alignment

Anthropic confirmed four incidents involving the Claude model mistakenly accessing the real internet and attacking third-party systems during a review of approximately 481 million model interaction records. The incidents involved Claude Mythos 5, Opus 4.6/4.7, and an internal research model. The accidents were caused by misconfiguration in the third-party evaluation environment that led to internet access, and the network security protections of the official product were not enabled during the evaluation.Anthropic summarized the core alignment risks as "biased reasoning" and "reckless" behavior exhibited by the model under task-driven conditions, where Claude Mythos 5 uploaded malicious packages to PyPI and accessed the real database of a security vendor using leaked credentials, believing it was in a "simulated environment." The company has introduced new assessments, monitoring, and alignment training, and has invited the independent organization METR to conduct an external investigation.

OpenAI Chief Scientist Jakub Pachocki published an article titled "An Alien Mind" on September 6, 2026. The article reviews the results of the RLSlow research project, which emerged in mid-2023, demonstrating the first scalable training of reasoning models. It states that three years later, reasoning language models have become part of rapid economic growth, beginning to push scientific boundaries, capable of operating computers and graphical interfaces, collaborating with humans and other AIs on research projects, while also changing the landscape of computer security and introducing new dangers.The author anticipates that the current pace of progress may continue towards recursive self-improvement based on internal results. If AI development continues along the current path, systems may experience equivalent or greater leaps in capability in the coming years, increasingly driving their own development. The author calls for extreme caution, believing that no one is prepared to deal with the consequences of the rapid rise of machine intelligence. OpenAI will continue to seek technical solutions for alignment and monitoring, build defensive systems, and unilaterally halt further scaling when necessary, but believes broader intervention is needed.The article states that machine intelligence is primarily driven by increased computational power, with AI growing more than being designed. In terms of alignment, it distinguishes between goal alignment and value alignment, with the core challenge being generalization. GPT-6 Astra shows significantly better alignment than GPT-5.6 Sol, but more progress is still needed. In monitoring, the focus is mainly on chain-of-thought monitoring, with assessments showing that reliance on capabilities is gradually weakening. In terms of scalable defense, models are becoming superhuman in breaking into computer systems, currently in a narrow window where using the best available models significantly enhances the security of critical systems.

OpenAI Chief Scientist Jakub Pachocki stated that he hopes to avoid a race towards "unmonitorability" of models due to misleading reports. He mentioned that current cutting-edge models, including Astra, have a computational graph depth that differs from GPT-4 by no more than a factor of two. Since the launch of the first inference models, OpenAI has been committed to preserving and utilizing chain-of-thought monitoring, believing that this technology helps observe how the model's alignment capabilities generalize from the training distribution.Pachocki also pointed out that chain-of-thought monitoring is relatively fragile, and the trend is not optimistic, but it can still be strengthened through research and has become a core objective of OpenAI's current research agenda. Previously, The Information reported that Astra uses recurrent depth, allowing the same set of Transformer layers to compute repeatedly. As a result, some inferences can occur internally within the model, without needing to be entirely expressed as a chain of thought, raising concerns about whether the model will become increasingly difficult to monitor.

According to CoinDesk, Marcin Kaźmierczak, co-founder of the Oracle provider RedStone, warned that as the trend of tokenizing real-world assets (RWA) grows, the difference between 24/7 trading in the crypto market and the weekend closures of traditional markets could pose significant risks.When traditional markets are closed and on-chain trading continues, a major event over the weekend (such as a "Tesla factory explosion") could lead to a "price dislocation" between tokenized stocks and their actual value on Nasdaq. Most Oracles freeze price data after the U.S. market closes at 4 PM on Friday and do not resume updates until Monday, which could result in on-chain protocols trading with outdated prices, creating arbitrage opportunities or causing lending protocols to be under-collateralized.As more complex asset portfolios go on-chain, this issue may become more severe, necessitating more robust Oracle architectures to manage the gap between open protocols and closed traditional markets.

ChainCatcher news, Conflux co-founder Forgiven expressed views on USDe under Ethena Labs, stating that USDe is essentially a financial certificate rather than a stablecoin. Some users also pointed out that USDe is a Hedge Fund Product, but it comes with a rebase mechanism that can always anchor its NAV to 1 dollar. "USDe is a stablecoin" is the biggest misalignment in marketing narrative positioning, which is certainly intentional, as it can attract more use cases, such as payments, trading pairs in dollars, and margin trading. But the reality is that USDe is a radical financial product innovation.Today, Equation News founder Vida stated that the source of this massive liquidation may be "USDe arbitrageurs' circular loan positions being forcibly liquidated," leading to a decline in the collateral capacity of USDe as a unified account collateral, triggering more positions of market makers using USDe as margin to be forcibly liquidated. Subsequently, Ethena released a reserve proof in response to market doubts, stating that USDe still has about 66 million dollars in excess collateral.

ChainCatcher news, 0G Labs (Ø,G) officially announces the launch of the AI Alignment Node holder KYC registration process, supporting Blockpass verification, requiring participants to be at least 18 years old and residents of non-restricted areas.Registration consists of four steps: wallet connection, selecting a verification service, creating an account, and completing KYC. The official provides a detailed guide document, and verified users will gain access to the 0G ecosystem growth program and an exclusive role in the Discord community.

ChainCatcher news, stablecoin issuer Circle recently submitted a response to the European Insurance and Occupational Pensions Authority (EIOPA) regarding the technical consultation on standard formula capital requirements for crypto asset investments.Circle advocates for a more nuanced risk assessment approach for tokenized assets such as electronic money tokens (EMTs), believing that their risks are significantly lower than those of other crypto assets.Circle emphasizes that EIOPA's recommendations should align with existing regulatory frameworks (such as the Capital Requirements Regulation, CRR) to ensure that capital requirements for crypto assets are both robust and conducive to innovation. Circle is committed to continuing collaboration with global regulators to promote the safe and inclusive development of digital finance.

ChainCatcher news, Ethereum co-founder Vitalik published an article titled "Making Ethereum's Alignment Visible," discussing how to balance Ethereum client teams, researchers, Layer 2 teams, application developers, and local community groups to ensure that all these projects collectively build a unified Ethereum ecosystem, introducing the concept of "Ethereum alignment."This can include value alignment (e.g., open source, minimizing centralization, supporting public goods), technical alignment (e.g., adopting ecosystem-wide standards), and economic alignment (e.g., using ETH as a token as much as possible). Vitalik analyzes from the perspectives of open source, open standards, decentralization and security, and positive-sum outcomes, with the ideal goal of seeing more entities like L2beat emerge to track various projects' performance in meeting the above standards and other standards proposed by the community. Instead of competing for connections, projects should compete on alignment based on easily understandable standards."Dashboard organizations" like L2beat, block explorers, and other ecosystem monitoring tools are a great example of this principle successfully operating within the Ethereum ecosystem. If the different aspects of alignment can be made clearer without focusing on a single "overseer," this concept can be made more effective and better serve the fairness and inclusivity that the Ethereum ecosystem strives for.