WOOFUN · AI Crypto Intelligence

Woofun AI reports that Bank of America CEO Brian Moynihan has joined other Wall Street executives in raising concerns regarding Anthropic's "Mythos" AI model. Moynihan noted the significant shift in workload and the rapid pace at which these tools can expose system vulnerabilities, necessitating swift responses. Bank of America is among the institutions authorized to use Mythos for internal system testing and peer information sharing. The model remains unavailable to the public, with JPMorgan Chase CEO Jamie Dimon previously warning that public access would be as dangerous as "giving a ballistic missile to an individual."

Woofun AI reports that Alpaca has completed a $135 million funding round dedicated to expanding its infrastructure for tokenized stocks, aiming to bridge traditional U.S. equities with blockchain settlement systems. The firm currently custodies or clears approximately 94% of all tokenized U.S. stocks and holds over $1.5 billion in underlying equities on behalf of its partners. Despite the tokenization process, regulated entities must still retain the underlying assets and manage corporate actions, a requirement that continues to serve as a primary constraint in the market.

Woofun AI reports that BASED has unveiled Based Alpha, a token issuance platform operating on the Robinhood Chain. The protocol employs a constant product bonding curve mechanism with a fixed total supply of 1 billion tokens. Of this amount, 793.1 million tokens are available for public sale to facilitate price discovery. Upon exhaustion of the bonding curve sales, the remaining 206.9 million tokens will automatically pair with accumulated funds to establish a locked liquidity pool on the Based DEX.During the bonding period, 0.3% of transaction fees are allocated directly to token creators, while the remainder stays within the BASED protocol. Once fees reach a specified threshold, a distribution system activates to share profits with BASED holders. BASED indicated plans to introduce additional features and trader incentive programs in the future. Previously, the platform enabled cross-chain asset exchanges within the Robinhood ecosystem.

Woofun AI data shows that US-listed memory chip equities are undergoing significant downward adjustments. At the time of reporting, Micron Technology declined 5.48%, Western Digital dropped 8.03%, and Seagate Technology fell 8.02%. SanDisk decreased by 9.70%, Kioxia ADR slid 11.28%, and SK Hynix retreated 8.46%.Calculations indicate that several memory chip stocks have surpassed a 40% correction range in this cycle. Western Digital has corrected by 41%, SK Hynix by 38.3%, and Kioxia ADR by 47.9%.

Woofun AI reports that the Financial Action Task Force (FATF) published its seventh special update on virtual asset regulation implementation. The document indicates that while legislative adoption is increasing, enforcement lags, allowing criminal groups to move billions in illicit funds. Data shows 83% of surveyed jurisdictions have enacted "travel rule" provisions, rising from 73% in 2025, with 11 others working on implementation.The report highlights persistent challenges in identifying Virtual Asset Service Providers (VASPs) and managing risks from offshore entities and DeFi platforms. Notable trends include the abuse of stablecoins by North Korea-linked actors and terrorist financiers, who are developing proprietary coins to evade asset freezes. Cited cases include a Cambodian group laundering $4 billion between 2021 and 2025, and Spain’s National Guard dismantling a €460 million crypto fraud network in June 2025. FATF Chair Gilles Thomson emphasized the need for strengthened public-private cooperation to close regulatory gaps.

Woofun AI reports that Democratic Senator Elizabeth Warren released a statement indicating Trump administration reforms at the Consumer Financial Protection Bureau (CFPB) imposed up to $26.5 billion in additional costs on American consumers. The assessment attributes approximately $22.5 billion to the repeal of limits on credit card late fees and bank overdraft charges, while roughly $4 billion stems from abandoned enforcement cases and restitution agreements.The report highlights that previous CFPB rules intended to cap most late fees at $8 and limit certain overdraft charges to $5, which were projected to save consumers about $10 billion annually. Warren criticized the administration's staff cuts and rule rescissions for weakening the agency's regulatory role, though the CFPB and White House have not yet responded to these allegations.

Woofun AI reports that Moonshot AI has updated its platform with the Kimi K3 model, offering users access to K3·Max and K3 Cluster·Max versions. Sources indicate the model contains between 2 trillion and 3 trillion parameters, positioning it as China's largest AI model by scale. Its performance is comparable to Anthropic's Opus 4.8, which industry estimates place at 1.5 to 2 trillion parameters.

Woofun AI reports that the Dark Side of the Moon website has updated a new model, KIMI K3, offering users access to "K3·Max" and "K3 Cluster·Max" versions. Sources indicate the model possesses a parameter scale ranging from 20 trillion to 30 trillion, positioning it as the largest AI model developed in China to date. Its capabilities are assessed to be comparable to those of Anthropic, a leading US AI laboratory. While Anthropic has not officially confirmed specific parameter counts for its models, industry estimates suggest its Opus 4.8 variant contains approximately 1.5 trillion to 2 trillion parameters.

Woofun AI reports that Andreessen Horowitz has invested in Runta, a startup focused on AI agent security. The specific investment amount remains undisclosed. Runta founder Guanlan Dai, formerly of Cloudflare and Kong, describes the company's approach as managing agents like children, requiring boundaries and supervision. Runta is developing "AI Agent guardianship" infrastructure to control permissions, limit risks, and prevent data leaks or financial losses. The system aims to serve as a "parental control layer" for businesses deploying autonomous agents, offering identity management and operational oversight.

Woofun AI reports that Sleepagotchi is developing a decentralized AI health application to reconstruct the Web3 health economy. Originally a 'sleep to earn' mobile game, the project has evolved into a comprehensive health platform utilizing AI to analyze smartphone and wearable data for personalized insights. CEO Kenny Wood emphasized that sleep issues are root causes of many health problems, prompting the integration of on-device AI and encrypted infrastructure to keep sensitive biometric data off corporate clouds and blockchains. The system employs a decentralized multi-agent architecture, including sleep, health, diet, and shopping agents, to process tasks locally and provide real-time recommendations.The project plans to utilize its native staking token, SLEEP, to support its marketplace and advanced features. While basic insights remain free, users must pay in SLEEP for additional AI queries after exhausting daily quotas. Revenue streams include subscription fees, marketplace listings, partner staking deposits, and affiliate earnings from the shopping agent. Sleepagotchi has raised $6.5 million from investors such as 6th Man Ventures, Collab+Currency, and GSR. The project currently boasts over 2 million users and generated more than $100,000 in revenue during a three-week testing period.