Sabitlenmiş Tweet
vish
988 posts
vish
@0x1379
cypherpunk | developer | anarcho-capitalist ✨
Bangalore Katılım Temmuz 2012
673 Takip Edilen232 Takipçiler
My personal take: One number fitting all markets is an assumption, not a law of nature maybe.
I'd push back on 'independent of any market.' The right liquidity buffer depends on withdrawal patterns, and those are market-specific. A USDC/ETH market during a crash will see correlated mass withdrawals as lenders rush to exit. A stablecoin/stablecoin pair almost never sees that kind of demand spike. 90% utilization leaves a 10% buffer. For the first case that could be dangerously thin. For the second it might be unnecessarily conservative.
English
Right, I get the tradeoff between liquidity and capital efficiency. The article explains that well. My question is different.
The entire thesis of the Adaptive IRM is that the system discovers equilibrium on its own without governance setting parameters. But 90% is itself a governance-set parameter, and it's the most important one. The controller treats it as ground truth. So why should we trust governance to pick the right target utilization but not trust it to pick the right interest rate? What makes one parameter fundamentally different from the other? If 90% turns out to be wrong for a specific market, there's no feedback loop that corrects it. The controller will just faithfully converge to the wrong equilibrium. I hope I am clear here.
English
The aim here is to balance liquidity and utilization in the market.
If the market is close to 100% utilization for a long time, then the lenders can't exit the market easily and they would have to wait for repayments on the borrows.
If it's too low, then it means that a smaller fraction of the pool is being productive and earning yield. Essentially, you are leaving money on the table.
The 90% utilization provides a decent middle-ground for these competing factors.
English
The 90% target utilization is doing a lot of heavy lifting here. Who decided 90% is the right number and why is it hardcoded? If Morpho's whole thesis is that governance shouldn't set equilibrium rates because markets know better, why does governance get to set the single most important parameter the entire controller is built around?
English
@flipdazed Smoothing the kink doesn't fix Morpho's IRM problem. The real vulnerability is the time-dependent speed component: someone can cycle liquidity in and out, repeatedly pushing utilization above target, and the rate ratchets up with each cycle.
(1/2)
English
@0xAneri @elevenlabs @reet_batra would be a good fit for the Developer Marketing Lead. You should reach out to her.
English
The leads from my last post were incredible, so we're doing it again.
🚨Dream job alerts if you love building massive growth engines and want to define how the world discovers the next generation of AI creative tools.
We're hiring for three incredible roles at @elevenlabs
- B2B Marketing Lead. You've built GTM motions for complex creative platforms, you know how to drive massive pipeline, and you love showing the world's biggest brands how to scale production with AI.
- Consumer Music Marketing Lead. You are obsessed with the intersection of AI and culture, you've scaled consumer apps to millions of users globally, and you have the data-driven playbooks to prove it.
- Developer Marketing Lead. You deeply understand what developers actually want, know how to speak their language, and have a proven track record of turning intent into self-serve revenue through high-performing campaigns and content.
Tag who we should hire or apply on our site.
aneri@0xAneri
🚨Dream job alerts if you love creating with AI and want to define how the world discovers the best creative tools. Hiring for two incredible roles at @elevenlabs 1) Short-Form Creator. Own IG, TikTok, and YT Shorts strategy and content creation. You've scaled accounts past 50k, you know what performs, and love telling the world about how to create better with AI. 2) AI Creative Producer - Define what world-class AI-generated content looks like. You are constantly experimenting with the latest image, video, and audio models and have the portfolio to prove it. Tag who we should hire, share your best work, or apply on our site.
English
@0xWenMoon Projects like Pumpdotfun and Zora have done real damage to crypto.
Turns out tokenizing every piece of dogshit you stumble across isn’t a sustainable industry.
English
Actually disagree with this take, as a builder
I feel like somewhere along the way “tokenise everything” became a misguided goal
When I joined crypto, there was an idea of “token minimalism”.
The idea that, not everything needs to be tokenised 😅
End of the day, a token is a tool. It should in theory provide the team a valuation boost thanks to liquidity.
But if the token is hurting the projects growth, which seems to be the case for Across, then I commend them for taking that bold move since they have the cash to do it
Not everything needs to be tokenised.
I actually hope MORE teams think carefully before racing to attach a shitcoin to their project in the future.
Ignas | DeFi@DefiIgnas
I dislike this Across proposal so much: converting $ACX tokens to equity. Huge failure of crypto, especially if it becomes a new trend. It feels like a betrayal of the crypto spirit: investment access for everyone, anywhere, globally. US investors under 5M ACX need accredited investor status to participate in the equity exchange. ACX stops being freely tradable on DEXs. If they ever IPO, that liquidity goes to TradFi instead. It should be the reverse: tokenize equity. That would be progress for the industry. Yet the reasons for conversion are telling: - They believe the token is undervalued - They can't do effective BD as a DAO without a legal entity The founder said: "having a token generally hurts more than it helps." I hope other DAOs won't follow them. Especially Polymarket if they skip token altogether.
English
I found a silent inflation bug in Aave v1's interest redirection that no auditor flagged yet.
When you redirect interest across a chain of 2+ hops, the protocol gifts retroactive yield on principal that didn't exist yet.
open.substack.com/pub/zpunk/p/aa…
English
I just shipped wad·nr
a wad fixed-point math for @NoirLang/@aztecnetwork
Two problems with 18-decimal token arithmetic in Noir:
• 1000×1000 tokens = 10^42, overflows u128
• integer division truncates 1/3 → 0
github.com/merkle-groot/w…
English