right. Security Analyst and Business Analyst have a high people skills requirement as well, you spend your time hovering and translating
270.
by mratsim 6y ago
I'd be skeptical of anyone trying to build regulated financial services on a technology that is still unregulated in large part of the world. Getting a financial service license is possible, getting a blockchain-based banking service r

More ›