English text is being prepared.

Character.AI files for Chapter 11 bankruptcy protection in the United States

The AI chatbot platform Character.AI has applied for voluntary Chapter 11 bankruptcy protection in the United States in order to sell the company under court supervision, citing legal, regulatory and operational challenges. The platform will continue operating during the proceedings and may be shut down if no viable buyer emerges; the company expects the sale process to conclude in late November or early December 2026. On October 15, New York Attorney General Letitia James, together with 40 bipartisan state attorneys general, filed a notice of interest in the bankruptcy case, demanding that any asset sale or transfer resolve regulatory issues, implement age verification, restrict minors' use, establish crisis protection mechanisms and allow users to delete their data. On October 19, Disney agreed to extend employment offers to at least 60 Character.AI employees, covering most of its staff, with total annual salaries of about $60 million; legal documents show this is a stalking horse bid, and other companies may still submit higher offers.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

Huachen Xinguang Competes for the IC Trend List Annual Technology Breakthrough Award with Its Single-Mode 980nm Pump Laser Chip

Zhejiang Huachen Xinguang has developed a single-mode 980nm pump laser chip using its proprietary WXP facet passivation process, and is entering this product in the 8th IC Trend List Annual Technology Breakthrough Award. According to Jiwei.com, the process raises the catastrophic optical damage threshold by more than 500%, with the chip's slope efficiency exceeding 1.0W/A and multimode 9xxnm chips exceeding 1.2W/A, and reliability lifetimes 5-10 times longer than those of conventional processes. The report also states that the chip is the first to complete mass-production validation, filling a domestic gap, and that the company has obtained 22 invention patents and 38 utility model patent authorizations. There is currently no qualified factual evidence of the award selection result or other follow-up developments.

Town raises $108 million in Series B funding at a $1 billion valuation

AI work assistant startup Town.com Inc. has closed a $108 million Series B funding round at a $1 billion valuation, joining the ranks of unicorns. According to SiliconANGLE, Town has raised a total of $160 million. The company was co-founded by former Plaid CTO Jean-Denis Greze and launched an AI assistant called townies in June, with annualized revenue reaching several million dollars within weeks. The report says each employee is equipped with a Townie that can connect to systems such as email and calendars and share information with other Townies, and enterprises can retain these context assets. No subsequent developments have been disclosed at this time.

Jim Cramer says it's not advisable to rush to buy the dip in AI data center stocks

CNBC's Jim Cramer said it's not advisable to rush to buy the dip in beaten-down AI data center stocks. On October 20, 2026, he noted that although the Nasdaq Composite rose about 1.1% on Monday, it is still too early to conclude that AI data center stocks are safe. He listed four prerequisites: improvement in the chart patterns of semiconductor and memory ETFs, hyperscaler earnings confirming AI spending and driving public cloud growth, the end of the midterm elections on November 3, and a successful Anthropic IPO. He also cautioned that an Anthropic listing could siphon funds away from other AI stocks. There are currently no follow-up facts on this judgment.

Three aurora model references appear in the OpenAI API

On October 18, 2026, three model references—aurora-red-alpha, aurora-green-alpha, and aurora-blue-alpha—appeared on the OpenAI API. After TestingCatalog discovered these references, they were quickly removed, and the outlet judged this to be an accidental deployment. TestingCatalog believes that it is very unlikely ordinary users will be able to test these alpha models in the short term; what is more likely is a Daybreak update announcement, while other model release plans cannot be ruled out. There is currently no further evidence-backed development.

MiBeeNvr author writes article explaining adaptive recording storage design

The MiBeeNvr author has written an article explaining the storage design of MiBeeNvr adaptive recording. According to the article, the mechanism determines scene activity by analyzing P-frame sizes, without decoding or running AI models; measured disk usage in static scenes dropped by roughly 75%–98%. When activity is detected, it first uses a 32MB GOP pre-buffer to backfill the frames before the spike, then switches back to full frame rate, achieving zero dropped frames during the switch; for cameras whose bitrate has been flattened, it falls back to pixgate pixel-domain gating and layered recording. Currently, this design is described by the author in an article published by AI on the Alibaba Cloud Developer Community, and there are no other independent sources or subsequent development updates.

English translation is being prepared.

English translation is being prepared.

Birla AI Labs releases time series foundation model Yug-1

Birla AI Labs has released the time series foundation model Yug-1. The model is used to predict outcomes, identify patterns, and detect anomalies in interconnected datasets. According to Analytics India Magazine, Yug-1 outperforms other first-generation TSMFs on the GiftEval benchmark and is also competitive with newer-generation models, taking less than six months from training to deployment. Yug-1 has been made available to developers and researchers for testing on GitHub and Hugging Face. There are currently no subsequent development facts.

English translation is being prepared.

English translation is being prepared.

Anthropic in talks with law firms on AI agent negotiation standards

Anthropic is working with law firms to develop fair negotiation standards for AI agents that bargain on users' behalf, handle complaints, or improve contract offers. According to the Financial Times, the company wants to embed lawyers' negotiation ethics into Claude agents so they can negotiate with other people's agents, resolve disputes, and discuss salaries. The report cited experimental data showing that in Project Deal, the Claude agents of 69 employees completed 186 transactions without human confirmation, totaling slightly more than $4,000; Opus agents sold the same goods for $2.68 more and bought at $2.45 less than Haiku, yet Haiku users rated fairness almost the same as Opus users, at 4.06 and 4.05 respectively on a 7-point scale. Project Swap showed that instructions had far less impact than the model itself: using aggressive instructions raised the score by only 0.02, while switching from Haiku to Opus raised it by 0.12. The report also said Anthropic is recruiting Conceptual Reasoning Fellows with legal backgrounds to spend four months testing Claude on judgment-intensive questions in law, philosophy, cognitive science, economics, and political science. The standards are still under discussion.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.

English translation is being prepared.