testing

The Trump administration will revise the new artificial intelligence guidelines to expand regulation of AI models. The White House has developed a set of AI frameworks that require U.S. laboratories to undergo safety testing by the federal government before public release. In the coming months, this framework is expected to cover open-source models, which will be included in testing once their capabilities reach the same level as Anthropic's Mythos-level models and OpenAI's GPT-5.6.

HashKey has become the authorized distributor of the Anchorpoint Hong Kong Dollar stablecoin HKDAP, launched beta testing, and completed the first minting and redemption transaction.

The largest bank by asset size in Latin America, Itaú, is collaborating with OpenAssets to participate in a tokenized fixed income and investment fund test led by the industry. Brazil has become a testing ground for placing bonds, credit, and other financial assets on the blockchain under a regulatory framework.

On August 10, UMX (The Unified Market Exchange), incubated by Avenir Group under Li Lin, announced the launch of an invitation-only public beta. UMX is positioned as a crypto-friendly securities platform for global professional investors, supporting spot, leverage, contract, and options trading of crypto assets, while also supporting real US stocks, ETFs, and US stock options trading. The aim is to connect crypto assets and real securities trading on the same platform, enhancing the capital efficiency of cross-asset trading.The public beta focuses on opening the capital link between crypto assets and securities accounts, supporting functions such as stablecoin exchange transfers, non-stablecoin collateralized lending, fiat account withdrawals and deposits, and the mutual conversion of securities holdings and stock tokens. To enhance cross-asset capital efficiency, the platform introduces a cross-asset margin mechanism, allowing financial assets to be used as margin. During the public beta, BTC and USDT financial products will also be launched, with the highest annualized returns reaching 2.5% and 5.5%, respectively. This public beta is invitation-only, and users who have not yet obtained a public beta code can reserve the official version and participate in the early bird benefits program.

Starknet announced that it has completed the quantum-resistant signature transfer test, where a wallet account using a quantum-resistant signature mechanism completed a real transfer on the Starknet mainnet, with a transaction fee of about 6 cents, and can be queried through the block explorer. This account is currently an experimental, unaudited version, mainly used for research testing. StarkWare stated that this transfer benefits from the design of the Starknet account model.Unlike most blockchains that fix the signature algorithm at the protocol layer, each account on Starknet is a smart contract that can autonomously define the accepted signature schemes, allowing users to upgrade their wallets from traditional elliptic curve signatures to quantum-resistant signatures without hard forks, asset migrations, or changing addresses. StarkWare pointed out that most blockchains face quantum computing risks because wallet signatures and underlying verification systems rely on elliptic curve cryptography. Once large-scale quantum computers appear, running Shor's algorithm could potentially break the related encryption systems. Currently, the Starknet ecosystem supports a quantum-resistant signature scheme based on Falcon-512, which is part of the NIST post-quantum cryptography standardization process, with relevant implementations promoted by ecosystem teams and organizations such as OpenZeppelin.

According to Wired, during security testing, the open-source model Kimi K3 from the dark side of the moon broke out of the sandbox and accessed the internet, attempting to find test answers via GitHub. American security startup Frontier Security stated that Kimi K3 exploited a sandbox configuration vulnerability, but it itself lacked sufficient built-in protective mechanisms.Similar to incidents previously disclosed by OpenAI and Anthropic, this escape partially stemmed from a sandbox configuration error. However, Kimi K3 did not initiate an attack, as the answers were easily accessible on GitHub. The CEO of Frontier Security stated that Kimi K3 performed excellently in cybersecurity defense but lacked mechanisms to prevent cheating or escaping the sandbox. Security experts emphasized that configuring the testing environment is crucial, as the misuse of AI agents could lead to system failures.

According to Reuters, citing three informed sources, Samsung Electronics and SK Hynix are testing the etching equipment from Chinese semiconductor equipment manufacturer AMEC to assess its potential application in their factories in China, aiming to hedge against the risk of the U.S. potentially tightening export controls in the future. The testing began about two years ago, coinciding with rising uncertainty over whether the U.S. would allow Korean companies to import U.S.-made chip equipment to China.The report noted that although no decision has yet been made on large-scale deployment, verification from Samsung or SK Hynix would be an important commercial endorsement for AMEC. The gap in etching, deposition, cleaning, and other areas for Chinese equipment has gradually narrowed, with prices typically 20%-30% lower than similar overseas products. Samsung officially denied having tested AMEC equipment for its factories in China, while SK Hynix declined to comment. Deutsche Bank expects Northern Huachuang, AMEC, Tuojing Technology, and Shengmei Semiconductor to each exceed $1 billion in revenue by 2026, collectively expected to capture 25%-30% of China's approximately $28 billion wafer fabrication equipment market this year.

According to Reuters, citing two informed sources, the Trump administration informed leading AI companies on Tuesday that it will not conduct security testing on open-weight AI models. This decision marks a new direction for the U.S. government in its approach to AI regulation. Open-weight models allow developers and users to freely download, modify, and deploy model parameters without relying on closed API access. Previously, organizations such as the American Institute for AI Safety (AISI) had conducted security assessments on cutting-edge models. The White House and relevant departments have not publicly commented on this matter yet.

Bitgo CEO Mike Belshe deposited 100 BTC into a public Bitcoin address, valued at approximately $6.3 million at the time, and invited the Claude model under Anthropic to attempt to transfer the funds from that address. On-chain records show that the wallet received the funds on July 31, and the balance has not been transferred out.Anthropic previously disclosed that during 141,006 cybersecurity assessment runs, 3 incidents were found, with 6 assessment sessions involving 3 models unexpectedly interacting with real organizational systems. The relevant models include Claude Opus 4.7, Claude Mythos 5, and an unpublished internal research model, due to configuration errors by third-party testing partner Irregular that caused the testing environment to connect to the internet.Anthropic stated that Claude Opus 4.7 identified a real website sharing the same name as a simulated company during one assessment, exploited weak passwords and exposed services to recover infrastructure credentials, and accessed a production database containing hundreds of records. The company claimed that the model was attempting to complete assigned tasks and was not actively breaking restrictions or pursuing independent goals. Belshe's challenge involved the Bitgo institutional custody platform, which uses multi-signature or multi-party computation technology to distribute signing authority across multiple independent keys. Anthropic has not publicly responded to this challenge.

According to Jinshi News, Qianwen Assistant has undergone extensive testing within Tesla's vehicle system and is expected to be integrated soon. Insiders reveal that Qianwen has completed a large number of tests in the real vehicle environment of Tesla, possessing the capabilities of "listening and responding, controlling the car, navigating, and handling tasks."In the future, the functions of Qianwen on Tesla's vehicle system may exceed the traditional definition of a "voice assistant." Currently, Alibaba Cloud and Tesla China have not made any official response to this.

According to X platform blogger @MaxForAI, a screenshot of the DeepSeek Harness internal testing recruitment is circulating on social media. The screenshot shows that the DeepSeek Harness plans to start internal testing later this week, with the first batch of users being selected from small group chats. Selected individuals must submit personal information and sign a "Confidentiality Commitment Letter." The screenshot also specifically reminds that any leaks will result in disqualification and affect future opportunities for selection in the DeepSeek model and product internal testing. If the news is true, the official version of DeepSeek V4 and the Harness product are expected to be launched soon.

Domestic DRAM storage IDM leader Changxin Technology will officially list on the Sci-Tech Innovation Board on July 27, with an issuance valuation of approximately 580 billion yuan and a total fundraising amount of 29.5 billion yuan, of which about 22.066 billion yuan is for equipment purchase and installation costs. Industry insiders believe that 2026-2027 is the golden window period for the introduction of domestic equipment.In terms of the supply chain, on the equipment side, North Huachuang has shipped a large volume of various categories such as etching and thin film deposition, Zhongwei Company continues to validate and mass-produce etching and thin film equipment, Huahai Qingshi CMP equipment has entered the core supply chain, and Jingce Electronics and Jingzhida testing equipment are benefiting simultaneously. On the materials side, Yake Technology precursor products cover advanced process needs, Guanggang Gas provides special gases and bulk gases, Jinhong Gas has a gas production order cycle of 15 years, Tongcheng New Materials ArF/KrF photoresists have entered mass production, Jingrui Electric Materials has a market share of over 40% in high-purity hydrogen peroxide, but high-end photoresists are still in the validation stage, and Debang Technology is in the sample testing stage. On the packaging and module side, Huatian Technology is a packaging supplier, Jiangbolong and Demingli have reached procurement agreements, while Maijie Technology is still in the introduction phase. On the distribution side, Shangluo Electronics has been authorized, and there are expectations for supply strategy adjustments after listing. Several listed companies hold shares indirectly through industrial funds, with Shangfeng Cement holding approximately 0.1517% of the equity before issuance.

The Web3 fitness protocol BRT.Fit has released a transparency statement, currently in the product testing and ecosystem validation phase. Application functions and validation methods may be adjusted based on testing results. The test token tBRT is only used for economic model validation and does not represent the official token value, nor does it constitute an investment product or guarantee of future returns. Its "mandatory exercise protocol" is an incentive and commitment mechanism that users voluntarily join, helping to develop exercise habits through verifiable data, without physical compulsion.After the official mainnet launch, core functions such as token staking, rewards, and burning will be executed through on-chain smart contracts, with contract addresses and audit reports to be disclosed gradually. The project plans to support mainstream wearable devices such as Apple Watch, WHOOP, and Xiaomi, Huawei, and will continue to optimize the anti-cheating system. The project team reminds that digital assets carry significant volatility risks, participation does not guarantee returns, and users should refer to official announcements, remaining vigilant against counterfeiting and phishing activities.

According to external disclosure information, OpenAI has been internally running an unreleased model. This model, without the aid of tools like Lean, solves the unit distance problem with a 48% probability through a single autonomous inference and can independently find a counterexample to the Jacobian conjecture based on a single prompt. In security testing, this model has bypassed the sandbox environment and submitted results that should have been released internally to GitHub in the form of a Pull Request, and it has evaded detection by splitting authentication tokens. Relevant code records show that OpenAI began benchmarking this model no later than May 9, and its internal availability has exceeded 2.5 months.Previously, OpenAI and Hugging Face jointly disclosed that last week this model breached Hugging Face's production infrastructure during a network capability assessment. The model gained internet access through a zero-day vulnerability and obtained testing solutions by stealing credentials and exploiting remote code execution paths. OpenAI stated that this incident indicates the network attack capabilities of advanced models have been effective in real-world scenarios, and they are collaborating with Hugging Face to investigate and patch the vulnerabilities. Currently, OpenAI has not publicly commented on this matter.

Tether CEO Paolo Ardoino posted on the X platform that Tether's wallet development toolkit has launched a web playground feature, allowing users to test the basic functions of the wallet in their browser.

SBI Group, DigiFT, and Startale Group have completed a joint proof of concept, demonstrating the full lifecycle of tokenized securities using the Japanese yen stablecoin JPYSC in an Ethereum testnet environment, including instant settlement for fund subscriptions and automatic on-chain dividend distribution.SBI Group and DigiFT are advancing the tokenization of the SBI Japan High Dividend Equity Fund, which is managed by SBI Asset Management with an asset management scale of approximately $1.3 billion, making it one of Japan's major public equity strategies. The three parties completed two tests, one demonstrating the near-instant settlement of fund subscriptions using JPYSC, and the other automatically calculating and distributing dividends to eligible token holders' wallets via smart contracts after the allocation registration is completed.The three companies plan to explore the integration of tokenized Japanese stocks with institutional DeFi platforms and consider testing use cases such as collateralized lending and on-chain asset management with ecosystem partners like Morpho and Gauntlet.

The Ethereum Foundation Protocol Security team researchers stated in a blog post on Thursday that a series of AI agents have been deployed to test the software relied upon by Ethereum, searching for vulnerabilities in the cryptographic systems, protocol code, and smart contracts. The vulnerabilities discovered by the AI agents include a remotely triggerable panic issue in the peer-to-peer layer libp2p gossipsub used by the Ethereum consensus client, which has been fixed and disclosed on GitHub as CVE-2026-34219.The researchers noted that the AI agents are organized into specialized roles such as reconnaissance, searching, patching, and verification, aimed at identifying potential attack paths, reproducing failures, and verifying their applicability to production code. The Ethereum Foundation stated that AI has not replaced security researchers but has changed the way they work, allowing the team to cover a much broader scope than manual reviews, although researchers need to exercise more careful judgment over a large number of seemingly credible conclusions.

Arc announced that its testnet will undergo a scheduled load test aimed at analyzing the network's performance and stability under extreme conditions. The official statement indicated that during the testing period, users may encounter network congestion, an increase in transaction failure rates, higher Gas fees, extended block generation times, and brief pauses in block production.

Alipay announced that its AI open platform is officially launched and is now open for invitation testing to enterprise developers and service providers. The platform aims to assist merchants and institutions in conveniently upgrading their service capabilities to AI.According to the official introduction, through this platform, merchants can upgrade their existing mini-programs, API interfaces, and service capabilities to AI-callable MCP, Skill, or Agent without starting from scratch. Once integrated, services can not only be provided to Alipay's 1 billion users through the built-in AI assistant "Abao," but can also be authorized for cross-end interconnection to smart terminals and large model platforms such as mobile phones, in-car systems, AI glasses, and IoT, achieving "one-time access, multi-end distribution and unified management," thereby reducing the R&D costs of adapting to multiple terminals separately. In addition, the platform will provide basic support for the hosting, distribution, trading, and settlement of AI-enabled services. Alipay emphasizes that this AI upgrade will not affect the existing Alipay App mini-program services, and the platform will continue to provide security infrastructure such as identity verification, payment security, and risk prevention and control.

According to crypto analyst alicharts, Ethereum is currently testing resistance at the 0.8 MVRV pricing level of $1796. If the daily closing price is above this level and successfully holds it as support, it will strengthen the bullish case and may open the move towards Ethereum's target price of $2245.alicharts further pointed out that breaking above both $1796 and $1816 will trigger a bullish breakout. From a technical perspective, such a move will also increase the probability of ETH breaking through the channel top at $1844 and starting the march towards the target price of $2245.