AI Daily News — 2026-10-09 (Fri)

AI news analysis report (last 12 hours)

Overview

OpenAI's annualized sales were reported to be about $50 billion, about $20 billion lower than previously reported, putting a damper on expectations for AI profits. Meanwhile, Google introduced a resident "Gemini agent" with its own email and calendar, further accelerating agent competition. AI-driven restructuring at companies such as Amazon, HSBC, and FICO is occurring one after another, and the impact on employment is becoming a reality.

Anthropic / Claude trends

Anthropic has committed $150 million to AI science initiatives backed by the White House. CNN reported on the company's "Claude-led biological discoveries," but scientists are divided. The site's usage policy, which prohibits "abusive or cruel behavior" towards Claude, also became a hot topic. According to Bloomberg, investors looking to IPO are struggling to price the risk of runaway AI. It has also been reported that the company plans to invest $518 billion in computing resources, highlighting its large scale and cash needs.

OpenAI trends

CNBC and the Financial Times reported annualized sales of about $50 billion, $20 billion less than Signal. This raises questions about the justification for the huge investment. USA Today sued OpenAI for copyright infringement of training data. Meanwhile, on the official blog, Oracle introduced a case study in which Oracle reduced work from days to minutes using ChatGPT and Codex, and a video showed a visual answer to GPT-6 using "Intelligent UI." The appeal of enterprise demand continues.

Google AI trends

Google Cloud has announced a resident "Gemini agent" that executes long-running tasks. It has dedicated Gmail, Calendar, and Drive areas, and also handles code creation and task execution. It is positioned as a countermeasure against OpenAI and Microsoft. Customer adoption continues, with The RealReal using Gemini Enterprise to scale its AI shopping agent. For developers, image search using EmbeddingGemma 2 was also introduced.

Trends of other AI companies

- xAI/Grok: SpaceXAI was hit with a biometrics-related lawsuit over images uploaded to Grok. Meanwhile, Grok also incorporates Suno, Claude, and Midjourney to manage Shopify stores. - Microsoft: Copilot can now work with local files on Windows. It also announced the Surface Laptop Ultra and RTX Spark. In Belgium, there have been complaints about Copilot's cross-selling. - DeepSeek related: Microsoft has added NVIDIA and DeepSeek models to Windows PC. Xiaomi has promoted Luo Fuli from DeepSeek, and the competition for talent is intensifying. - Meta: There was an article discussing the impact of open weight AI, but there were no new announcements of note.

Statements and trends of AI celebrities

- Sam Altman (OpenAI CEO): He drew criticism for saying that ``the world should accept that some bad things will happen'' because of the benefits of AI. - Elon Musk (xAI Founder/Tesla CEO): announced that he would use Suno's music model for Grok. Reaffirming control over Terafab AI chip facility, Intel also remains involved. He is expected to receive the Medal of Science from President Trump. - Jensen Huang (NVIDIA CEO): Dismissed AI's ``apocalyptic narrative'' as ``irresponsible'' and said that success requires a willingness to struggle. - Dario Amodei (Anthropic CEO): Although he did not make a direct statement, he was reported in connection with the company's $518 billion computing resource plan and the AP-NORC poll (many Americans perceive AI development to be too fast). - Geoffrey Hinton (AI Godfather/Nobel Laureate): Called for FDA-type safety approval before releasing models, saying public pressure like climate change drives action. - Mark Zuckerberg (Meta CEO): CNN reported on his ability to engage Mr. Trump and get his intentions across on AI policy. It has also been pointed out that smart glasses have issues. - Demis Hassabis (Google DeepMind CEO): An article introducing the person has been published. I cannot confirm his new statements. - Ilya Sutskever (SSI Co-Founder): There is no information about him, and his name was only mentioned in a report related to the movie ``Artificial,'' which depicts OpenAI. - Sundar Pichai (Google/Alphabet CEO), Yann LeCun (Meta): Specific statements and actions cannot be confirmed.

Breakthrough and innovation in AI technology

Goodfire announced that it can detect runaway agents at low cost with an "inside-out" monitor that monitors internal conditions. In a local environment, it is said that a PC equipped with RTX Spark can run a model with 284B parameters, and some claim that it outperforms GPT-5 in coding and inference (benchmark verification is required). EmbeddingGemma 2's image search and artificial life research also became a hot topic. Scientific substantiation remains a challenge for the biological discoveries led by Claude.

AI Regulation/Policy/Social Impact

While Hinkton called for FDA-style pre-approval, The Economist argued that liability laws alone cannot control AI. SemiAnalysis analyzed that China has adopted an AI safety system that prioritizes speed and is unable to keep pace with the frontier. In the US, an AP-NORC survey found that many people see AI progressing too quickly. Trust in corporate agents has also declined, with the number of companies entrusting them with production changes dropping from 75% to 56%. On the employment front, AI-induced layoffs were widespread, including at Amazon (approximately 1,000 people), HoneyBook (14%), FICO (15%), and HSBC (up to 70% of wealthy advisors in the UK). In the EU, there have been complaints about Copilot's cross-selling, and legal issues surrounding model distillation have also emerged.

Integrated analysis: What is happening now

Chart of power: OpenAI's downward revision in sales is a sign that the first mover's advantage is waning. On the other hand, Google is introducing a resident agent centered on Workspace, and Microsoft is also strengthening Copilot and local execution on the device side. Anthropic will increase its presence with science and government collaboration, but its $518 billion computational resources plan and IPO risk assessment demonstrate the severity of its capital needs. Grok's use of another company's model, Claude, is evidence of the growing interdependence of model layers.

Technology Trend: The focus has shifted from chat to autonomous agents with their own accounts and privileges. At the same time, agent monitoring (Goodfire, Darwinium) and fraud detection and reliability are emerging as demands. The decline in corporate trust (75% → 56%) indicates tension with expanded adoption. I can also see a trend toward moving large models locally.

Investment/Employment: It is clear that increased AI investment will lead to restructuring as a source of cost reduction funds. The gap between Bezos' "three-day work week" theory and the actual reductions symbolizes the disconnect between narrative and reality. Office jobs in finance, retail and healthcare are among the first to be affected.

Geopolitics: The United States deepens its honeymoon with AI promotion and industry, and China adopts a security system that prioritizes speed. The competition for talent (Xiaomi) and the issue of distillation have become new points of contention in the US-China conflict.

Future attention points

Aftermath of OpenAI's earnings review: Focus on how it will affect funding, computational resource contracts, and IPO-related valuations. 2. Gemini agent adoption status and security: The focus will be on the enterprise battle with Microsoft and OpenAI and the risk of accidents in agent privilege management. 3. Debate on pre-approval regulations and AI-induced job insecurity: We need to see whether Hinton et al.'s proposals will be reflected in policy and how the growing public opinion will affect regulations.