AI Daily brief · Sunday, 8 November 2026
STARHOT
Industry
QbitAIAnthropic's three internal models exposed, strongest Model 4's AECI exceeds Opus5.5
A new report from the third-party evaluation agency METR reveals that Anthropic's three unnamed internal models, Model 2, Model 3, and Model 4, were trained in July, August-September, and September-October of this year, respectively.
QbitAI
AGENTS.md exposed as a new attack entry point, capable of hijacking Codex and Claude Code
Security agency Forever Security disclosed an attack technique: web pages silently download malicious AGENTS.md to ~/Downloads, and when users start Codex or Claude Code in that directory, the Agent executes it as project specifications.
ITHome · AIAnthropic must continue to defend against the Claude Max plan usage misleading lawsuit, as the judge did not adopt its arguments
U.S. federal magistrate judge Xie Wenhui ruled that Anthropic must continue to defend against the main claim that the Claude Max plan usage is misleading to consumers, and dismissed several other claims in the case. The plaintiff claims that the $100 and $200 per month Max plans, when calculated on a per-week basis, actually provide only 3.5 times and 6 times the usage of the Pro plan, respectively, rather than the advertised 5 times and 20 times.
Research papers
Anthropic ResearchOfficial
Project EVE: Claude trading and teaming up in EVE Online and EVE Frontier.
Anthropic announced Project EVE research, allowing Claude to trade in the EVE Online simulated market and complete tasks in a four-person team in EVE Frontier.
Anthropic ResearchOfficial
Anthropic disassembles the composition of intelligent agent memory and the dreaming mechanism.
Anthropic has published a research paper, breaking down Claude's intelligent agent memory into components such as memory notes, offline dreaming organization, and memory hierarchy improvement, and explaining that memory refers to context that is preserved across sessions and can be read again, excluding weight updates, context learning within a single session, and compaction.
Tips & opinions
Tomasz TunguzThe cost of AI breaking into corporate networks has decreased by 99% in five months.
Tomasz Tunguz points out that the cost of AI breaching corporate networks has decreased by 99% in five months, and corporate security budgets are accelerating to catch up.
More worth reading
The DecoderVals AI benchmark shows the gap in network attack capabilities between open-source and closed-source models continues to narrowVals AI's CyberOffense benchmark allows AI agents to autonomously attack simulated enterprise networks on isolated Kali Linux machines, GPT-6 Sol Daybreak in OpenAI restricted projects completes 80.8% of the target, open-source weighted model Deepseek V4.1 Flash 1017 reaches 72.8%, with a single run cost of about 3 dollars instead of about 22 dollars
Hacker News · 300+Claude Code's Don't Ask Again option leads to passwords leaking into 1,122 public GitHub filesApiiro's research indicates that Claude Code's Don't Ask Again option makes passwords and other sensitive information appear in 1,122 public GitHub files. The article received a score of 582 and 394 comments on Hacker News.
36Kr · AIGoogle released the Gemini Omni 1.2 Flash video generation model, increasing the duration of a single segment to 15 seconds.Google has launched the AI video generation model Gemini Omni 1.2 Flash, natively supporting the longest 15-second video, with a duration limit increased by 50% compared to the previous generation's 10 seconds. Users can set the generation duration from 3 seconds to 15 seconds in 1-second increments.- ITHome · AIApple reveals the humanoid robot system HALO: using the Qwen model to validate on the Yushu G1, achieving an average success rate of 91% for single tasks.Apple's research team reveals humanoid robot autonomous control system HALO, with an average success rate of 91% in 360 single task tests in 6 real environments; the system calls the Qianwen model for verification on the Yushu G1, achieving an average success rate of 70% in 10 tests of 3 combinations in a single long sequence compound task, and a success rate of 50% in the most complex task.
- QbitAIHow does AFK AI Comma achieve continuous context and multi-Agent parallel collaboration?
TechCrunch · AIOpenAI hires Senator Schumer's daughter and seeks to mend ties with the Democratic PartyOpenAI has hired Senator Schumer's daughter and calls its data center a 'compute park'. After the midterm elections, tech giants are trying to repair their relationship with the Democratic Party to deal with stronger AI regulation, but more bargaining may occur at the state government level and on specific issues such as data center location.
