testing

Leifeng Network · Whale Rhino has learned that OpenAI is secretly developing a screenless AI hardware product in Shenzhen. Industry insiders say that the project's confidentiality level is extremely high.Engineers came to Shenzhen for a business trip without even bringing their phones. The product is planned to be launched in 2027.

Arya.ag, an agricultural warehousing and loan company in India, is testing a system for tokenizing warehouse receipts for stored grains on the Avalanche dedicated Layer 1 blockchain. Arya.ag is collaborating with Finternet to connect grain storage, warehouse receipts, collateral commitments, and loan statuses through this network. Devika Mittal, head of Ava Labs India, stated that the testing is underway, with each tokenized warehouse receipt representing ownership of the stored goods.Sanmesh Kalyanpur, a director at Finternet Labs, mentioned that Arya.ag's samplers collect information on stored grains and input it into the company portal. Finternet will integrate farmers, commodities, warehouses, and insurance information into a "composite token" for banks to assess collateral risks. Arya.ag stores approximately $2 billion worth of agricultural products in its warehouse network and supports loans of about 120 billion Indian Rupees (approximately $1.26 billion) annually, with its loan department, Arya Dhan, disbursing around $230 million in loans each year.The concept of Finternet originated from a 2024 paper by the Bank for International Settlements (BIS), co-authored by Infosys co-founder Nandan Nilekani and then-BIS General Manager Agustín Carstens, proposing the establishment of an interconnected unified ledger for tokenized assets. In 2024, the Indian government launched a 10 billion Rupee credit guarantee scheme to encourage financing against electronic transferable warehouse receipts. The two companies have not disclosed the expected launch date or initial deployment scale.

According to "Du Jia," Alibaba Cloud is currently beta testing a new AI investment research tool called Qovest, positioned as an AI financial research assistant for professional individual investors. It is currently open for experience in batches to professional investors, with the first batch focusing on verifying A-share condition searches, Deep Research, and continuous tracking and research of asset accumulation.Qovest supports research on companies, industries, themes, and markets. After users pose questions, it can help find data, read financial reports and announcements, analyze research reports, generate reports, and consolidate research into the investment research space, continuously tracking subsequent changes in research subjects. The product standardizes common workflows for individual investors' fundamental research into modules, such as organizing the operational trends of a single company, horizontally comparing the business structures and financial differences of two targets, breaking down business models, and generating visual charts and research cards. The product is aimed at individual investors with independent research habits, but ordinary retail investors can also use it.

The supply chain reports that the capacity for high-end testing interfaces for AI chips is tight. NVIDIA has significantly raised prices to secure the capacity of mainstream testing interface vendors' probe cards and test sockets, causing some R&D and engineering validation demands to be squeezed out.Including wafer foundry TSMC and chip manufacturers, multiple supplier validations have been initiated to cope with the supply pressure after capacity has been locked.

According to a report by the Financial Associated Press, the Chief Financial Officer of OpenAI stated that they are considering reforming the charging model, testing a new pricing model based on performance, gradually moving away from the usage-based billing method, and competing with lower-priced open-source models.

Douyin is currently conducting an internal test of the AI Workshop, aimed at ordinary creators and developers, providing a complete link from AI application building, debugging, and operation to platform distribution, deeply integrating with Douyin's internal traffic system. The AI development framework, interface capabilities, and cloud storage services are all built into the creator's backend.The Douyin App's recommendation page has added a new content form interest card, used to showcase AI applications created in the AI Workshop. Users can directly interact with these cards while browsing the recommendation page, without needing to switch to a mini-program or leave the main app, allowing them to complete Q&A queries, fun interactions, and light tool experiences within the information flow. The interest card relies on Douyin's personalized recommendation distribution, supporting custom gameplay and styles.Creators who do not understand coding can use natural language to describe their ideas for building, while technically skilled developers can modify underlying files for deep customization. Debugging, real-device preview, and publishing can all be completed in the same environment. The produced works can be directly converted into interest cards for placement in the recommendation stream. This product is currently in the invitation-only internal testing phase.

According to The Block, ZKsync developer Matter Labs has open-sourced the permission engine of its distributed ledger platform Prividium for financial institutions, allowing organizations to run, inspect, and modify core code in their own environments. The German central bank has become the first institution to test and deploy Prividium.Matter Labs CEO Alex Gluchowski stated that feedback from regulators regarding commercial core means reliance on a single vendor, and that open-sourcing the core "structurally eliminates adoption barriers."Other components of Prividium, such as the ZKsync OS core and Atlas sequencer, have previously been open-sourced. The permission engine is planned to operate as an independent component, allowing institutions to deploy it without a commercial agreement, but management tools and integration with existing systems remain commercial products. Gluchowski said, "The central bank running public code is the clearest signal in this direction."

DeepSeek has started internal testing of version 4.1 Flash intermediate version. The official WeChat group notification states that the new model has a new architecture, natively supports multimodal capabilities, is more powerful, faster, and cheaper. The current billing is the same as V4 Flash, with a maximum of 20 concurrent sessions per account.Users can now directly call it, and there is no need to change the base_url, just replace the original model ID with deepseek-v4.1-flash-expires-on-0910.

Licheng, Jingyuan Electronics, and Silergy all reported year-on-year and month-on-month revenue growth in August, with Licheng and Silergy setting new historical highs for two consecutive months. Licheng's consolidated revenue in August was 8.558 billion New Taiwan Dollars, a month-on-month increase of 3.48% and a year-on-year increase of 24.46%. After breaking the 8 billion mark for the first time in July and setting a historical high, August saw another record. The cumulative revenue for the first eight months reached 61.258 billion New Taiwan Dollars, a year-on-year increase of 30.83%. Jingyuan Electronics reported revenue of 4.08 billion New Taiwan Dollars in August, a month-on-month increase of 2.23% and a year-on-year increase of 31.58%. The cumulative revenue for the first eight months was 29.404 billion New Taiwan Dollars, a year-on-year increase of 35.53%. Silergy's revenue in August was 2.07 billion New Taiwan Dollars, with a month-on-month increase of nearly 3% and a year-on-year increase of nearly 29%. The cumulative revenue for the first eight months reached 15.26 billion New Taiwan Dollars, a year-on-year increase of over 20%.Licheng's main source of growth this year remains the rebound in demand for memory packaging and testing. Licheng previously anticipated that revenue in the third quarter would show single-digit quarter-on-quarter growth, with the fourth quarter having the potential to exceed the third quarter. The momentum for memory orders is expected to continue until the end of the year, with a positive market outlook at the beginning of next year. In addition to its core memory business, advanced packaging has become a key focus for the next phase of development. The panel-level packaging has already been adopted by American clients, and the production capacity of the P11 plant has been secured by clients, with plans to enter the mass production phase of integrated ASIC and HBM AI chips by mid-2027.Jingyuan Electronics has been actively expanding its high-end testing capacity in recent years, with AI-related products becoming an important driving force for its operations. The testing demand for AI GPUs, ASICs, and high-performance computing chips remains its main source of growth. Silergy pointed out that the demand for AI and high-speed interconnects is strong, including an increase in demand for high-performance computing chips such as CPUs, GPUs, ASICs, and AI accelerators. After completing the cleanroom construction at the newly acquired Hukou plant in February, Silergy began mass production in July, currently achieving 40% of the total capacity of the plant and gradually taking on new demand from overseas clients.

According to the official announcement, Binance has launched the AI feature suite "Binance AI," which is currently in Beta testing. The aim is to help users interpret complex market data, identify market dynamics, and reduce information noise through AI.It is reported that Binance AI will provide a personalized information interface, AI market insights covering both crypto assets and traditional assets, as well as concise and timely market dynamic updates. The first batch of users can apply to participate in the testing, and users who qualify for the whitelist will receive access instructions via email.Currently, the Binance AI Beta is only open to Binance users who have completed identity verification. The number of spots is limited and will be distributed in batches. Submitting an application does not guarantee testing eligibility. Binance also stated that user feedback will be used to further improve the product.

Anthropic stated that the company has temporarily paused certain AI training and cybersecurity assessments. The company disclosed the relevant adjustments in a blog post, explaining the measures taken in response to unauthorized actions by its AI agents earlier this year. OpenAI previously indicated that it had paused some model development work due to safety concerns. Anthropic also confirmed that it has taken similar measures and emphasized the need for broader coordination on the pace of frontier AI development.Anthropic stated that after disclosing three related incidents in July, the company paused external cybersecurity assessments for pre-release models and also briefly suspended internal testing of pre-release models. Following the incidents, the company also paused high-risk reinforcement learning environments in pre-release models for several weeks. Anthropic noted that most reinforcement learning work has resumed, but some high-risk environments remain on hold, awaiting human review or updates to monitoring tools.

According to a report by the Economic Daily, ASE Technology Holding benefited from the increased capacity of Google TPU, significantly adding advanced packaging and testing orders. Intel is accelerating the promotion of its EMIB advanced packaging platform while simultaneously expanding its outsourcing orders to ASE Technology Holding, with the previously raised testing quotes showing benefits. Google continues to increase its AI capital expenditures to expand data centers, using TPU to support the Gemini large model, AI agents, and Google Cloud services. The market estimates that Google TPU will enter a large-scale production phase by 2028, with shipments expected to reach 12 to 15 million units. ASE Technology Holding has mastered 2.5D, 3D advanced packaging, and high-end testing technologies, and Google is also expanding its use of TSMC's advanced processes and CoWoS capacity.Intel is racing to advance its EMIB packaging platform, with major cloud companies like Meta and Google successively adopting the EMIB and EMIB-T platforms. ASE Technology Holding, positioned as a pure testing and packaging foundry, can directly procure organic embedded silicon bridge substrates to undertake backend wafer-level assembly and high-end testing services. Wu Tianyu, the Chief Operating Officer of ASE Technology Holding, believes that Intel's EMIB and TSMC's CoWoS are not zero-sum competitors. As the testing time for AI chips lengthens, the costs of equipment, materials, and labor are rising, and high-end packaging and testing capacity is becoming increasingly tight. Major testing and packaging factories have gradually raised prices based on products, capacity, and customer conditions, with increases of about 5% to 10%.

XRP has dropped 6.23% in the past 24 hours to around $1.38, making it the token with the largest decline among the top ten cryptocurrencies by market capitalization. Previously, XRP surged from around $1 on August 18 to $1.69, an astonishing increase within four days, briefly touching the psychological level of $1.7.The pullback is mainly due to leveraged liquidations. XRP has still risen 35.55% over the past week, second only to Hyperliquid's 38.65% weekly increase. Notably, XRP-related ETFs have recorded net inflows for nine consecutive days, indicating that institutions have not withdrawn, and the issue mainly lies in the leverage aspect rather than capital outflows.After breaking through $80,000 on Tuesday, Bitcoin has retreated to around $78,000, dragging down the overall altcoin market. The market is awaiting tonight's core PCE inflation data and Nvidia's earnings report; whether XRP can hold the $1.4 support level will determine if this rally is a healthy correction or a further decline.

According to Techcrunch, AI Agent testing infrastructure startup Arga Labs announced the completion of a $10 million seed round financing, led by General Catalyst, with participation from institutions such as BoxGroup, Emergence Capital, Gradient Ventures, and SV Angel.The company aims to address the challenges of testing, training, and adapting complex business systems during the deployment of enterprise-level AI Agents by creating a complete digital twin environment for enterprise software, replicating permission systems, Webhooks, and system interaction logic, allowing AI Agents to train in an environment close to real work.

According to the "Science and Technology Innovation Board Daily," Alibaba Qianwen is currently testing a multi-Agent collaboration feature on the Qianwen creation platform in a small scope, mainly aimed at professional creative scenarios such as AI short dramas, comic dramas, marketing advertisements, and image design. Currently, only a small number of professional creators have been granted testing qualifications.This feature enables multiple Agents to collaborate, linking together the stages of creative planning, script writing, visual design, storyboard generation, and final production. Different Agents can cooperate based on task division to complete the entire production process from concept to final product. Users can participate in discussions throughout the process, adjust the creative direction at any time, or delegate some tasks to AI for independent completion.Recently, Qianwen has launched Qianwen creation for the professional content production field, and it has already integrated Alibaba's latest models such as Wan 3.0, Wan 3.0 Prime full version, and Qwen-Image 3.0 Pro.

Recently, multiple users discovered an image model named luna-lisa-alpha appearing on the LMArena testing platform, suspected to be a new generation image generation model that OpenAI is testing. Although not officially confirmed, there is speculation that it could be an early version of GPT Image 2.5 or the next generation of GPT Image.From the testing feedback, the new model shows significant improvements in character consistency, image realism, text generation capability, and noise control. Some testers compared it with models like GPT Image 2.0, Nano Banana Pro, and Grok Imagine, believing that luna-lisa-alpha can better maintain the appearance, clothing, and style of the same character across multiple generations. However, some testers pointed out that there are still issues with repeated textures in complex texture areas. On the same day, OpenAI announced that the GPT Image 2 API preview supports transparent background generation, which can be used for product images, web design, and other scenarios. The exposure of multiple checkpoint names has also led the community to speculate that OpenAI may be conducting A/B testing. Whether this model will be officially released remains uncertain, and we need to wait for official information from OpenAI.

According to X user MaxForAI, Tencent has begun gray testing its brand new flagship model, Hunyuan Hy4. Some users have discovered that the model selection list in the Tencent Yuanbao App now includes Hy4, labeled as an expert-level model, ranking above Hy3 and DeepSeek. The official account @TencentHunyuan also shows this model option at the relevant entry.Last week, Tencent confirmed in its Q2 financial report that the larger parameter-scale Hy4 will be launched soon, further enhancing model performance and multimodal capabilities. Currently, Tencent has not officially released Hy4, and the outside world cannot confirm whether this is a small-scale gray test or an early opening of the entry, but based on the progress, the model's launch seems imminent.

The Ethereum Foundation's DevOps team announced the launch of the Platåberget testnet as an early public testing environment for the Glamsterdam (Gloas + Amsterdam) upgrade. This testnet is open to the community and is planned to run for several months, providing developers with a stable experimental platform to identify and fix issues before the upgrade is deployed to long-term testnets like Sepolia and Hoodi. The Glamsterdam hard fork is scheduled to activate on August 20 on this testnet.The Glamsterdam upgrade includes several significant changes to both the consensus layer and execution layer, including built-in proposer-builder separation (ePBS), block-level access lists (BALs), gas repricing targeting approximately 200 million gas, an increase in the maximum contract deployment size from 24KiB to 64KiB, an increase in the initcode limit from 48KiB to 128KiB, and forward-compatible consensus data structures. Relevant EIPs are summarized in meta EIP-7773. Gas repricing will affect wallets, indexers, and gas estimation tools, and any tools with hard-coded maximum gas limits will need to be updated; EIP-8037 also introduces an independent state gas dimension, where new accounts or writing to new storage slots will be charged by state bytes.The Platåberget validator set is small and allows public participation, supporting deposits from validators or builders submitted through the Dora browser. The testnet provides one-click resources to add networks, faucets, and client images. The community can provide feedback on issues in the Ethereum R&D Discord and related specification repositories.

Bitcoin infrastructure development company Blockstream announced on Twitter that due to Boltz's ongoing automated detection and multiple vulnerability attacks leading to the suspension of its exchange services, Blockstream will launch its own Blockstream Swaps service.Currently, Blockstream Swaps is in the beta testing phase, aimed at providing a toolkit for the ecosystem and will complement existing service providers.