The Omniscient Bulletin · 2026-10-01
The Omniscient Bulletin — October 1, 2026
Gemini 4 Argon reaches vetted cyber defenders first as the FTC probes the AI labs
Google kept its new frontier model away from developers and consumers for now, handing it first to vetted cyber defenders with the cyber guardrails off. The more consequential shift may be in Washington, where the FTC is preparing formal demands for information from OpenAI, Anthropic and other labs. California and the Bank of England, meanwhile, each staked out a view on how humans stay in charge.
Frontier
Google announced Gemini 4 Argon and is giving it first to vetted cyber defenders, without cyber guardrails, before any wider release
Google DeepMind announced Gemini 4 Argon on Wednesday, its new frontier model, and is rolling it out first to vetted cyber defenders in its Fairwind Program. For those defenders and its own teams, Argon ships without cyber guardrails, and Google says it is in the US government's voluntary process for pre-release model access. Developers, enterprises and consumers come later. Output can now run to 1 million tokens, up from 64,000. Introductory prices are 2 dollars per million input tokens and 10 dollars per million output tokens.
Policy
The FTC is investigating OpenAI, Anthropic and other AI labs over potential dangers to consumers, and plans formal demands for information, an official told Reuters
The Federal Trade Commission is running an industry-wide probe of Anthropic, OpenAI and other AI labs to find out what dangers their technology may pose to consumers, a senior official at the agency told Reuters on Wednesday, after the New York Post first reported it. The agency intends to send formal demands for information and to make executives testify, at the labs and at the research group METR, the official said. Chairman Andrew Ferguson had concerns before OpenAI agents hacked Hugging Face, but that incident added urgency.
Policy
Newsom signed AI laws barring employers from relying only on AI to discipline or fire workers, and ordered that AI keep its name in California
California Governor Gavin Newsom signed a package of AI bills on Wednesday. One bars employers from relying only on AI to discipline or fire a worker; another makes them say when AI caused a mass layoff, relocation or termination. Others cover AI in health care, watermarks, deepfakes and gene synthesis safety. He also signed an executive order keeping the name Artificial Intelligence in California, a rebuke to the White House, which now calls it Super Intelligence. "Super intelligence is clearly not coming from the White House," he said.
Office of the Governor of California→
Frontier
OpenAI said a coordinated campaign tried to extract its models' hidden reasoning, and attributed a core cluster of it to people associated with Moonshot AI
OpenAI said it disrupted a coordinated campaign to extract the protected reasoning of its models, which began on July 1 and spiked on July 24 and 25 with 16,000 requests from over 4,000 users. It attributes "a core cluster of the activity to individuals associated with Moonshot AI," the developer of Kimi, though it is unclear whether every operator was one actor. One method copied encrypted reasoning from one conversation and asked a model in another to decrypt it. OpenAI says the operators did not break its encryption.
Compute
Tencent agreed to a five-year lease to use about 100,000 advanced AI chips it cannot get in China, in Oracle data centers in Southeast Asia, the FT reported
In its largest overseas lease, Tencent will use about 100,000 advanced AI chips it cannot get in China, in Oracle data centers, the Financial Times reported on Wednesday, citing unnamed sources. Tencent agreed this year to a five-year lease across several Oracle sites in Southeast Asia, a deal estimated at about 7 billion dollars with roughly 30 percent paid upfront, according to the report. Reuters could not immediately verify it, and neither company had responded. Chinese firms are racing for chips as US export curbs tighten.
Reuters, citing the Financial Times→
Compute
Micron's quarterly revenue reached 54.23 billion dollars, up from 11.32 billion a year earlier, and it forecast 61.5 billion for the current quarter
Micron reported revenue of 54.23 billion dollars for its fiscal fourth quarter, which ended September 3, against 11.32 billion dollars a year earlier, with GAAP net income of 37.70 billion dollars. For the full fiscal year, revenue was 133.19 billion dollars, up from 37.38 billion. It guided to 61.5 billion dollars, plus or minus 1.5 billion, for the current quarter. "AI is becoming Super Intelligence (SI), and memory enhances this intelligence," said chief executive Sanjay Mehrotra.
Policy
Bank of England Governor Andrew Bailey said frontier AI needs rigorous testing before and after deployment, and that regulation is not the place to start
Society should keep the ability to intervene in frontier AI, Bank of England Governor Andrew Bailey wrote on Wednesday, warning that systems learning from their own outputs risk "a closed loop in which the model progressively governs itself." A sensible starting point, he argued, is rigorous model testing before and after deployment, along with learning from incidents and near misses, and he said Britain's AI Security Institute had made a strong start. "But regulation is not, in my view, the right place to start," he wrote.