0xkydo

Other than zkSync, I don't believe anyone has made significant progress on a complete privacy stack.

The checks and balances in L1s are beautiful.

Stakers: Congressional - Control financial resources

Operators: Judicial - decides what is the network by picking version of the node software

Network Developers: Executive - Implement the changes to node software

If Hyperliquid has shown anything is that product obsession outweighs narrative in trading.

Why? Because your target customers are sophisticated traders who can smell bullshit from miles away.

The next new exchange won’t be built on retardio narrative but on real product differentiations.

At CRCL’s P/S ratio, Hyperliquid would be trading at ______

At CRCL’s P/S ratio, Ethereum would be trading at ______

At CRCL’s P/S ratio, Solana would be trading at ______

had an interesting chat with my legal friends about "credible neutrality"

here's the twist: "being maximal censorship-resistant and credibly neutral" might be the least neutral stance

why?

by prioritizing censorship resistance, you're taking a strong political stance valuing individual sovereignty over shared morality

consider this: imagine irrefutable evidence of some address's owner committing horrible acts like child trafficking

do you ban their address or not?

not censoring at the base layer is an extreme position in the tradeoff between individual rights and collective ethics

some might argue it leans towards moral relativism - by not banning them, it is suggesting no absolute right or wrong exist

i'm not entirely sold on this view, but it offers an intriguing perspective on credible neutrality

thoughts?

who has the best data platform for indepth l2 analysis?

i am looking for
- cost structure of transactions (base/prio split) across time and periods
- fee volatility
- med/avg fee by txn type (swaps/transfers)

50MB/s might not seem like much at first glance
but when you scale it up, it's mind-blowing

for 1 yr, that's 1,576,800,000 mb of data, enough to support
- 1.3 trillion uniswap swaps
- 7.8 trillion erc20 transfers

for context, visa only processed 0.233 trillion txns in 2024

تحويل مؤسسي المستهلكين إلى العملات المشفرة بهذه الحيلة البسيطة

أثناء قراءتي للأوراق البيضاء لبروتوكولات تدريب الذكاء الاصطناعي اللامركزية/على غرار السرب هذا الأسبوع، صُدمتُ بما لم يُذكر فيها: الحتمية.

يشتهر استدلال وحدة معالجة الرسومات (GPU) بعدم استقراره - إعادة ترتيب fp32 صغيرة، والذرات في التلافيف، وتخفيضات نواة الموتر، وسباقات التدفقات المتعددة - جميعها تُنتج لوجيتات مختلفة على نفس "المرور الأمامي". تعجّ الأدبيات بالحلول البديلة (أوضاع cuDNN الحتمية، ونوى قفل التذكرة، وبناءات المحرك المُجمّدة)، ومع ذلك، لا يظهر أيٌّ من ذلك في أوراق dtrain اللامعة.

ما المانع؟ إذا كان كل نظير في الشبكة الشبكية يُنتج تدرجات مختلفة قليلاً، فحظًا سعيدًا في الوصول إلى إجماع على السلسلة أو إثبات مساهمة صادقة. سترتفع تكاليف التحقق بشكل كبير، مما يُقلل من انقطاعات المنطق، وسيبدأ شعار "التدريب المُقلّل من الثقة" بأكمله في الشعور بأنه مثالي أكثر من كونه تطبيقًا.

إذن، تويتر crypto-ml: من الذي يُعالج عدم الحتمية في بيئة مُوزعة ومُنافسة؟ هل هناك أي أوراق بحثية أو مدونات أُنصح بقراءتها؟ هل هناك تشبيهات بطبقات إجماع أخرى؟ أنشر الروابط أدناه.

اكتشفتُ للتو أن EigenDA هو مزود alt-da الوحيد الذي يدعم كلاً من USDC وUSDT، تحية لـ @celo.

من الواضح الآن أن طبقة DA مهمة للمؤسسات، وأن EigenDA جاهزة لاعتماد الدفع في العالم الحقيقي!

just spent an awesome day at stanford

half the time was giving a lecture on intersubjectivity, the other half meeting folks buzzing about crypto

shoutout to @0xfishylosopher for building a solid undergrad crypto scene and competing with ai

everyone's pumped, curious, and full of passion

feeling super optimistic about stanford crypto's future!

i think your feelings on this sentence reveal a lot

if you're disgusted, you likely view crypto as a political tool first, tech second

but if you're grateful for stopping the hack, you see it as a tech tool first, political tool second or not at all

NYC is the capital of crypto

feeling optimistic about crypto today

- stablecoin legislation passed in senate
- sec taskforce doing gods work
- solana announces new consensus protocol
- coinbase fighting to keep circle from ripple

good day

Crypto needs to rely less on carbon-based intelligence and more on silicon-based intelligence

Less multisigs, more AI

Dream bigger

Crypto needs to rely less on carbon-based intelligence and more on silicon-based intelligence

Less multisigs, more AI

as xrp gets its etf approval, the prestige of having a crypto etf is mostly gone

this is net negative for ethereum and hugely positive for bitcoin

the first crypto etf and the rest of them