Over the past 24 hours, the crypto market has witnessed a multitude of dynamics, ranging from macroeconomic discussions to the development of specific ecosystems. Mainstream topics focused on the leaps in small-model capabilities and the trend towards local deployment, controversies surrounding the boundaries of cooperation between AI companies and governments, and energy price volatility and risk premium reassessment triggered by unforeseen events in the Middle East. Meanwhile, breakthroughs in quantum computing have also been incorporated into discussions of long-term technological competition. In terms of ecosystem development, Ethereum is advancing its ePBS and anti-MEV roadmap, while Base is accelerating the deployment of AI Agent infrastructure and on-chain incentive quantification mechanisms. Prediction markets are intensifying competition in terms of trading volume and rule transparency, with Nasdaq signaling its entry. Perp DEX, leveraging its 24/7 trading capabilities, is strengthening its institutional advantages amidst macroeconomic volatility.

I. Mainstream Topics

1. Qwen 3.5 emerges: an open-source model running on mobile phones, beating competitors with 4 times the size.

Alibaba's Qwen team released the Qwen 3.5 small model series, covering 0.8B, 2B, 4B, and 9B parameter versions, emphasizing "more intelligent, less computation," and supporting multimodal capabilities, architecture optimization, and extended reinforcement learning training. The official statement emphasizes its focus on edge device scenarios: 4B is suitable for lightweight agent applications, while 9B's performance approaches that of larger-scale models. Developer tests show that the 6-bit quantized version of 2B can run smoothly on iPhones via MLX acceleration and supports inference on/off switching; many users interpret this as "powerful agents that can run locally," recommending deployment with tools like LM Studio. Data comparisons also indicate that 4B's performance is close to the previous generation 80B A3B, while 9B is compared to higher-parameter open-source models, and 0.8B and 2B are considered suitable for offline mobile phone operation scenarios.

The debate in overseas communities centers on whether there is a gap between the actual capabilities of the models and the hype surrounding them. Some supporters believe that the leap in performance of small models signifies that AI is truly moving towards localized and decentralized deployment, and that edge computing and personal devices will unleash unprecedented capabilities. Critics, however, point out that benchmark test scores do not equate to stable performance in complex real-world scenarios, especially in long-chain inference and complex decision-making tasks, where small models still have significant limitations.

Behind these differing perspectives lies a long-standing tension: the lack of a unified evaluation standard between parameter size, benchmark scores, and real-world application capabilities. As "small but strong" becomes the focus of the narrative, the industry both yearns to break free from the worship of scale and struggles to shake off its reliance on test scores; this structural contradiction persists.

2. Aftershocks of the Anthropic DoD Controversy: Sam Altman Shares His Principles, Packy Discusses "Hype Tax," and ChatGPT Uninstalls Surge.

Anthropic's decision to refuse cooperation with the U.S. Department of Defense has caused ongoing turmoil within the industry. Subsequently, Sam Altman released internal emails clarifying OpenAI's cooperation boundaries, including not developing unsupervised autonomous weapons, not participating in domestic surveillance targeting U.S. citizens, and not providing services to specific intelligence agencies. He also acknowledged that the previous contract disclosures were too rushed and urged against viewing Anthropic as a supply chain risk. Commentators have cited the view that Anthropic's excessive use of "nuclear weapons" analogies to elevate its own position has actually increased government vigilance, a form of "hype tax." Some analysts also point out that this controversy is not accidental but rather the result of long-term strategic communication failures, with potential costs including policy influence and lost revenue. Meanwhile, media reports indicate that after Anthropic launched its "memory transfer" tool, some users have shown a trend of migrating to Claude.

The core of the disagreement lies in how AI companies should balance national security and business ethics. Supporters of OpenAI emphasize that "responsible collaboration within a democratic framework" is the realistic path, arguing that Anthropic's position is overly idealistic and even arrogant. Supporters of Anthropic see it as a symbol of upholding privacy and principles, criticizing OpenAI for constantly adjusting its boundaries under commercial pressure. There are also relatively neutral voices who believe that while the government's response was indeed strong, the company's strategic pronouncements also lacked prudence.

This controversy has exposed the unavoidable structural tensions between AI companies and governments: when technology is given "strategic" significance, any excessive self-mythologizing or opaque contractual arrangements amplify public distrust. The industry urgently needs a clearer, more auditable ethical framework; otherwise, the gap between commercial expansion and value commitment will only widen further.

3. Details of Khamenei's assassination revealed, oil prices plummet, and geopolitical risk premium in the cryptocurrency market recedes.

The Financial Times revealed that Israel had infiltrated Tehran's traffic camera system and monitored individuals for years, ultimately leading to the precision strike against Iran's Supreme Leader. Subsequently, some analysts pointed out that the US military posture in the Gulf region is being restructured, potentially pushing Iran further into China's sphere of influence, while Israel deepens its integration with Gulf states. Others believe that in a future multipolar world, cryptocurrencies may become a new geopolitical settlement layer. Following the incident, international oil prices fell significantly, and the geopolitical risk premium previously priced into the cryptocurrency market subsided.

The crux of the debate lies in whether the so-called "geopolitical restructuring" truly exists. One side argues that this is a sign of an accelerated formation of a multipolar order, with the marginal shrinking of US military influence and a reshuffling of regional power, making the logic of crypto assets as an alternative settlement layer clearer. The other side questions the lack of direct evidence for this grand narrative, arguing that the market is more of a hedge against short-term risks than entering a new phase of "meticulous planning."

The differing interpretations reflect a deeper issue: in a highly pervasive intelligence warfare environment, any sudden action could trigger dramatic fluctuations in energy and financial markets. The frequent inclusion of crypto assets in the "geopolitical hedging" narrative itself demonstrates that the limitations of the traditional financial system in sanctions and settlements remain unresolved, and the multipolar transition remains fragile and fraught with uncertainty.

4. Quantum computing has surpassed a threshold considered impossible 30 years ago.

A lengthy article titled "SUPERPOSITIONED: Quantum Situational Awareness" points out that quantum computing will cross technological thresholds considered insurmountable decades ago by 2025, including key milestones such as the optimization of error correction algorithms. It also outlines the billions of dollars invested by governments and the private sector, the landscape of major players, and prospects its potential applications in drug discovery and complex optimization. The authors position it as a continuously updated industry draft.

The community's disagreement centers on the priority of quantum technology versus the AI ​​narrative. Some believe that quantum computing has crossed a critical threshold and will develop in parallel with AI, especially in the areas of quantum security and high-complexity computing, where it holds strategic significance. Others argue that the discussions lack actionable investment and implementation pathways, remaining largely at the level of macro-level visions.

The controversy reveals a fragmented information structure: the quantum field is highly specialized, with scarce and scattered publicly available information, leading to a lag in public understanding. In an AI-dominated media environment, quantum narratives struggle to gain sustained attention and lack a platform for systematic integration; this structural imbalance may delay cross-disciplinary collaboration.

5. Claude introduces voice mode.

Anthropic's Claude Code has launched a voice mode, rolled out to users in phases. Users can speak by holding down the spacebar, and the system will automatically transcribe the speech when released. It supports mixed voice and text input, and there are no extra charges. Transcription costs are not included in the credit limit. It is available for Pro, Team, and Enterprise plans.

Community discussions focused on functionality and scalability. Supporters argued that this seamless embedded voice input significantly reduced prompting costs and improved workflow efficiency; many users, however, wanted voice output functionality added, and even open SDKs and APIs for integration with other products.

Discussions from different perspectives all point to one trend: AI tools are shifting from pure text interaction to multimodal experiences, but the gradual release and phased deployment have also raised questions about fairness and accessibility. As features are distributed through "gray-scale releases," user experiences become stratified, and striking a balance between the pace of innovation and user inclusion has become a new challenge in product evolution.

II. Mainstream Ecosystem Dynamics

Ethereum/Base

1. Vitalik explains the Glamsterdam roadmap: ePBS decentralizes block building.

Vitalik Buterin further dissects the ePBS mechanism in the Ethereum Glamsterdam upgrade, proposing to decouple block building power from proposers and outsource it to a permissionless, open marketplace to prevent builder power from infiltrating the staking layer. The roadmap also incorporates the FOCIL mechanism, where 16 random attesters enforce the inclusion of transactions to enhance censorship resistance, and envisions expanding its reach and further weakening the role of builders through Big FOCIL. Additionally, it includes a cryptographic mempool to mitigate toxic MEVs (such as sandwich attacks), network-layer anonymization (such as privacy routing like Tor and Mixnet), and a long-term vision for distributed block building aimed at reducing the cost of transactions dependent on global state.

The community generally views ePBS as a key signal of Ethereum's strengthened decentralization and MEV resistance, believing that the path from ePBS to FOCIL and then to Big FOCIL reflects a systematic decentralization of block-building power. Some commentators argue that this evolution signifies a shift in block building from being dominated by a few participants to a more commoditized market structure; others point out that the implementation of network-layer anonymization and cryptographic mempools still requires broader client and protocol support. One perspective summarizes it as: "The evolutionary path from ePBS to FOCIL and then to Big FOCIL clearly demonstrates that Ethereum is systematically weakening the power of centralized participants until block building becomes a commodity."

This roadmap signals the continued evolution of Ethereum infrastructure toward censorship resistance and distributed architecture, but uncertainties remain regarding the complexity and practicality of the related mechanisms.

2. Venice AI became a provider of recommendation models for OpenClaw (recommendations have now been removed), accelerating the Base AI Agent ecosystem.

Erik Voorhees previously announced Venice AI as the recommended model provider for the OpenClaw framework on the Base chain, emphasizing its privacy-first local inference capabilities and recommending the use of GLM 4.6 instead of the default Llama 3.3 model; however, this recommendation has since been removed from the official documentation. Meanwhile, the Base Builder Quest concluded, attracting over 125 OpenClaw AI Agent projects and awarding 5 ETH in rewards. Projects covered areas such as prediction markets, bounty mechanisms, free token issuance tools, AI lifecycle NFT tokenization, digital clone interaction, and liquidity management.

The community generally views these actions as a signal that the Base AI Agent ecosystem is moving from the conceptual stage to the construction of a standardized technology stack, believing that the combination of privacy-preserving inference and on-chain autonomous agents is forming a replicable path. Some commentators believe this provides space for open-source collaboration and framework integration; others point out that Venice is not a completely private architecture, and data can still be read by network nodes, and the removal of recommendation information has also sparked discussions about the stability of collaboration. One comment mentioned, "Hopefully, it can be integrated into our crypto-native OpenClaw fork version."

Base is building an AI Agent infrastructure with privacy as its core, but model selection and cooperation stability still pose potential variables.

3. Brian Armstrong: Base Builder Codes quantifies the on-chain traffic contribution of each application.

Coinbase CEO Brian Armstrong introduced the Base Builder Codes mechanism (based on ERC-8021), which allows applications and protocols to quantify user referral contributions by tagging on-chain user origins with transaction suffixes. Developers need to register a unique code NFT, bind metadata and a yield address, and complete the integration through the official repository. This mechanism is set to be the sole basis for future reward programs, and projects such as Aerodrome and Moonwell have already adopted it.

The community generally interprets this move as a structural upgrade of the Base incentive mechanism, shifting from subjective judgment to on-chain verifiable data, believing that quantifying influence can provide a more transparent basis for reward allocation. Some commentators see it as an attempt towards meritocracy; others point out that the privacy issues brought about by historical contribution tracing and transaction tagging still need careful evaluation. A representative response is: "based".

Builder Codes signals a long-term direction centered on data-driven growth and incentive restructuring, but the definition of privacy costs and fairness remains to be verified.

[Market Prediction]

1. Kalshi Election Center launched, with weekly trading volume of $2.72 billion, surpassing Polymarket.

Kalshi launched an election real-time tracking hub, integrating Associated Press voting data with platform transaction information to achieve dynamic monitoring across multiple sectors. Data shows its weekly transaction volume reached $2.72 billion, surpassing Polymarket's $2.4 billion to rank first in the market. Simultaneously, Kalshi's CEO announced settlement rules for markets related to Khamenei, clarifying that settlement will be based on the final transaction price before death and all fees will be reimbursed, emphasizing avoiding direct profit from death outcomes.

The community generally views the leading trading volume as a sign of accelerated integration between traditional prediction markets and on-chain transactions, and believes that public settlement logic helps improve transparency. Some commentators criticized the previously unclear rules, potentially affecting trust; others see it as a public test under high-intensity event conditions. One comment stated, "Kalshi and Polymarket are leading the market, and a significant gap has formed between the third-place competitor and the top two."

Prediction markets are evolving toward real-time data integration and rule transparency, but the boundaries of institutional design remain uncertain.

2. According to Bloomberg, Nasdaq is attempting to enter the prediction market (binary options).

Bloomberg reports that Nasdaq plans to launch binary options products linked to the Nasdaq-100 index, using a yes/no contract structure with a price range of $0.01 to $1, to reflect market probability expectations. This will be Nasdaq's first formal foray into prediction market-related products, and still requires approval from the U.S. Securities and Exchange Commission.

The community generally views this as traditional exchanges' endorsement of the prediction market product model, believing that institutions are incorporating probabilistic trading into their compliance frameworks. Some commentators see this as a sign that prediction markets are entering the mainstream financial system; others point out that it's closer to a compliant version of the crypto platform path. One comment stated, "Nasdaq is replicating Polymarket with more complex steps."

Nasdaq's entry is an important variable in predicting the institutionalization and mainstreaming of the market, but the progress of regulatory approval still determines its final form.

【Perp DEX】

1. Bloomberg cited Hyperliquid data to demonstrate the value of 24/7 trading.

Bloomberg reported that Hyperliquid has reached a new high in open interest in traditional asset futures markets, highlighting the role of its 24/7 trading mechanism during geopolitical events, allowing traders to maintain continuous exposure to assets such as oil prices when traditional markets are closed on weekends. The report describes it as a complement to the time constraints of traditional trading for on-chain derivatives.

The community generally views this as a structural signal that crypto-native trading platforms are absorbing the demand for traditional asset trading, and believes that 24/7 trading capabilities demonstrate institutional advantages in a volatile macroeconomic environment. Some commentators point out that this may mean that 24/7 trading is becoming the new normal. A representative comment is: "As open interest (OI) of traditional asset futures on perpetual DEXs hits record highs, geopolitical volatility is also intensifying, and this is gradually becoming the new normal."

Hyperliquid's expansion reinforces the trend of perp DEX evolving into a global macro risk hedging tool, but the sustainability of its growth still depends on market conditions and the density of risk events.

Click to learn about BlockBeats' job openings.

Welcome to the official BlockBeats community:

Telegram subscription group: https://t.me/theblockbeats

Telegram group: https://t.me/BlockBeats_App

Official Twitter account: https://twitter.com/BlockBeatsAsia