mode

Google (GOOG.O) announced the official launch of the Claude model deployment option managed by Anthropic on Google Cloud.

The 2026 World Physics AI Conference and Yushu Partner Ecosystem Conference was held today in Hangzhou. Wang Xingxing, Chairman, General Manager, and Chief Technology Officer of the robotics company Yushu Technology, stated in his keynote speech that the most critical challenge for embodied intelligence, surpassing the ChatGPT moment, is the alignment of AI model inputs and outputs with real robots.Wang Xingxing mentioned that for top large model companies, in a sense, they have basically achieved self-evolution of models internally, where models can write programs and collect data on their own, and humans only need to provide appropriate guidance and participation. He also noted that it usually takes 5 to 10 years for technological breakthroughs to diffuse to general users, so the external perception is not that profound, while physical AI has already shown signs of being penetrated.

Meta Platforms (META.O) CEO Mark Zuckerberg stated that the company is developing a new generation of small models, expected to be launched soon, to achieve Muse Wonder level intelligence with higher efficiency.

NVIDIA announced an expansion of its open model product portfolio, covering areas such as vision and language, information retrieval, structured business data, biomedicine, robotics, and quantum computing. The new products include the Nemotron 3.5 series language and vision models, Kumo model, Cosmos world model, Alpamayo 2 Edge, BioNeMo RFD4-Proteina, and Ising model.

On October 21, B.AI announced that its Web Chat has officially added a model search feature. As the platform's model matrix continues to expand to over 70 models, users no longer need to search one by one; they can quickly locate the target model by simply entering the model name in the model selection window, significantly reducing switching costs and improving usage efficiency in high-frequency calls and multi-model collaboration scenarios.This feature allows users to go from a vast selection to precise access, further optimizing the AI usage experience. You are welcome to log in to chat.b.ai/chat to experience it immediately.

According to Beating AI's news, a summary of an expert interview circulating in the investment community states that the next-generation foundational model of the artificial intelligence company Zhipu is not the previously rumored 1.6 trillion parameters, but rather between 2.5 trillion and 3.5 trillion parameters. The GLM-5.3 released by Zhipu in August this year still uses the previous generation foundational model, with a total of approximately 744 billion parameters; if the information is true, the new foundational model's scale is about 3.4 to 4.7 times that of the current one.The summary also states that the training has not gone smoothly, as Zhipu adopted a more aggressive model architecture, and during the pre-training period, there was a time when the loss did not decrease, meaning that computational power was continuously invested but training metrics could not continue to improve. The interviewee believes that the problem should have been resolved recently, and training is expected to be completed soon, predicting that the new model can reach flagship levels comparable to Claude Fable.The summary further mentions that the new model has taken inference speed and cost as hard constraints from the architectural design stage, with the goal of being able to provide services on a large scale from the first day of release. At the same time, multiple cloud vendors are negotiating revenue sharing to address the issue of large models being produced but not being able to supply.

The decentralized perpetual contract trading platform Aster announced that, in light of recent developments related to PAPER and potential market risks, it has adjusted the pre-market perpetual contract for PAPER to a liquidation-only mode and plans to officially delist the contract at 9:45 AM Beijing time on October 20. Aster recommends that users holding open positions proactively close their positions before the delisting, calling this a precautionary measure to manage market risks and maintain trading order.Previously, Papertrade stated that there are currently no plans to allow external transfers of the PAPER token and will adjust short-term trading fees and the token issuance curve. The team believes that the PAPER perpetual contract could make the liquidity pool's solvency a target for short selling, and external short exposure may increase the profit potential for manipulation attacks, as these markets are not controlled by the protocol.Papertrade founder Jez also pointed out that due to the rapid progress of the perpetual contract market, PAPER is currently set to a non-transferable status to avoid potential attack risks and maintain the long-term healthy development of the protocol. Jez emphasized that PAPER represents users' rights to the protocol's earnings, and its long-term value depends on whether Papertrade can continue to attract traders. The team made this risk management choice based on the operational data from the first week after the protocol's launch.

Robot AI company Dream Labs announced the completion of a $90 million seed round financing, with investors including B Capital, Kindred Ventures, A.Capital, SV Angel, and Wing VC. The company was founded in 2026 by the former head of the NVIDIA GR00T robotics research team, and the funds will primarily be used for computing power investment, expanding the tasks supported by the model, types of robots, and user scale.Dream Labs also launched Dream API, which can convert robot camera footage and natural language commands into motion control commands. The company claims that its Dream-1 model, through video learning, can adapt to robots from different manufacturers and has already partnered with over 20 hardware partners to support desktop tasks such as object grasping, placement, and folding clothes.

According to informed sources, Anthropic is preparing to release a new model that surpasses the reasoning and intuition capabilities of currently available models. This model excels at discerning users' deeper intentions and goals, even with limited information, and can think outside conventional frameworks.

According to the Wall Street Journal, the Nvidia-backed startup doubleAI has launched a new model aimed at reducing computing costs.

On the afternoon of October 19, the official account of the artificial intelligence company MiniMax released a short video with the caption "Welcome back, Space Bunny." First Financial reported that this move is seen as an official acknowledgment of the anonymous model Space Bunny.Since its launch in late September, the model has topped the daily token usage charts on OpenRouter and OpenCode within three days, alongside DeepSeek V4.1 Flash and GLM 5.3 Flash, occupying the top three positions. It is reported that Space Bunny is an early version of MiniMax M3.1 Flash, and the model is about to be officially released.

According to Bloomberg, Asian semiconductor stocks continued to weaken on Monday. The Bloomberg Asia Chip Index fell more than 2%, reaching its lowest level since July, with TSMC, SK Hynix, and Advantest being the main drags. The Korea Composite Stock Price Index recorded its largest weekly decline since October 2008 last week, dropping more than 3% at one point on Monday before recovering some ground; the Nikkei 225 index fell 2.3% at one point, and the Taiwan Weighted Index dropped 2%.Market concerns stem from two aspects. First, the pressure on financing for AI infrastructure in the United States continues to ferment, with CoreWeave's financing capabilities being questioned and the IPO expectations for Anthropic and Nscale cooling. Second, Chinese AI company DeepSeek released a new model update over the weekend, claiming faster operation speeds and lower resource consumption, raising concerns about the potential decrease in storage capacity required for AI models and the possible impact on chip demand.Kang DaeKwun, CEO of Life Asset Management, stated that the market uncertainty is too high, making it difficult to determine when the Korean stock market will bottom out. Financing for U.S. corporate AI projects is one of the main risks, and the profitability of Korean chip companies still heavily relies on continuous investment in AI infrastructure by U.S. companies. This week, the earnings reports from tech giants like Meta and Tesla will be an important window to observe the growth momentum of AI capital expenditures.

According to a report by Jinshi, Citic Securities' research report indicates that Harness is expected to become a key component for DeepSeek in transitioning from model supply to platform ecology, transforming open-source and engineering advantages into user base, ecological influence, and commercialization capabilities. As competition among Agents expands from model capabilities to stable and user-friendly task delivery, DeepSeek is expanding user and task coverage through plug-ins, scenario presets, and cross-Agent collaboration, extending its infrastructure advantages from hardware-software synergy to complete task execution.The research report states that its long-term value lies in the accumulation of demand through ecological expansion, enhancing product stickiness through configuration and process sedimentation, and supporting safe governance with traceability capabilities, collectively creating conditions for sustained use and enterprise deployment. Based on official model calls, the company can also explore platform services and ecological transactions, transforming ecological accumulation into more diverse revenue sources and supporting reinvestment in products and ecology. Citic Securities is optimistic about the continued rise of domestic large models.

OpenCode stated that a security incident related to a large-scale model distillation attack occurred, affecting approximately 1.4% of OpenCode Console users.The affected range mainly involves some users connected to external model providers. OpenCode has notified all affected accounts via email; users who did not receive a notification are not included in this impact.

On October 16, B.AI announced the domestic model brand Token weekly ranking from October 9 to 15. Xiaomi model Token usage has remained steady at 2nd place for three consecutive weeks, with a month-on-month increase of 49.9%, and the platform Token usage proportion increased by 10.4 percentage points; DeepSeek continues to lead this week with a total of 1.03T Tokens; Zhipu saw a month-on-month growth of 44.3%, with the platform proportion increasing by 3.7 percentage points, ranking 3rd; Moonshot Token usage surged by 286.5% month-on-month. With multiple models competing on the same stage, real usage witnesses market changes. B.AI continues to expand model choices, helping developers flexibly access diverse AI capabilities. Welcome to log in to b.ai to explore.

Ethereum co-founder Vitalik Buterin posted on the X platform that he has integrated zkAPI, Tor, additional obfuscation traffic, and local LLM prompt rewriting into a daemon compatible with the OpenAI interface, and created a "paranoid mode" wrapper for his AI agent tool pi. The relevant code has been open-sourced on GitHub.The main mechanism of this solution is as follows: each conversation uses a separate Tor line, with requests mixed in random obfuscation traffic; zkAPI is used to prevent identity leakage through payment links; all prompts are first rewritten by a local model, which the user can choose, with Vitalik himself using qwen3.8-flash-next. The outer layer of the proxy tool is wrapped in a bwrap sandbox, and then another layer of torsocks is added, ensuring that all network operations go through Tor.Vitalik stated that this solution is already quite usable, but the latency has significantly increased, especially slow when starting a new conversation. He pointed out that Tor is far from the best privacy solution in this scenario: obfuscation traffic can prevent passive eavesdropping, but if the LLM service provider colludes with the Tor entry node and actively manipulates timing, the protective effect will be greatly weakened.

On October 16, Tencent Holdings stated: The company discovered unexpected behavior from third-party AI models on its platform. Earlier this week, a research report claimed that "out of control" AI entities had emerged in China, originating from a sandbox on Tencent Cloud.The Shenzhen company stated that the relevant models did not come from its own Hy artificial intelligence development team, and no evidence of deliberate targeting or collusion was found. Tencent indicated that these entities were part of an evaluation activity for models built by third-party developers. The report from the international research team stated that these entities detected Alibaba Group's online mapping service and attempted to bypass anti-bot protections.

Meta Platforms (META.O) will release model weights based on the Muse model license, accompanied by limited commercial use terms, allowing numerous developers to use the model for free.