Sabitlenmiş Tweet
🏴 LevS (of ether) 🏴 💙💛🇮🇱☀️🦁🗡️
9K posts

🏴 LevS (of ether) 🏴 💙💛🇮🇱☀️🦁🗡️
@levs57
recreational cryptographer
Tel Aviv, Israel Katılım Nisan 2012
609 Takip Edilen1.2K Takipçiler

@banteg You could try to selfhost - but centralized mirrors sometimes will not serve your stuff. Ipfs is gigantic failure.
English

@CPerezz19 @StatelessEth I mean currently benchmark is on qemu vm, which probably means that exec is relatively slower compared to disk...
English

@levs57 @StatelessEth No there's no commit whatsoever. I just benchmarked with the EIP one.
The plan once the entire thing is quite final is to also showcase hasher diffs in performance and MGas/s.
Although except for poseidon I agree would be the same more or less (maybe a small improvement for B3
English

@StatelessEth has the next strawman goal in the scope already: Binary Trees for Ethereum.
Our duty is to prove they are performant too, not just a nice tool to enable ZKEVMs and smaller proofs.
Thus, here's the 1st part of our work: ethresear.ch/t/the-path-tow…
English

@CPerezz19 @StatelessEth And sha2 will be the same as blake3 or slightly worse
English

@CPerezz19 @StatelessEth Ok, just want to know there is no hard commit to this particular hash.
(I suspect I can improve hashcaster significantly, and on the other hand, my blake3 ideas while outperform sota are still worse)
English

@levs57 @StatelessEth I have more on that here: cperezz.github.io/pbt-spec/
But to answer: We get sequential reads in stems. But overall I agree that the random-access-pattern is never your friend when reading.
Nevertheless, wit hall improvements we have so far (not merged yet), We are good (100Mgas/s)+
English

@StatelessEth @CPerezz19 If keccak turns out better from provability perspective, are there any downsides still? (I am currently investigating efficient batch hash proving systems and I suspect keccak will be the winner due to uniformly boolean ops, ie no add mod 2**32).
English

@levs57 @CPerezz19 Sha256 isn't necessary. Blake3 is definitely an option we are considering.
English

@CPerezz19 @StatelessEth Also, another dumb question: if sequential reads are 40x more efficient, maybe random access design is bad and we actually *should* colocate contract data?..
English

@CPerezz19 @StatelessEth Hi. Is using sha2 absolutely necessary? I think from provability standpoint both blake and keccak would fare better...
English
🏴 LevS (of ether) 🏴 💙💛🇮🇱☀️🦁🗡️ retweetledi

@malekanoms In fact, mental model probably is:
diverse validators => more censorship resistance
diverse nodes => more liveness in case of major internet distruption
diverse wallets / light clients => more resist to enforced hardfork
English

@malekanoms Vice versa, "permissionless" blockchains can very well pivot if most of their users use wallets that switch to new chain version (think collusion of rpcs + wallets + major exchanges).
Validators being permissionless does not stop it, at all...
English

@lucidzk :p
Honestly, I dislike few parts about new EF mandate, mainly in the "we are not a certifying body" part, I honestly think they could do some advisory.
But it seems they have finally realized that wallet infra / privacy integration need their active involvement; that's good.
English

idk man ~60% of this industry still settles around those “kids in the sandbox” called ethereum
so if you’re betting against that, you’re kinda just betting against crypto itself
ergo i’ll take the kids stubbornly iterating toward decentralization, fundamental human privacy, permissionlessness, and credible neutrality with a network spirituality mandate
over whatever the shiny permissionless non-permissionless corpo-run re-building not re-building finance is currently being shoved in my face
milady
Ig Mos@IgMosqueira
The point folks have a really hard time grasping is that the entire @VitalikButerin cypherpunk ethos was kids playing in the sandbox thinking they are running the universe. x.com/lucidzk/status…
English

@tcoratger I think they need agg. scheme that hides full pubkey and salt (pubkey can be hidden in snark and only address being public input, salt is nonce so it is derived from address). Correct?
English

@Roniakia Не, я понимаю теорию, но она мне кажется оч сильно базируется на том, что этот вариант зайдёт усреднённому конечному пользователю.
Русский

@levs57 Банально придем к ситуации когда у тебя не будет ГПУ на рынке у которой будет достаточно блоков чтобы старые игры нормально рендерить ибо они под ии рендер заточены. Хуанг обещает это, амд подтянется как обычно попозже. Пропихнут это добро за пару поколений консолей
Русский

Большое спасибо за релиз реально успешной батлы которая всем понравилась, а теперь мы запустим новый сезон в ней который будет хуйней и бонусом разгоним всех кто смог создать этот успех. Ну а че
Knoebel@Knoebelbroet
EA has laid off an unknown number of individuals from across its Battlefield teams, including workers at Criterion, Dice, Ripple Effect, and Motive Studios. 'Battlefield 6' launched back in October 2025 and was the best selling game of 2025 in the US. ign.com/articles/ea-la…
Русский

@Roniakia Я бы в эту хуйню играть не стал, пойду лучше третьих геройчиков запущу.
(но в целом вот как раз этого пока indication-а нет, всё как будто универсально; так-то то что ты говоришь весьма логично)
Русский






