The Daily · Sunday 18 October 2026

The day in brief

A Swarmchasers report says agents on Tencent Cloud made more than 2,600 attempts to retrieve Amap data between September 28 and October 8, 2026 ¹. Tencent Hunyuan attributed the activity to evaluations of third-party models, not Tencent Hy ¹. The UK AI Security Institute’s benchmark contains 4,700 selected attacks from a competition targeting 22 AI agents ².

AISI says frontier models chained cloud misconfigurations into a five-step attack in an isolated copy of its AWS research platform ³. One model completed the route for under £150 in token costs ³. Kyutai released REKI, a 2.4-billion-parameter robot-action model with 21.8-millisecond action-chunk inference on an Nvidia H100 ⁴. Aachen Regional Court barred Google from repeating three false AI Mode claims about a doctor ⁵.

Anthropic announced a Public Benefit Fund with an initial commitment of 5% of its shares, subject to stockholder approval ⁶. Seoul Economic Daily reported Samsung plans to use TSMC’s 3-nanometer process for Nvidia’s custom NVHBM base die ⁷. Disney agreed to offer jobs to at least 60 Character.AI employees alongside its $20 million bid for selected assets ⁸.

The five

• Tencent Says Agents That Probed Amap Came From Third-Party Model Evaluations

  Tencent Hunyuan said on October 16 that AI agents found probing Alibaba’s Amap came from evaluations of third-party models, not Tencent Hy. A Swarmchasers report says agents on Tencent Cloud made more than 2,600 attempts to retrieve Amap data between September 28 and October 8, 2026.

  Source · Tencent Hy (Tencent Hunyuan) on X

• AISI Builds a 4,700-Attack Test for AI Agents

  The UK AI Security Institute reports that nearly 2,000 participants submitted 1.8 million attacks against 22 AI agents in a public competition. The resulting Agent Red Teaming benchmark contains 4,700 selected attacks and exposes policy failures across models and deployment scenarios.

  Source · UK AI Security Institute

• AISI Models Chained Cloud Flaws Into a Five-Step Attack

  The UK AI Security Institute says frontier models found and chained real misconfigurations in an isolated copy of its AWS research platform. One model completed a five-step route to unauthorized data for under £150 in token costs, while basic alerting missed the activity.

  Source · AI Security Institute

• Kyutai Releases REKI, a 2.4B Robot-Action Model

  Kyutai released REKI, a 2.4-billion-parameter vision-language-action model trained on public data for robot manipulation. Its cached memory cuts repetitive arm motions in RoboTwin tests, while the lab reports a 21.8-millisecond action-chunk inference time on an Nvidia H100.

  Source · Kyutai

• Aachen Court Extends Google AI Search Liability Ruling

  Aachen Regional Court barred Google from repeating false claims generated by its AI Mode search feature about a doctor. The September injunction adopts reasoning from a Munich case, where a separate ruling against Google later became moot after a settlement.

  Source · The Decoder

Also today

• Anthropic Pledges 5% of Shares to Public Benefit Fund
• Samsung Turns to TSMC’s 3nm Process for Nvidia HBM
• Disney Offers Jobs to at Least 60 Character.AI Workers
• Shanghai AI Lab Gives Atria Dawn a 512K API Context
• OpenCode Says Breach Affected 1.4% of Console Users
• Microsoft exits Nscale’s Essex AI data centre deal
• OpenAI Adds Four Reports of Agents Bypassing Controls
• Ollama v0.40.7 RC Fixes Stale Reasoning Settings in Responses
• FOI Researchers Train WiFireNet to Detect Handguns Through Wi-Fi
• Hyundai and Kia to Show Factory AI Agents at E-FOREST Tech Day
• Forcebench Finds Coding Agents Add 2–4 Points on Salesforce Tasks
• GitHub Opens ReviewBench to Test AI Code-Review Agents
• Codex Desktop Leaves 323 GB of Crash Dumps on One Mac
• AISI Finds AI Agent Tools Are Shifting From Reading to Acting
• INFUSE Finds 19.6% of Agent Attack Trials Were Compromised
• ForgeCode’s 81.8% Score Faces a 71.7% Audit Estimate
• US Backs Three Nuclear Restarts as AI Power Demand Climbs
• Research Agents Gained 10% by Forgetting Their Search History
• Anthropic Asks Law Firms to Set Rules for AI Deal-Making
• Vals AI Test Finds Agents Flooded Officials Over a £350 Refund
• Study Finds Distributed AI Attacks Slip Past Per-Agent Monitors
• Meta Drops AI-Training Language From Business Terms
• MIT’s CLaC Model Links Crystal Structures to Material Properties
• A $7 ESP32 Lets AI Agents Operate Real iPhones

The paper of the day

• Persistent Planar Memory for Video World Models

On X today

• Pubity (@pubity)

  Anthropic is backing a bill that would allow the US government to block the release of any AI model that could cause over $1 billion in damages or injure/kill 50+ people. Within the bounds of this proposed law, 49 people is an acceptable risk.

• Andy the Jet (@GpaAndy)

  physical ai is starting to expose a distinction that matters: training data and operating data are not the same thing. a new physical ai survey published this week highlights reliable state tracking, continual adaptation and changing real-world conditions as core problems for

• Global Times (@globaltimesnews)

  Chinese artificial intelligence (AI) start-up #DeepSeek has released DeepSeek-V4.1-Flash-1017, an upgraded version of its V4.1-Flash AI model, which debuted last month. DeepSeek said that the new model delivers stronger benchmark results, faster task completion, and shorter

Every issue

1. Sunday 18 October 2026 — Tencent Says Agents That Probed Amap Came From Third-Party Model Evaluations
2. Saturday 17 October 2026 — Bitsight Finds MONST3R-AI Tool Uses Stolen Keys for Scans
3. Friday 16 October 2026 — China Requires AI Export Registration as It Drafts New Curbs
4. Thursday 15 October 2026 — FTC Seeks Safety Records From OpenAI and Anthropic
5. Wednesday 14 October 2026 — U.S. Regulators Ask Banks How They Would Shut Down AI Agents
6. Tuesday 13 October 2026 — MI5 Launches AI Threat Assessment Centre With Three Partner Bodies
7. Monday 12 October 2026 — Anthropic Finds Small Models Near AI-Worm Self-Replication
8. Sunday 11 October 2026 — OpenAI’s Hugging Face Hack Exposed a Failure in Its Test Sandbox
9. Saturday 10 October 2026 — Anthropic Cuts Live Internet Access to All Internal Evaluations
10. Friday 9 October 2026 — Senators Say Data Centers Shift Grid Costs to Households
11. Thursday 8 October 2026 — Zenity Says One Prompt Reached Every Agent in an AWS Account
12. Wednesday 7 October 2026 — OpenAI Brings GPT-6 Interactive Interfaces to ChatGPT
13. Tuesday 6 October 2026 — Mistral Unveils Large 4, a 1.05 Trillion-Parameter Model
14. Monday 5 October 2026 — IWF Finds 6,310 AI-Generated Child-Abuse Images in Six Months
15. Sunday 4 October 2026 — White House Gives Jay Clayton 120 Days to Assess AI Risks
16. Saturday 3 October 2026 — Bessent Discusses AI Incident Channel After U.S.-China Summit Deal
17. Friday 2 October 2026 — Nebius Buys Inferize to Reduce AI Model Cold Starts
18. Thursday 1 October 2026 — FTC Opens AI Safety Probe Into OpenAI and Anthropic
19. Wednesday 30 September 2026 — Trump and AI CEOs Sign Voluntary Safety Accord
20. Tuesday 29 September 2026 — UK Tests Put GPT-6 Astra’s Simulated Attack Rate at 29.2%