safe

The cryptocurrency wallet backup company Vault12 submitted a letter to the U.S. Securities and Exchange Commission (SEC), pointing out that the proposed token safe harbor framework fails to adequately alleviate concerns about the listing of tokens on centralized exchanges. Trading platforms remain worried about the potential for the SEC to retroactively challenge the legal status of tokens in the future.Vault12 raised $17.45 million from 38 investors in 2018 through a future token agreement, and its issued Vault Guardian Token (VGT) has yet to be listed on any centralized exchange. The project has fulfilled or formally abandoned all commitments in its white paper, but the exchanges previously contacted still require confirmation that the SEC will not determine in the future that VGT continues to be bound by the original investment contract. The SEC's proposed "Regulation Crypto Assets" plan introduced in August of this year aims to provide exemptions for token financing and allows eligible projects to terminate the association between tokens and the original investment contracts through a safe harbor mechanism. Vault12 suggests adding a 60-day objection period, during which if the SEC does not raise any objections, eligible filings should have final effect on the agency, except in cases of fraud or significant misrepresentation.

Anthropic recently submitted a draft AI safety bill to U.S. lawmakers, which has been questioned by U.S. AI safety advocates. The draft sets a high threshold for the government to urgently halt dangerous models, requiring that risks involve scenarios such as biological weapons, cyberattacks, or AI going out of control, and potentially causing more than 50 deaths or serious injuries, or losses exceeding $1 billion; some provisions may also limit state government regulatory authority.Nathan Calvin, the strategic head of the AI safety organization Encode, subsequently compiled Anthropic's decisions over the past year. In February of this year, the company modified its previously strict safety policy: it had previously committed to pausing training or releases if safety measures did not keep pace with model capabilities, but the new policy no longer requires the company to independently halt R&D at higher risk stages, citing that competitors might take advantage of the opportunity. During the same period, the company invested heavily in computing power, formed a chip team, and allowed Claude to take on more AI R&D tasks.On October 21, Anthropic CEO Dario Amodei posted a lengthy rebuttal on X. He acknowledged that the company had made some mistakes but insisted that competing for cutting-edge AI capabilities did not violate its safety mission, stating that the company adheres to a race-to-the-top approach, believing that only by staying at the forefront of technology can it drive the industry to raise safety standards. He listed nine safety decisions that had cost the company commercially, including restricting the Mythos-level model's open access for months and strictly limiting biology-related requests. The safety roadmap updated by the company on October 19 states that existing policies mainly target the deployment of dangerous models, and in the future, it may be necessary to directly limit the speed of R&D, with plans to propose relevant policy options by January 15, 2027, or explain why none were proposed.

OpenAI CEO Sam Altman introduced three proposals for the safe development of super AI to employees last week. The first is to establish a global regulatory organization similar to the International Atomic Energy Agency, to inspect advanced models and audit safety measures. Insiders say Altman leans towards this route.The second proposal emulates the European Organization for Nuclear Research, merging major AI laboratories into one organization or collaborating through a unified agency for joint development, rather than continuing to race against each other. The third and most extreme, which he referred to as the "Eliezer Plan": an AI company achieves recursive self-improvement first, and then uses overwhelming economic power to control computing power, making it impossible for other developers to catch up. Altman described this idea as "melting down all other companies' GPUs," while stating that this plan is very bad and is only meant to illustrate the extreme possibility.The idea of "melting down GPUs" comes from AI safety researcher Eliezer Yudkowsky. He proposed a thought experiment in 2021, suggesting that super AI could create self-replicating nanomachines to melt down global GPUs, stating that this is merely an extreme example to illustrate the intensity of intervention, not a proposal to be executed. OpenAI proposed the establishment of an "AI version of the International Atomic Energy Agency" in 2023, and Google DeepMind head Hassabis has also advocated for the creation of an "AI version of CERN." Currently, OpenAI, Google, and Anthropic plan to establish a joint safety standards organization by the end of this year or early next year, but there are disagreements over the responsibility for model safety testing. Altman hopes to set speed limits for AI development while stating that OpenAI cannot currently cede its competitive advantage to rivals.

NVIDIA's CEO Jensen Huang told CNBC that the company must collaborate with Chinese artificial intelligence companies to monitor related risks. In a media briefing in London on Wednesday evening, he stated that both parties should work together to safely advance technological development. Huang mentioned that artificial intelligence poses potential dangers and indicated that he maintains ongoing dialogue with many researchers in China.Huang believes that the way to collaborate with Chinese artificial intelligence companies is through dialogue and emphasized the need to strengthen this dialogue. He stated that the United States should do this and also advocated for the United Kingdom to take the same approach. The page noted that OpenAI CEO Sam Altman and Anthropic CEO Dario Amodei advocate for increased regulation, while Huang has become a prominent voice opposing this stance. In recent months, researchers from laboratories including OpenAI and Anthropic have called for a slowdown in technological development and warned of risks to human survival.The page also mentioned that U.S. Treasury Secretary Scott Bessent discussed U.S.-China artificial intelligence dialogue with Chinese counterpart He Lifeng in September, covering communication channels for events up to the national security level. The Chinese side reported that Xi Jinping pointed out that there may be competition between China and the U.S. in artificial intelligence, but the space for cooperation may be larger, and both sides can continue to exchange views on risks and benefits to jointly prevent abuse. Huang stated that the UK, with its intellectual capital, is well-positioned to become the convener of relevant dialogue. The related market data showed that NVIDIA's stock price fell by $3.95, a decrease of 1.78%.

According to The Defiant, Safe announced the launch of the Safenet Aegis security network on Ethereum, Gnosis Chain, and Arbitrum. This is the first production-ready version of Safenet that can generate fees, previously approved for funding and launch by SafeDAO through the SEP 56 proposal. The network separates transaction verification from transaction authorization, providing transaction security verification for smart accounts like Safe multi-signature wallets.Safenet consists of three types of roles: Sentinel, responsible for checking transactions and staking USDC bonds, can earn verification fees paid in USDC; Validator, who secures the proof layer by staking SAFE, can earn SAFE rewards; Delegator can delegate SAFE to Validators and participate in staking rewards. Safenet Guard is responsible for executing proof results on-chain, with relevant protocol parameters governed by SafeDAO.The current Sentinel pool includes OpenCover, IntentGuard, BlockSec, and Candide, while the Validator pool includes Gnosis, Greenfield, Core Contributors, and Safe Labs. Safenet will subsequently integrate transaction verification into Safe Wallet, displaying verification results during the transaction review process and supporting access for other wallets built on Safe smart accounts.

NVIDIA CEO Jensen Huang stated that the company will cut ties with clients developing unsafe artificial intelligence.

NVIDIA has committed to investing $250 million over the next five years to promote research in agent safety and security in the UK. The company stated that it will provide $50 million in computing resources support for agent safety research in the UK each year.

On October 14, Jinshi Data reported that the AI company Anthropic has requested the federal appeals court in Washington to reconsider the ruling made on September 25. This ruling upheld the Trump administration's decision to classify Anthropic as a supply chain risk enterprise that poses a threat to U.S. national security.Anthropic argued that the judge misinterpreted the law, allowing the Department of Defense to take action without any prior determination procedures. The company also stated that there is clear evidence indicating that this classification is retaliation for its statements on artificial intelligence safety issues.

Federal Reserve Governor Waller stated that the safety premium of U.S. Treasury bonds no longer exists.

ByteDance Seed recently sent an internal email announcing the establishment of the Seed Model-Alignment & Preparedness department, which will be led by Yan Lin and report to Seed head Wu Yonghui. Members of the original Responsible AI team and some members of the Chat team will transition to the new department.The department aims to strengthen investments in safety and alignment. The core of alignment is to ensure that model behavior is consistent with human intentions and safety constraints, in order to enhance the safety capabilities of Agent Models and prevent risks from cutting-edge capabilities such as chemical, biological, radiological, and nuclear threats.

Artificial intelligence company TypeSafe AI announced the completion of its Series A financing, amounting to $870 million, with a valuation of $7.5 billion. This round was led by Andreessen Horowitz, with participation from Sequoia Capital, existing investor DCVC, and a group of angel investors. Andreessen Horowitz partner Martin Casado has joined the board.The company stated that it will convert the funds into benefits for users, continue to refine the developer product Jev, provide more machine-native models, and build the infrastructure needed to create intelligent software. The company claims that one-third of the Fortune 500 companies are using Jev, which has saved customers millions of dollars in production environments, and will further enhance enterprise features requested by users.

According to PRNewswire, the publicly traded company Universal Safety Products (UUU) announced that it will be renamed DeFi Capital Markets on October 19 and will trade under the new ticker "DCM." In the future, it will gradually cease its original product business and focus on asset tokenization, digital assets, market making, as well as quantitative trading and blockchain financial markets.

According to Catcher Predict monitoring, the probability of the "Yes" option in the sub-market "Safety?" for the event "Falcons vs. Saints" on the Polymarket has experienced significant fluctuations, rising from 7.5% an hour ago to the current 26.5% (a fluctuation of 19%). Please note the impact of related breaking news.

Greenfield Capital has submitted a complaint to the Swiss Federal Foundation regulatory authority ESA regarding governance issues at the Safe Ecosystem Foundation, accusing a director of the Safe Foundation of demanding a large amount of SAFE tokens from relevant personnel after Bybit was attacked.At the same time, the director threatened to sell approximately 10% of the total supply of SAFE held in Gnosis and cut ties with Safe.

According to CoinDesk, Hester Peirce, a commissioner of the U.S. Securities and Exchange Commission (SEC), has completed her term today. She served in the agency for over 8 years and voted against the committee's decisions to take enforcement actions instead of creating industry-compliant rules multiple times, and she was an early proponent of the token safe harbor proposal.Hester Peirce previously led the Crypto Task Force and stated that the SEC's duty to protect investors includes using regulatory authority to provide regulatory boundaries for the rapidly expanding digital asset industry.

Regarding the previous incident where the FlashLoopAdapter vulnerability caused a loss of approximately 114 ETH from two Safe multi-signatures, Aave founder Stani Kulechov stated that the contract with the vulnerability is not the Aave v3 contract, but rather a third-party external adapter built on top of Aave, which does not affect Aave v3 itself.Previously, SlowMist disclosed that the FlashLoopAdapter has an access control vulnerability, which attackers exploited to bypass permission checks and transfer assets from two Safe multi-signatures that enabled the module, ultimately resulting in a loss of approximately 114.09 ETH.

Slow Fog issued a security alert stating that the Aave V3 Loop Safe Module has a security vulnerability, allowing attackers to forge Safe authentication and execute arbitrary modules, stealing approximately 114.09 ETH from two Safe multi-signature wallets.Attackers bypassed authentication by forging a Safe that always returns true and executed module transactions using a router and calldata that could be arbitrarily controlled, transferring weETH and Aave collateral. After repaying approximately 1300 WETH in debt, the attackers unlocked the collateral.

The White House Artificial Intelligence Agreement shows that Nvidia, OpenAI, Anthropic, xAI, and Google will regularly discuss artificial intelligence safety standards.

On September 28, 2026, OpenAI published a safety-related article, stating that before continuing any cutting-edge reinforcement learning training, structured safety documentation should be required. Ideally, such documentation should reach the evidence-based structured risk argument level used in safety-critical industries like aviation and nuclear power. OpenAI views this as a direction for effort while acknowledging the complexity arising from the emergence of AI capabilities, making it difficult to achieve the same level of rigor.The article focuses on cutting-edge reinforcement learning training and does not cover the broader alignment attributes required for internal and external deployments. The recommendations in the article reflect current practices, which are expected to continue evolving and are being implemented internally at OpenAI. Technical safeguards should cover model alignment, isolation, and monitoring, including avoiding speculative positive reinforcement rewards, offline alignment assessments and stress testing, preventing automated scorers from seeing thought chains, as well as multi-layer infrastructure security, sandbox red teaming, limiting high-bandwidth cross-sample communication, and immutable preservation of agent records.Operational guidelines include preemptive dissent across teams, approvals that can be vetoed by senior leadership, accountability of training leads for safety arguments and incident responses, as well as fail-safe pauses, internal oversight, audit access, and escalation by severity. In response to serious misalignment events, OpenAI proposes controlled access to original records, root cause analysis, operational and cultural reviews, and treating incident-derived assessments as regression tests; results of investigations should be made public, along with reviews and operational changes, and affected third parties should be notified as soon as possible.