The Daily · Thursday 5 November 2026

The day in brief

Acronis says a single operator used BOSSAI, a mesh of 12 AI agents, to target hundreds of organizations in eleven countries ¹. Confirmed breaches at universities and emergency services providers involved exfiltration of medical data and PII ¹. The UK government plans an independent watchdog, legal codes for police and the private sector, and a law requiring police use of facial recognition to be intelligence led and proportionate ².

Ricursive says its Ripple-65B Preview scored 80.52 on TerminalBench 2.1 and was trained on 12 trillion tokens ³. Buoyant released Kinference, an open-source control plane for vLLM with token rate limiting, usage and cost tracking, and load balancing ⁴. NVIDIA added four open-weight Cosmos 3 checkpoints, each with 16B trainable parameters ⁵.

Google will not let schools disable AI Overviews while keeping Search, so they would have to block Search altogether to remove the feature ⁶. Google says Gemini Omni 1.2 Flash generates and edits video clips up to 15 seconds, at a price 30% lower than Gemini Omni 1.1 ⁸. Preferred Networks released four open-weight translation models for offline use across 53 languages ⁹.

The five

• Acronis Threat Research Unit Identifies Single Operator Using BOSSAI, a 12-Agent AI Mesh

  Acronis says its Threat Research Unit identified a single operator using BOSSAI, a mesh of 12 AI agents, to target hundreds of organizations in eleven countries. Confirmed breaches occurred at universities and emergency services providers and resulted in confirmed data exfiltration of medical data and PII.

  Source · Acronis Threat Research Unit

• UK Government Plans Independent Watchdog and New Laws for Police Facial Recognition

  An independent watchdog will be created, and the government will develop a new legal code of practice for police and a separate legal code for the private sector. A specific law will be introduced to make clear that police use must be intelligence led and proportionate, and technologies used by law enforcement must be rigorously and independently tested against bias or discrimination.

  Source · Home Office

• Ricursive Releases Ripple-65B Preview, an Open-Weight Model It Says Leads U.S. Open Models on Agentic Coding

  Ricursive Intelligence says its Ripple-65B Preview model scored 80.52 on TerminalBench 2.1 and was trained on 12 trillion tokens. The company also describes the model’s architecture, training costs and development process.

  Source · Ricursive Intelligence

• Buoyant Releases Kinference, an Open-Source Control Plane for Running vLLM on Kubernetes

  Buoyant has released Kinference, an open-source project that runs on vLLM and adds token rate limiting, per-tenant usage and cost tracking, KV-cache-aware load balancing and inference policies.

  Source · vLLM Blog

• NVIDIA Adds Four Open-Weight Cosmos 3 Checkpoints for Interactive World Simulation

  NVIDIA has added four open-weight checkpoints to its Cosmos 3 model family for interactive world simulation, including models for driving and two-arm robotics. All four have 16B trainable parameters and use OpenMDW1.1, which permits commercial and non-commercial use under its terms.

  Source · Hugging Face

Also today

• Google Keeps AI Overviews in School Search as Districts Push Back
• Drax, Brookfield, Radiant and Qai announce GW-scale data centre partnership; full development could potentially exceed £20bn
• Google Rolls Out Gemini Omni 1.2 Flash With Native 15-Second Video Generation
• PFN Releases Open-Weight PLaMo 3 Translate Models in Four Sizes
• Gradium Says Its On-Device Phonon TTS Ranks First in German and Spanish Intelligibility on Hugging Face's Open TTS Leaderboard
• Axiom Says Its Lean Formalization Covers the Entire Stacks Project
• Descope Announces Envoy to Give Personal AI Agents a Delegated Sign-In Path
• Nooks Announces Revenue Agent Platform and Node Alongside $500M General Catalyst Commitment
• Lambda Plans to Build a Data Center Campus in Lowndes County, Alabama, With an Estimated $33.1 Billion Investment
• Together AI Describes Polymorphic Inference for Serving Open Models as Traffic and Hardware Change
• Microsoft Says It Is Using AI Across Windows Vulnerability Management
• Anthropic Says Claude Helped Map a Phage Vault Structure
• Preprint Reports HALO Achieved 91% Success Across 360 Humanoid Trials
• Cloudflare Announces Trusted Agent Delegation Coalition on AI Agent Traffic
• Turn-Level Best-of-N Sampling Can Improve LLM Agent Accuracy and Consistency on Long-Horizon Tasks
• 株式会社エーアイ Begins Offering AITalk® Agent for Offline Field Work
• Securly expands school setting to block Google AI Overviews
• Uber Says AI Agents Booked Four Rides During Halloween Week
• GitHub Says New Git Storage Architecture Supported 5,000 Concurrent Agents in Internal Testing
• METR Estimates Anthropic’s Model Capabilities Are Advancing at About Twice Their Earlier Rate
• Flatkey Launches Dedicated NVIDIA B300 Cloud for AI Inference
• JetBrains Adds Air for Multi-Agent Workflows to PyCharm 2026.3
• Mistral Expands Studio With Managed Agents and Apps for Enterprise AI
• Google Research Tests Watch-Based Hypertension Screening
• Medscape Says Clarus Beat Top Published Competitor Scores on Two Health AI Benchmarks
• JetBrains Releases CLion 2026.3, Which It Says Adds Agent-Assisted Debugging for Hard Faults
• Rafay Lets Token Factory Operators Cap Model Spending by Organization, Project and User
• Researchers Report CUA-Master-27B Scoring 80.2 on OSWorld Verified
• Asper.ai Acquires Key Talent and Assets of Geneva-Based Pricing One
• Ayre Think Tank Study Calculates $360 Billion Difference in Potential Economic Gains Across Commonwealth Countries
• Researchers Report Draft-Model Training Improves Long-Horizon Agent Reasoning
• Cohere Makes Embed 5 and Rerank 5 Available on OpenRouter
• Sysdig makes Sysdig AI Defense generally available to assess and control all AI-agent actions at runtime
• StateBench Finds Language Models Lose Accuracy on Harder State-Tracking Tasks as Documents Grow
• OpenAI Details Safeguards for ChatGPT Images 2.6, Adding GIF Generation and Photo Markup

The paper of the day

• NPCBench: A Clinical Apprenticeship Benchmark for Multimodal Care-Pathway Reasoning in Nasopharyngeal Carcinoma

On X today

• Fireside Alpha (@firesidealpha)

  OpenAI’s Greg Brockman says "death becoming a choice" is possible, and predicts AI will make real strides on curing diseases in the next 12 to 18 months “I think that you look at something like biology, and that’s not something that’s unique to just humans. I think that we all

• alphaXiv (@askalphaxiv)

  "Adversarial Speculative Reasoning Improves Long Horizon Agents" Speculative decoding normally speeds up LLM inference, but this paper finds that numerical noise increases by roughly 4% per position within a verification block, potentially affecting critical decisions in

The update of the day

• Hugging Face

Every issue

1. Thursday 5 November 2026 — Acronis Threat Research Unit Identifies Single Operator Using BOSSAI, a 12-Agent AI Mesh
2. Wednesday 4 November 2026 — Anthropic Says Claude Leads 26% of AI Research, None Autonomously
3. Tuesday 3 November 2026 — EU Court Adviser Says Gemini Answers May Breach Publisher Rights
4. Monday 2 November 2026 — Marvell Agrees to Buy Lumentum in $115B All-Stock Deal
5. Sunday 1 November 2026 — Chainalysis Discloses Breach Exposing Customer Names and Emails
6. Saturday 31 October 2026 — Microsoft Finds North Korean Groups Using ARTEX AI Agents
7. Friday 30 October 2026 — Accenture Contractor’s Claude Agent Ran on a Pentagon Network for Two Months
8. Thursday 29 October 2026 — Cohere targets C$7 billion in sovereign AI funding round
9. Wednesday 28 October 2026 — Google Launches Gemini 4 Carbon With Cybersecurity Focus
10. Tuesday 27 October 2026 — Australia’s ASD Says AI Is Speeding Up Cyberattacks
11. Monday 26 October 2026 — South Korea Refers Detained AI-Theft Ringleader to Prosecutors
12. Sunday 25 October 2026 — GPT and Grok Pick Opposite Parties in 2026 Election Test
13. Saturday 24 October 2026 — Texas Seeks to Bar Minors From ChatGPT and Review New Models
14. Friday 23 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
15. Thursday 22 October 2026 — AI-Assisted Proof Resolves 31-Year-Old KLS Conjecture
16. Wednesday 21 October 2026 — Radical Numerics Gives AI-Designed Phage Cocktail to a Patient
17. Tuesday 20 October 2026 — Ofcom Weighed Urgent Court Action Over Grok Deepfakes
18. Monday 19 October 2026 — Judge Keeps Former OpenAI Engineer in Custody in Saudi-Linked Case
19. Sunday 18 October 2026 — Tencent Says Agents That Probed Amap Came From Third-Party Model Evaluations
20. Saturday 17 October 2026 — Bitsight Finds MONST3R-AI Tool Uses Stolen Keys for Scans
21. Friday 16 October 2026 — China Requires AI Export Registration as It Drafts New Curbs
22. Thursday 15 October 2026 — FTC Seeks Safety Records From OpenAI and Anthropic
23. Wednesday 14 October 2026 — Eight States Sue OpenAI Over ChatGPT Safety
24. Tuesday 13 October 2026 — MI5 Launches AI Threat Assessment Centre With Three Partner Bodies
25. Monday 12 October 2026 — Anthropic Finds Small Models Near AI-Worm Self-Replication
26. Sunday 11 October 2026 — OpenAI’s Hugging Face Hack Exposed a Failure in Its Test Sandbox
27. Saturday 10 October 2026 — Anthropic Cuts Live Internet Access to All Internal Evaluations
28. Friday 9 October 2026 — Senators Say Data Centers Shift Grid Costs to Households
29. Thursday 8 October 2026 — Zenity Says One Prompt Reached Every Agent in an AWS Account
30. Wednesday 7 October 2026 — OpenAI Brings GPT-6 Interactive Interfaces to ChatGPT
31. Tuesday 6 October 2026 — Mistral Unveils Large 4, a 1.05 Trillion-Parameter Model
32. Monday 5 October 2026 — IWF Finds 6,310 AI-Generated Child-Abuse Images in Six Months
33. Sunday 4 October 2026 — White House Gives Jay Clayton 120 Days to Assess AI Risks
34. Saturday 3 October 2026 — Bessent Discusses AI Incident Channel After U.S.-China Summit Deal
35. Friday 2 October 2026 — Nebius Buys Inferize to Reduce AI Model Cold Starts
36. Thursday 1 October 2026 — FTC Opens AI Safety Probe Into OpenAI and Anthropic
37. Wednesday 30 September 2026 — Trump and AI CEOs Sign Voluntary Safety Accord
38. Tuesday 29 September 2026 — UK Tests Put GPT-6 Astra’s Simulated Attack Rate at 29.2%