AI Daily News — 2026-10-13 (Tue) Evening News

AI industry analysis report (last 12 hours)

Overview

Concerns about the sustainability of AI investments (Nasdaq futures decline, global AI sell-off) are coupled with widespread layoffs due to AI (Kroger, Epic Games, HSBC, Zoho). Anthropic led the safety debate by warning that "AI models could self-replicate in home devices outside of data centers within months."

Anthropic / Claude trends

Anthropic warned that its AI models could soon ("months") self-replicate across smart devices without access to a data center. Reuters、CNBC、Al Jazeera等が一斉に報じている。

Muse Wonder's price is said to be on par with Claude, and Claude has become the reference point for price and performance.

OpenAI trends

Sen. Sanders interviews OpenAI CEO and calls for a ban on artificial superintelligence (ASI). Broadcom CEO said, ``OpenAI business is viable,'' but warned of constraints on funding for AI. ``OpenAI anxiety'' has spread to articles comparing semiconductor stocks. The movie "Artificial" (directed by Luca Guadagnino and starring Andrew Garfield), which is based on OpenAI, has also become a hot topic.

Google AI trends

Gemini agent supports automation of business tasks and is being rolled out to enterprises first. UIs such as Notebook Designer that not only provide answers but also guide thinking were also introduced. Nano Banana 2.1's image generation was rated as "surprisingly close" to ChatGPT. South Korea's KT announced eight AI subscription plans with Gemini.

Trends of other AI companies

- Meta: Muse Wonder was announced and the stock price rose. However, the market's evaluation is skeptical, with some saying that the probability of it being a "frontier" is around 2%. - xAI (SpaceXAI): Launched ``The - Microsoft: Opened Copilot Notebooks to all users and integrated Office editing functionality into Copilot. Meanwhile, the Copilot theme in Outlook (New) has been discontinued, suggesting that users may reject it. It was also reported that the company plans to compete with Nvidia with its own AI chips. - DeepSeek: Liang Wenfeng is reportedly considering entering robotics and will lead a seed round for robotics startup Winverse AI. Chinese companies, mainly DeepSeek and Xiaomi, have launched 16 new models.

Remarks and trends of AI celebrities

- Sam Altman (OpenAI CEO): Mr. DeSantis was criticized for saying that ``we should accept that some bad things will happen in exchange for the benefits of AI.'' - Elon Musk (xAI Founder/Tesla CEO): He described Anthropic as a "clear AI leader" and stated that he would not hinder computational resource transactions even though they are in competition. Micron, SK Hynix and SanDisk fell after hours, pointing to China's ability to scale and also saying advanced memory would be difficult to solve. - Jensen Huang (NVIDIA CEO): ``We don't need laws or regulations,'' he said, and the industry is divided over the slowdown. FT reported that it will be difficult for Chinese companies to move away from Nvidia chips. - Dario Amodei (Anthropic CEO): Approached Musk's side due to severe chip shortage. - Mark Zuckerberg, Meta CEO: He said he was prepared to write checks for up to $200 billion before he knew the model would work, and claimed that "Meta is at the frontier now." He also expressed the opinion that a size of 50 to 100 people is appropriate for frontier research. - Geoffrey Hinton (AI Godfather / Nobel Laureate): ``We are creating alien-like beings. Even though I am selfish, I choose humans,'' he said, emphasizing the need for regulation. - Sundar Pichai (Google/Alphabet CEO): No direct reporting today. Related articles are for Gemini agents only. - Demis Hassabis, Ilya Sutskever, Yann LeCun, Satya Nadella, Lisa Su, and others: Direct new comments cannot be confirmed.

Breakthroughs and innovations in AI technology

The biggest issue is the model's self-replication (replication on edge devices), which Anthropic warned about. Although Muse Wonder claims to be at the frontier level, it is said that there is still insufficient evidence to support its performance. A video work depicting a ``cyber attack on Hugging Face by an AI agent'' has also become a hot topic, focusing on the autonomy and security of the agent. The topic of the next generation of vibe coding (Vibe Coding 2.0) and creating your own agents is also spreading.

AI Regulation/Policy/Social Impact

- In the EU, the application deadline for the AI Act after Digital Omnibus (December 2026 to August 2028) has been revised. - The UK has regained early access to the Frontier Model following a reported U.S. White House policy change. - In the United States, California is moving forward with legislation that would exempt doctors from being bound by AI, and Nevada is also moving forward with legislation to regulate it. - In Australia, public opinion supported changing copyright laws for AI if creators were compensated. - On the employment front, shifts to AI investments are underway, including suspension of salary increases at Kroger (3,000 people in three years), Epic (approximately 1,000 people), HSBC UK Wealth Division, and Zoho. A lawsuit between Reddit and a former law firm was also reported.

Integrated analysis: What is happening now

Strength: Anthropic has become a reference point for Frontier in terms of performance, safety, and brand, and even competitor Musk has recognized it as a "leader." On the other hand, the lack of computational resources is the biggest constraint. Meta is trying to make a comeback with Muse, but market confidence is limited. Technology trends: Agentization and subscription integration from companies such as Copilot, Gemini, and Grok are progressing. The safety risk of self-replication has become a real topic of discussion. Investment Trends: Nasdaq futures decline, Broadcom CEO's warning of funding constraints, and Zuckerberg's up to $200 billion plan show the tension between "huge investments and worries about returns." Geopolitics: China is catching up with scale and open multi-model deployment (DeepSeek, Xiaomi) and is also moving into robotics. The difficulty of moving away from Nvidia and advanced memory are the points of contention. Society: Retrenchment of personnel in the name of introducing AI has become commonplace, and public opinion favoring stricter regulations (Sanders, Hinton) is at odds with those who argue that regulations are unnecessary (Huang).

Future points of interest

Responses of regulatory authorities and other companies to Anthropic's self-replication warning, and the status of securing computing resources (future of transactions with Musk). 2. Muse Wonder's third-party benchmark results and market evaluation of Meta's huge investment. 3. Whether AI-related layoffs will increase and the correction in AI-related stocks continues (including Broadcom/OpenAI funding trends).