avatar
Haotian | CryptoInsight
41,221 Twitter followers
Follow
独立研究员| Researcher | 以技术和商业视角解读区块链前沿科技 | ZK、AI Agent、DePIN ,etc | 硬核科普 | Previously：@ambergroup_io | @peckshield | DMs for Collab ｜ 社群只对Substack订阅会员开放
Posts
avatar
Haotian | CryptoInsight
Thread
Over the past two days, discussions around @Solana's 10W TPS have resurfaced, as @cavemanloverboy indeed achieved over 100,000 TPS on the Solana mainnet. However, most people haven't understood the significance behind this data: 1) First, Cavey's experiment was essentially an extreme test under "ideal conditions". This means this is not the normal performance of the Solana mainnet, and differs from testnet laboratory data, but not by much. Because he used a noop (no operation) test program, which, as the name suggests, only performs basic signature verification and returns success without executing any computation, changing any account state, or calling other programs, with each transaction only 200 bytes, far below the normal transaction's 1kb+. This means the 10W TPS test was calculated in a non-normal transaction environment, testing Solana's network and consensus layer's ultimate throughput, rather than the actual processing capacity of the application layer. 2) Another key to the success of this experiment was the Frankendancer validator client. Simply put, Frankendancer is a "hybrid test version" of the Firedancer validator being developed by Jump Crypto - grafting Firedancer's completed high-performance components onto the existing Solana validator. It essentially reconstructs Solana's node system using Wall Street's high-frequency trading technology stack, achieving performance improvements through fine-grained memory management, custom thread scheduling, and other underlying optimizations. Just by replacing partial components, a 3-5 times performance boost can be achieved. 3) This test experiment shows that Solana can achieve 10W+ TPS under ideal conditions, so why is the daily TPS only 3000-4000? Briefly, there are approximately three reasons: 1. Solana's PoH consensus mechanism requires Validators to continuously vote to maintain it, and these voting transactions occupy over 70% of block space, narrowing the performance channel for normal transactions; 2. Solana's ecosystem activities often involve massive state competition behaviors, such as Minting new Non-Fungible Tokens or new MEME releases, which might have thousands of transactions competing for write permissions to the same account, leading to a high proportion of failed transactions; 3. Arbitrage bots in the Solana ecosystem may send numerous invalid transactions to capture MEV benefits, resulting in resource waste. 4) However, the upcoming full deployment of Firedancer and the Alpenglow consensus upgrade will systematically solve these issues. One key point of the Alpenglow consensus upgrade is moving voting transactions off-chain, effectively releasing 70% space for normal transactions, while reducing confirmation time to 150 milliseconds, making Solana's DEX experience infinitely close to CEX. Additionally, enabling a local fee market can avoid the awkward situation of network congestion caused by a single program's FOMO. Besides performance optimization, the benefit of Firedancer is primarily achieving client diversity, allowing Solana to have multiple clients like Ethereum's Geth and Nethermind, directly improving decentralization and single node failure resistance. Therefore, for those in the know, the Solana 10W TPS discussion is actually about confidence in Solana's future client and consensus protocol upgrades. Those unfamiliar attempt to gain visibility through a TPS arms race (although TPS comparison is already outdated). However, understanding the experiment's meaning is quite insightful. A popular science explanation, shared with everyone.
GETH
0%
avatar
Haotian | CryptoInsight
$PLUME's upcoming listing on Binance has once again brought market focus to this RWAFi newcomer. From a TVL of over $330 million in two months to leading with 177,000+ RWA holders, has @plumenetwork used data to reframe the narrative for the RWAFi track? 1) At first glance, 177,000 RWA holders and $330 million might seem abstract, but if we consider that these 177,000 RWA holders exceed the total RWA holders of ETH and Solana, and the $330 million TVL is backed by over 90% capital utilization rate, it might give a new perspective on this "top player" in the RWAFi frontline; 2) If the previous bull market was dominated by "public chain narrative", this round has clearly shifted to "yield narrative". The yield triangle formed by Plume, Ethena, and Ondo represents three different paths: @ethena_labs focuses on synthetic dollar stable yields, @OndoFinance concentrates on government bonds and traditional fixed income, while Plume chose a hybrid yield - combining TradFi returns and DeFi strategies. This approach combines the stability of TradFi government bonds with DeFi's aggressive circular leverage, with a real-time circulation of over $40 million in RWA funds supporting its model; 3) Even more intriguing is how giants are participating. Traditional financial behemoths like Apollo, Galaxy, and Fidelity are not simply investing and waiting for returns, but directly issuing and deploying assets on Plume. This "shareholder and customer" deep binding essentially solves RWA's most challenging cold start problem - having assets, liquidity, and endorsement, a trinity of sorts. Of course, challenges in the RWAFi track remain, primarily regulatory uncertainty and traditional institutional players' acceptance. Plume has chosen a pragmatic path to first solidify "capital efficiency" and "real yield" - the key concerns of traditional institutional players, using a novel yield narrative to prove the value of underlying infrastructure, which is perhaps more convincing than any grand narrative or vision.
ETH
2.8%
avatar
Haotian | CryptoInsight
08-14
Here's the translation: Regarding the dedicated chains launched by Strip, Circle, and Tether, here are two perspectives: 1) Impact on Ethereum layer 2: Layer 2 solutions have been consistently trying to inherit mainnet security safely, yet they overlook a fact that for major clients like Strip, Circle, and Tether, the core demand for Mass Adoption is not decentralized security, but full-stack control from minting to settlement. Moreover, the commercial interests such as Sequencer revenue, MEV, and gas fees that can be directly pocketed have no reason to be shared with L2. More critically, when facing regulatory inquiries or "compliance" issues requiring urgent handling, dedicated chains can obviously meet TradFi risk control requirements more quickly and efficiently. Therefore, this is absolutely another blow to the Ethereum layer 2 strategy. L2 originally hoped to introduce real users and transaction volume through stablecoins and RWA assets, but these asset issuers are directly bypassing them. Ironically, the more "orthodox" L2 is technically, the less commercially attractive it becomes, as these technical innovations seemingly solve Ethereum community concerns but not stablecoin issuers' pain points. 2) Impact on Ethereum mainnet: The impact on Ethereum mainnet depends on the perspective. In my view, stablecoin giants creating dedicated chains are essentially building an efficient payment settlement layer, which precisely confirms Ethereum's position as a global financial settlement layer. These dedicated chains can indeed optimize point-to-point payment throughput and latency, but they lack true interoperability. When complex cross-asset financial operations are involved, the required atomicity and composability can only be achieved in Ethereum's unified state machine. Crucially, DeFi derivatives market innovation depends on permissionless liquidity aggregation. For instance, Uniswap V4's Hook mechanism, Aave's cross-pool risk management, GMX's synthetic asset model, etc., all require access to multi-source liquidity, which cannot generate synergy on closed stablecoin chains and naturally cannot showcase the innovative charm of permissionless DeFi infrastructure. Therefore, Ethereum will ultimately play a dual role: both a neutral settlement layer between these dedicated chains (similar to SWIFT's clearing function) and the foundational layer for DeFi innovation (providing composability for complex financial products).
UNI
2.9%
avatar
Haotian | CryptoInsight
08-13
Thread
<Monero> $XMR, a privacy coin project with a market value of $6 billion, was actually 51% hash rate attacked by a small project @_Qubic_ with a market value of only $300 million? WTF, it's not because of superior technology, but because the situation is so absurd. Let me explain: —— Who is Qubic? Before telling this magical story, let's first understand what Qubic is. Qubic's founder is Sergey Ivancheglo, known in the circle as Come-from-Beyond, a technical madman - he created the first PoS blockchain NXT and the first DAG architecture IOTA. In 2022, Qubic's mainnet went live, claiming to do three things: build a super-fast chain with 15.5 million transactions per second (2,000 times faster than Visa), transform mining hash rate into AI training power, and ultimately achieve AGI (Artificial General Intelligence) by 2027 (even OpenAI wouldn't dare to claim this). Sounds fantastical and outrageous, right? Why such ambition? Everyone knows traditional POW mining is criticized for wasting electricity, as most mining mechanisms involve miners solving mathematical problems with power-consuming equipment to grab reward blocks, essentially wasting hash rate to exchange for rewards. Qubic's new consensus is Useful Proof of Work (UPoW), allowing miners to mine for POW chains while training their AI system AIGarth under the dispatcher's coordination, essentially earning two incomes from one hash rate. This is why they could easily buy out Monero's miners, as the rewards once reached three times the direct XMR mining income. Think about it, with miners able to eat multiple fish, what "loyalty" exists in the face of "interests"? Alright, the underlying logic of Monero being attacked by this "vampire" has been explained with zero technical complexity. —— A science popularization: Why Monero, not Bitcoin? The answer lies in the difference in mining methods. Bitcoin uses ASIC miners, specialized machines only for mining BTC, capable of solving SHA-256 mathematical problems or mining similar algorithmic coins. However, BTC mining hash rate competition is so intense that miners are stretched thin (24/7 operation), and using ASIC for AI training is impossible. Monero is different, using the RandomX algorithm that allows mining with general CPUs, meaning miners can mine today, train AI tomorrow, and render videos the day after, truly multitasking. Qubic's brilliance is targeting CPU miners, enabling "dual-use" machines, thus leading to this 51% hash rate attack or control event. Bitcoin's moat remains stable, with miners locked to ASIC miners, only able to guard Bitcoin; —— Hash rate becomes mercenaries How terrible are the consequences? It tears off the last fig leaf of some POW chains, because we always say "hash rate" is a chain's moat, with more hash rate meaning more security. But Qubic's eye-opening experiment tells us: for CPU/GPU mining coins, hash rate is just mercenaries, following whoever pays more. More interestingly, after proving they could take down Monero, Qubic voluntarily withdrew. Why? Afraid of completely crashing Monero and affecting their own income. Because a significant part of the 3x income still comes from mining XMR, with $QUBIC only providing additional token rewards. If Monero crashes, Qubic can't escape either. Better to withdraw gracefully, create a spectacular marketing event, and humiliate POW's staunch followers - isn't this "I can kill you but won't" feeling as unbridled as their AGI slogan? [The translation continues in this manner, maintaining the original tone and technical nuances while translating to English.]
ETC
3.13%
avatar
Haotian | CryptoInsight
08-12
Thread
Here's the translation: Recently, compared to the steady upward movement of $ETH, $SOL's performance seems slightly lackluster. $4,300 vs $175, what mystery lies behind this price difference? In my personal understanding, it's fundamentally a covert battle about "who is the institutional darling": 1) ETH has obtained its "pass" into the traditional financial world - after ETF approval, cumulative net inflows have exceeded $10B, allowing off-market funds to enter compliantly, which is equivalent to opening a formal door for institutions. While SOL's ETF application remains uncertain, the current reality is that it lacks funding channels, directly impacting price performance. Of course, this can also be interpreted as SOL having room for catch-up, as its ETF is not entirely impossible, just requiring more time to go through compliance procedures. Crucially, under the purchasing power of companies like SharpLink and BitMine, ETH's MicroStrategy has already demonstrated a certain institutional FOMO effect, which will drive more enterprise Treasury fund allocations, creating massive Wall Street off-market funding momentum; 2) Currently, the stablecoin scale difference between ETH and SOL is still stark, with data showing 137B vs 11B. Everyone might wonder, with both having American blue-chip genes and on-chain Nasdaq characteristics, why has SOL fallen so far behind in this stablecoin war guided by U.S. stablecoin policies? Actually, it's not SOL's fault. It's an ultimate test of chain infrastructure's decentralization, security, and liquidity depth. On Ethereum, USDC (65.5B), USDT, and DAI firmly control the stablecoin market, reflecting absolute trust from institutions like Circle and Tether; Although SOL's VCs are all American capital, these new Wall Street institutional buyers might not consider too much, simply looking at the realistic data gap, which SOL might find hard to bridge short-term. However, objectively, SOL's stablecoin growth is actually not bad, with PayPal's PYUSD also choosing to focus on Solana, providing considerable imagination space, just requiring patience; 3) Once, SOL's on-chain economic vitality was explosive, with PumpFun's daily trading volume breaking millions, and various MEME on-chain memes flying everywhere. However, the issue is that we're currently in an institutional chip accumulation period, where large funds care more about compliance channels, liquidity depth, and safety records - "hard indicators" - rather than how many MEMEs are in PVP; In other words, it's not yet the narrative cycle dominated by retail investors. Conversely, this on-chain vitality is precisely SOL's differentiating advantage. When the market cycle shifts and retail FOMO reignites, SOL's accumulated innovative gameplay and user base might become the next market's trigger point; 4) As SBF's "favorite child", SOL might still be experiencing FTX collapse effects, with the tragic fall from $260 to $8 still vivid. Although technically SOL has completely independent, in institutional memory, this association is like a scar, occasionally brought up; Moreover, rising from $8 to $175 itself proves SOL ecosystem's resilience. Those teams building during the darkest moments became new forces reconstructing SOL's public chain, and this phoenix-like rebirth experience might be beneficial long-term; 5) ETH follows a layer2 segmentation route, criticized for liquidity fragmentation, but precisely meeting institutional risk isolation needs. SOL's integrated high-performance route, with everything running on one chain, appears as concentrated risk in institutional eyes. Thus, Robinhood partnering with Arbitrum is evidence. From an institutional perspective, ETH's high gas weakness becomes an advantage for filtering high-value transactions, contrary to Mass Adoption, but the current main theme isn't Mass Adoption - it's who can win institutional favor; 6) Lastly, there's a difference in time consensus accumulation. ETH has 9 years of history, SOL only 4. Although native projects like Jupiter and Jito have shown world-class product capabilities, compared to DeFi giants like Uniswap, Aave, and MakerDAO, there's a gap in market education, ecosystem sedimentation, and trust accumulation. In summary, the pain of E-guards might breed S-guards in a new market FOMO, but this contest is essentially a periodic mismatch between institutional and retail narratives. After all, ETH wasn't built in a day, and SOL's growth speed is actually already quite remarkable.
UNI
2.9%
avatar
Haotian | CryptoInsight
08-08
Recently on the YouTube channel "The Rollup", @TrustWallet CEO Eowyn Chen and @OpenledgerHQ core contributor Ram Kumar had an in-depth discussion about their collaboration. Here are some valuable insights: 1) Throwing Cold Water on the "Fat Wallet" Theory During the interview, Andy mentioned the popular "fat wallet" theory - that wallets with user registration channels can vertically integrate various services. However, Eowyn Chen's response was interesting. She frankly stated that consumer retail user business is actually very difficult, involving extensive customer support, higher security responsibilities, and frequent product route adjustments. Many people see Trust Wallet's 200 million downloads and think wallet business is lucrative, but the CEO herself emphasizes the pain of serving retail users. This suggests that a wallet's "fatness" isn't achieved simply by wanting it, and while user relationships are valuable, maintenance costs are also high. This perspective realistically explains the current situation of many wallet service providers. More critically, she mentioned that not all value is concentrated at the front end, and value chains should develop fairly across all parts. This view somewhat dampens the "fat wallet" theory and explains why Trust Wallet is willing to collaborate with infrastructure projects like OpenLedger. 2) Has the Inflection Point for Specialized AI Arrived? Ram Kumar's assessment of AI development path is worth noting. He believes AI is evolving from generality to specialization, similar to how Google derived vertical applications like LinkedIn and YouTube from general search. ChatGPT-like general AI will be like an operating system, with more specialized models for specific use cases emerging in the future. This aligns with my previous analysis of web3 AI industry trend evolution. Trust Wallet discovered that general models cannot solve specific problems in the crypto domain, which precisely confirms this trend. Importantly, building specialized AI models requires high-quality vertical domain data, which is exactly what OpenLedger aims to solve. 3) The Dilemma of "Unpaid Labor" in Data Contribution Ram Kumar bluntly criticized AI as a "trillion-dollar industry built on unpaid labor". AI companies train models by scraping internet data, yet data contributors receive no compensation, which is indeed a structural problem. OpenLedger's solution is to allow data contributors to receive long-term profit sharing from AI models, rather than selling data one-time. Combined with the wallet's global payment capabilities, this theoretically enables frictionless cross-border value distribution. However, a core question remains: how to ensure data quality? Ram himself acknowledges that 90% of open-source contributions on platforms like Hugging Face are useless. If contributed data itself has limited value, even the best incentive mechanism becomes futile. Eowyn Chen used the "gun rights" analogy for self-custody, emphasizing that AI functions are optional, allowing users to choose between convenience and security. This product philosophy is correct, but clearly presenting options requires strong product design skills. Ram also made an interesting observation: crypto wallets might be the only way for users to receive data contribution compensation globally. This suggests wallets may evolve from mere asset management tools to fundamental infrastructure for digital identity and value distribution. Note: To learn more, visit The Rollup's YouTube channel to watch this interview. twitter.com/tmel0211/status/19...
loading indicator
Loading..