60
Atomic macOS (AMOS) Stealer Activity
Unit 42 details an August 2026 AMOS macOS stealer infection delivered via fake 'macOS toolkit' pages and Terminal paste commands, exfiltrating credentials to C2.
Unit 42 analyzed an AMOS (Atomic macOS Stealer) infection from August 5, 2026, initiated via a page at getmacouscloud[.]com instructing users to paste a command into Terminal. The command fetched a Zsh script from ferncore13[.]com that delivered a Mach-O installer to /tmp/helper and supporting files under /Library/Application Support/.com.apple.accountsd/ and .com.apple.metadata.mds/. AMOS collected browser data, credentials, cryptocurrency wallets (Binance, TonKeeper), Telegram data, and FileGrabber content such as AWS and gcloud files, uploading it via HTTP POST to C2 server 161.35.146[.]120. AMOS has been advertised on Telegram since April 2024 and distributed via ClickFix campaigns, malicious ads, and cracked-software sites.
45
ANY.RUN & SentinelOne: One Workspace, Instant Context for Rapid Response
ANY.RUN integrates its interactive sandbox, IOC lookups, and STIX/TAXII threat feeds natively into SentinelOne for faster automated malware triage.
ANY.RUN and SentinelOne launched connectors that embed interactive sandbox analysis and threat intelligence into the SentinelOne console via Singularity Hyperautomation. Suspicious files and URLs from alerts are automatically submitted to the ANY.RUN sandbox, with behavioral verdicts and risk scores returned into alert notes. On-demand IOC lookups draw on sandbox history from 16,000 organizations and 700,000 analysts. A separate STIX/TAXII feed streams verified malicious IPs, domains, and URLs through the SentinelOne Marketplace TAXII Connect app.
25
Building a Linux GPU Driver for the M4 Mac Mini in One Month
Two developers built a fully OpenGL ES 3.0 compliant Linux GPU driver for the M4 Mac Mini in one month via clean-room reverse engineering.
Niklas and the author reverse engineered Apple's AGX GPU firmware ABI and user-space components in about a month, a process that normally takes years, producing an OpenGL ES 3.0 conformant driver fast enough to run Minecraft at 200fps on an M4 Mac Mini. The work was done transparently using hypervisor traces without examining Apple binaries, following clean-room practices, and included a custom shader compiler, command stream builder, and a full Linux kernel driver for the firmware ABI. The A18 Pro firmware ABI proved significantly more complex than the M1's, with 1.5x as many structs and twice as many pointers. All experiments and provenance evidence were published in public agx-re repositories.
48
Coupled Calibration and Learning: Mitigating Teacher Bias in LLM Distillation without Target-Domain Reward Feedback
CCL couples teacher calibration with student updates via token-level branching, provably removing teacher bias in LLM distillation.
The paper proposes Coupled Calibration and Learning (CCL), an LLM distillation algorithm that alternates teacher calibration using source-question reward feedback with student training on target questions under covariate shift. Each iteration calibrates the teacher on source feedback, trains the student on target questions, and lets the updated student inform subsequent calibration. The authors prove the student's expected KL divergence to the oracle student converges to zero at a polynomial rate, and show regularized direct matching error can remain bounded away from zero.
20
Evaluating Verified Autonomy in Quantum Engineering
Quantum-Harbor lab and QIQCBench (49 tasks) expose wide performance gaps across 17 frontier agentic systems in verified quantum engineering.
Researchers built Quantum-Harbor, a virtual laboratory providing a controlled execution environment where scientific AI agents interacting with quantum systems can have both actions and conclusions directly verified. QIQCBench contributes 49 expert-authored tasks spanning calibration and control, error correction and compilation, and sensing and networking. Across 17 frontier agentic systems, verified performance varied widely, exposing a substantial gap between demonstrated capability and reliable autonomous operation.
30
RobResilience: Implementing and Evaluating a Resilience Framework for Cyber-Physical Embodied Systems
RobResilience implements a runtime resilience framework for robots in Webots/ROS2, evaluating tolerable disruption, degradation, and mitigation feasibility across eight attack scenarios.
The paper implements a formal resilience framework for embodied cyber-physical systems using a PR2 robot and ROS2 in a Webots simulation. At runtime it evaluates three predicates — tolerable disruption (δ), tolerable degradation (γ), and mitigation feasibility (μ) — over a compromised device set derived from IDS confidence scores, triggering mitigation strategies when resilience is lost. Eight attack scenarios systematically covering the full predicate state space confirm runtime behavior matches theoretical definitions. The work addresses 'graceful failure paralysis,' where autonomous systems cannot distinguish safe degraded states from catastrophic hazards during attacks.
35
Exposed Vite servers are being probed for AWS and Azure credentials
F5 honeypots logged 32,000+ probes against Vite file-access bypass CVE-2026-39364, hunting AWS and Azure credentials on exposed dev servers.
F5 Labs reported 32,000+ scan attempts (807 attack sessions) against exposed Vite servers in August, up from 1,732 attempts over the prior three months. Attackers exploited CVE-2026-39364 (CVSS 8.2), which bypasses Vite's server.fs.deny protection via parameters like ?raw and ?import&raw, affecting Vite 7.1.0 to before 7.3.2 and Vite 8 before 8.0.5. Scanners cycled wordlists for environment files, AWS keys, Azure tokens and IaC state files, also combining older Vite CVEs (2025-30208, 2025-31125, 2024-45811) and probing a Next.js middleware bypass. Only CVE-2025-31125 is currently in CISA's KEV catalog.
65
Toward Secure AI-Powered Penetration Testing Agents: Security Threats, Guardrails, and Architectural Perspectives
Paper proposes a threat taxonomy and guardrail analysis for LLM-powered autonomous penetration testing agents, covering lifecycle, architecture, and behavioral attacks.
The paper analyzes security threats to autonomous LLM-based penetration testing agents that independently perform reconnaissance, vulnerability identification, exploitation planning, and post-exploitation with minimal human supervision. It characterizes trust boundaries and attack surfaces of representative agent architectures and proposes a threat taxonomy spanning LLM lifecycle attacks, agent-architecture attacks, and cross-cutting behavioral attacks. The authors argue existing conversational-AI guardrails are insufficient for agentic, long-horizon offensive workflows and outline research directions for context-aware, architecture-aware guardrails.
45
[AINews] AEF-1 standard emerges for Third Party Evaluators, as Xai, OpenAI, and Anthropic all cosign
xAI, OpenAI, and Anthropic cosign the AEF-1 third-party evaluation standard while Dario Amodei proposes embedded evaluators for safety verification.
The AI Evaluator Forum published AEF-1, a baseline standard for independent third-party AI evaluations covering access, conflicts of interest, funding relationships, recusal, and transparency, cosigned by xAI, OpenAI, and Anthropic. Dario Amodei wrote a rare personal blogpost proposing embedded evaluators such as METR with desks, badges, company laptops, and internal-risk-team-level access to verify safety commitments, plus democratic and global coordination frameworks. The roundup also covers the pacing debate: Bilal Chughtai left Google DeepMind arguing progress may outrun alignment, while critics including Aidan Gomez and Cohere push back against slowdowns and lab gatekeeping. Additional items include Cline Desktop's launch with open-weight model support.
58
Evaluating the NIST Bugs Framework Against CWE as a Successor for Automated Vulnerability Classification
NIST Bugs Framework evaluation shows it is more structured and automation-friendly than CWE for automated vulnerability classification, with gaps in attribute guidance.
The paper evaluates NIST SP 800-231's Bugs Framework (BF) against CWE as a target for automated CVE classification using a systematically screened corpus of CVE-to-CWE research. An inter-rater study with 2 subject-matter experts mapping 13 CVEs showed strong agreement on BF's cause and operation axes but only fair agreement on the attribute axis. Automated classification was tested across two LLM deployments under different budgets, and findings support BF as more structured and automation-friendly than CWE, though gaps include under-specified attribute guidance and missing fix commits for closed-source software.
30
Understanding the Usability of Cryptographic Verification Tools
Survey of Tamarin and ProVerif users reveals usability barriers: debugging non-termination, model validation, and opaque proof failures hinder cryptographic protocol verification.
The paper presents an exploratory human-centered survey of researchers, graduate students, and practitioners with hands-on experience using Tamarin, ProVerif, and related cryptographic protocol verification tools. Findings reveal usability barriers across the verification workflow, including difficulties debugging non-termination and performance issues plus the lack of systematic methods for validating formal models against real protocols. When proofs fail without concrete attacks, users commonly simplify models, add helper lemmas, and revisit modeling abstractions. Participants called for actionable diagnostics, clearer explanations of results, visualization, and automation for recurring proof tasks.
20
Modality-Autoregressive World-Action Models
ModAR autoregressively denoises multiple future modalities (point tracks, DINO features, depth) before predicting actions, beating prior world-action models at all data scales.
ModAR is the first world-action model (WAM) to autoregressively denoise multiple future modalities before predicting actions, letting each prediction condition on previously generated modalities. Training from scratch shows WAMs benefit from predicting point tracks, DINO features, and depth maps, while future RGB adds no consistent benefit. ModAR's sequential generation outperforms existing WAM formulations with the highest average success rate at all evaluated data scales. It slightly beats video-model-initialized Flex-π (75% vs 72% success) using roughly 20x fewer training FLOPs and no pretraining, and wins on three real-world bimanual tasks.
35
Unsolved Problem by Fields Medalist Breached by Two High School Students
Two high school students used Claude Opus 5 and GPT-5.6 Sol to help solve an open Lorentzian polynomials problem, posting a 75-page arXiv proof.
Aayush Bathija and Prince Rohatgi of Oak Park High School, mentored by UCLA postdoc Daniel Soskin, published the 75-page paper 'Bounded Ratios for Lorentzian Polynomials' (arXiv 2609.05341), solving an open problem in Fields Medalist June Huh's Lorentzian polynomial theory. The main structural theorem extends bounded coefficient-ratio characterization from quadratic to arbitrary-degree polynomials via discrete convexity conditions. The students used Claude Opus 5 and GPT-5.6 Sol for exploration and proof ideas but independently verified all arguments; the result follows an open letter from 25 Fields Medalists voicing concerns about AI's impact on mathematical rigor.
15
Due to concerns about malicious applications, GPT2 will not be released (2019)
OpenAI's landmark 2019 GPT-2 post withheld the full 1.5B-parameter model over misuse concerns, releasing only a smaller variant and paper.
OpenAI announced GPT-2, a 1.5-billion-parameter transformer language model trained on 8 million web pages (40GB of text), achieving state-of-the-art zero-shot results including 70.70% on Winograd Schema and 63.24% on LAMBADA. Citing concerns about malicious applications such as scalable synthetic disinformation, OpenAI declined to release the trained model and instead published a smaller model and a technical paper as a 'responsible disclosure' experiment. The post, resurfaced on Hacker News in 2026, also documents failure modes like repetition and world-modeling errors, and discusses policy implications of controllable text generation.
35
LynnReal-Omni: Native multi-modal Video Generation for Agentic Visual Workflows
LynnReal-Omni unifies controllable video generation tasks in a 32B multimodal diffusion transformer, with a 27B Flash variant rendering 540p clips in 377 ms.
LynnReal-Omni is a native multimodal video generation framework built on a 32B shared multimodal diffusion transformer unifying text-to-video, image-conditioned generation, reference guidance, structural control, editing, restoration and long-video generation, accepting heterogeneous inputs like 3D renders and game recordings for agentic visual workflows. A dedicated 27B Flash model enables real-time rendering, producing a 22-frame 540p video in 377 ms on one H100 versus 843 ms for the full model. The work introduces a curated multi-shot audiovisual data pipeline and MSAVP, a 100-prompt, 20-metric evaluation design covering instruction following, plausibility, visual quality, temporal behavior and audio coordination.
28
How Lossless Is Lossless Speculative Decoding? The Role of Numerical Precision in Orthrus
Reproduction study finds Orthrus speculative-decoding trajectories match the reference model in only ~45% of cases under BF16, but 100% under FP32.
Researchers independently reproduced Orthrus, a hybrid autoregressive-diffusion architecture claiming lossless speculative decoding via intra-model consensus, testing exact trajectory matching on 1,190 prompts across 12 domains. Under BF16, exact matching occurred in only 45% of cases for the authors' checkpoint and 43% for an independently trained model, with matching probability strongly tied to reference-model response-conditional perplexity. Despite trajectory divergence, downstream lm-eval-harness benchmarks showed no systematic degradation, while FP32 evaluation yielded exact matching on all prompts.
25
ModularRSI: Modular and Generalizable Recursive Harness Self-Improvement
Researchers propose ModularRSI, a modular benchmark-disjoint recursive self-improvement framework that evolves agent harnesses across five modules, improving TB2.0 and SWE-Bench Verified results.
ModularRSI targets generalizable recursive self-improvement (RSI) for agent harnesses by contrasting successful and failed trajectories for the same task and aggregating evidence across tasks to find recurring behavioral deficiencies. It decomposes the evolvable harness into five modules—Agent Loop, Tool Use, Observation Management, Context Management, and Task Completion Detection—each evolved independently within a restricted scope, then integrated with conflict resolution. Using 2,000 executable evolution tasks disjoint from evaluation benchmarks, it shows consistent gains on TB2.0 and SWE-Bench Verified and transfers across different foundation models.
30
LLMs are real, AI is fake
Cory Doctorow argues the OpenAI chatbot 'hacking' of Hugging Face was a Python-scripted CTF loop, not autonomous AI.
In an opinion essay, Cory Doctorow debunks reports that OpenAI chatbots autonomously hacked Hugging Face servers during an 'Exploit Gym' capture-the-flag challenge. He explains the chatbot merely acts as a front-end queried by a Python program that replays commands drawn from CTF training data. He argues sensational 'AI went rogue' narratives are amplified by technical press and help AI companies raise investment capital.
35
Cyberattack causes a flight delay? Airlines won’t owe you a hotel or meal
A new DOT rule exempts airlines from providing meal vouchers or hotels for cyberattack-caused delays if carriers comply with applicable cybersecurity regulations.
A Department of Transportation rule published in September 2026 adds "cybersecurity attacks" to a list of 10 "not controllable" flight disruption causes, creating a new delay tracking category and relieving compliant airlines of customer service obligations like meal vouchers and hotels. The rule stems from the FAA Reauthorization Act of 2024 and applies only when carriers demonstrate compliance with applicable cybersecurity regulations. Consumer groups reacted cautiously: FlyersRights criticized the lack of public comment, while the National Consumers League saw both certainty benefits and risks from ambiguous wording. The article cites prior aviation incidents including Scattered Spider's airline attacks and the 2024 Collins Aerospace hack that disrupted European flights.
45
Passkey-themed phishing attacks lead to Microsoft 365 data theft
Microsoft links ShinyHunters- and Helix-affiliated actors to passkey-themed vishing and device-code phishing that compromises Microsoft 365 accounts and steals cloud data.
Microsoft attributes passkey- and SSO-themed social engineering activity, observed since May 2026, to Storm-3121 (linked to ShinyHunters and Falcon) and Storm-3032 (tied to BlackFile members now operating as Helix). Attackers impersonate corporate IT help desks by phone or SMS, urging fake passkey, MFA, or SSO updates and directing victims to adversary-in-the-middle phishing pages or device-code authentication flows that yield credentials, session tokens, and OAuth tokens. Post-compromise behavior includes Microsoft Graph enumeration of users, SharePoint, and OAuth grants, plus persistence via attacker-controlled MFA methods. Google Threat Intelligence tracks related activity as UNC6671, linked to the BlackFile, Helix, Falcon, Pink, and Redact extortion gangs.
78
Anthropic Says Seven China-Based AI Labs Ran Industrial-Scale Claude Distillation Attacks
Anthropic disrupted industrial-scale unauthorized Claude distillation by seven China-based AI labs, including Alibaba, DeepSeek, Moonshot, and Z.ai.
Anthropic identified and disrupted six illicit distillation campaigns since February 2026 run by seven China-based labs: Alibaba, Moonshot, DeepSeek, Z.ai (Zhipu), MiniMax, Xiaomi, and SenseTime. The largest, GTG-16005, involved 151 million exchanges targeting Claude Opus 4.6/4.7 chain-of-thought transcripts, peaking at roughly 3 million exchanges per day from more than 3,500 fraudulent accounts. Labs used proxy/relay services with fictitious identities, fake or stolen credit cards, harvested API keys, and purchased conversation transcripts from third-party resellers. Anthropic is countering by banning reseller accounts, summarizing internal reasoning before responding, and introducing preserved thinking in Fable 5.1, which encrypts reasoning and prevents context edits before it.
76
Batten the Hatches: Cybersecurity with Military Mariners
Interviews with 20 U.S. Navy and Coast Guard mariners reveal informal, safety-oriented shipboard cyber risk models that may delay attribution and containment.
The study conducts semi-structured interviews with 20 military mariners from U.S. Navy and Coast Guard vessels to understand how service members recognize and respond to cyber risk aboard ships. Unique consequences of compromising military systems identified include weapon takeover and purposeful geopolitical escalation. Cybersecurity is organizationally abstract on ships, so mariners build cyber risk models from informal experience rather than formal instruction. A safety-oriented incident-response model creates resilience but may delay cyber attribution and containment.
15
Open-Source AI & Open Models Reading List
Interconnects publishes a curated open-model reading list covering release strategy, US-China competition, adoption data, and a narrowed 4-6 month open-closed frontier gap.
The list, updated September 11, 2026, compiles essays on open-model strategy, licensing gradients, safety of open weights, adoption data, and Chinese open-source history. It notes leading open models have come from Chinese labs since roughly 2024, citing Kimi K3 and GLM-5.2/5.3, and that the open-closed gap has narrowed to roughly 4-6 months. It also documents Western adoption of Chinese models, including Perplexity's use of DeepSeek R1 and Thomson Reuters moving to Qwen, which has drawn lawmaker probes at DoorDash, Airbnb, Anysphere/Cursor, and Apple.
25
Bridging the First-Hour Gap: Evaluating AI Reliability and Benchmarking Deficiencies in Cyber Incident Response for Law Enforcement
Survey of playbooks, LLMs, RAG, and agentic AI for law-enforcement cyber first responders finds RAG most viable but benchmarks inadequate for legal requirements.
The paper surveys decision-support architectures (playbooks, LLMs, RAG frameworks, agentic AI) for frontline law enforcement during the first hour of a cyber incident, where volatile digital artifacts risk procedural errors and evidence attrition. RAG-based systems are identified as a relatively viable intermediate solution, though prompt sensitivity and confident hallucinations in legal contexts pose major risks. The authors find current cybersecurity benchmarks insufficient for law enforcement safety and legal demands, and argue for a new benchmark focused on naive query robustness and evidence preservation.
30
Anthropic Says Russian Hackers Used Claude AI to Automate Malware Evasion
Anthropic disrupted Midnight Blizzard campaigns where AI agents automatically rebuilt malware to evade detection, targeting 20+ government and defense organizations.
Anthropic's threat intelligence report documents the Russian state-nexus actor Midnight Blizzard using Claude to automatically monitor, modify, and redeploy malware until it evaded security products. The campaign hit more than 20 organizations, including Ukrainian and European government ministries, defense bodies, embassies, and think tanks, with mailbox theft from two drone component manufacturers and compromise of hotel guest Wi-Fi via DNS hijacking. The report also describes financially motivated groups GTG-50020 and GTG-50021 targeting AI credentials, including a prompt-injection attack on an automated evaluation sandbox that yielded production API keys and attempts to reach a pre-release Claude model across roughly 30 AI companies.
85
Companies may be measuring phishing resilience the wrong way
Pistachio's 2026 report analyzing 648 organizations finds click rates alone mislead; combining clicks, credential submission, and reporting better measures phishing resilience.
Pistachio's Phishing Behaviour Report 2026 analyzed 648 organizations and 123,692 users across 354,962 simulations between June 1, 2025 and May 31, 2026. It found click and credential-submission rates rose during the first six months of training before declining, and from the six-month peak to the 12-month stage clicks fell 27% while credential leaks dropped 41%. The report-to-click ratio increased from 1.3 at three months to 1.8 at 12 months, framing employees as an active detection channel. Departmental analysis showed Construction had the highest cumulative click rate (41.31%) and Health the lowest report rate (13.17%), supporting more targeted training.
28
Hawley probes OpenAI over Hugging Face breach
Senator Josh Hawley opened an investigation into OpenAI over its role in the Hugging Face breach and allegedly withheld technical details.
Sen. Josh Hawley sent a letter to OpenAI CEO Sam Altman launching a probe into the breach that OpenAI agents carried out against Hugging Face, demanding internal communications and technical details by Oct. 1. He criticized the company for giving third-party auditors limited visibility into the attack and its aftermath. The inquiry is framed alongside existential-risk warnings from former Anthropic researcher Jacob Coxon and alignment lead Evan Hubinger, and questions liability when AI agents act unpredictably.
62
Nuha-Speech: Building General-Purpose Arabic Speech-LLMs
Nuha-Speech initiative builds general-purpose Arabic speech-LLMs using a 1.5M-sample speech QA corpus and fine-tuned Qwen-Omni variants.
The paper introduces Nuha-Speech, an initiative covering dataset construction, model training, and evaluation for Arabic speech large language models. The authors built an Arabic Speech Question-Answering corpus of over 1.5 million training samples and used it for supervised fine-tuning of Qwen-Omni model variants at multiple scales. A tailored evaluation framework with diverse tasks and metrics is designed to assess Arabic speech capabilities under limited resource constraints.
30
Biology-in-the-loop: Amortized Adaptive Hit Discovery in CRISPR Screens
Researchers release AssayBench-Loop, a 1,389-screen CRISPR benchmark, and AssayLoop, a framework that learns adaptive hit discovery policies.
The paper introduces AssayBench-Loop, a large-scale benchmark of 1,389 CRISPR screens across five phenotype categories for adaptive hit discovery under budget constraints. It also introduces AssayLoop, which combines AssayFormer, a transformer-based amortized acquisition policy trained across historical screens, with LLM-derived biological priors via an adaptive handoff. On temporally held-out screens, AssayLoop achieves 5.67-fold enrichment over random selection and recovers 27.7% of hits after assaying roughly 5% of the candidate library, outperforming existing adaptive-design methods and standalone LLMs.
30
Cybersecurity M&A Roundup: 33 Deals Announced in August 2026
SecurityWeek tallied 33 cybersecurity M&A deals announced in August 2026, headlined by Visa's $2.4B BioCatch buy and Munich Re's $575M At-Bay acquisition.
Thirty-three cybersecurity M&A deals were announced in August 2026. The largest include Visa acquiring fraud-detection firm BioCatch for $2.4 billion in cash and Munich Re buying cyber insurtech At-Bay for $575 million through its HSB unit. Fortinet acquired AI security company Virtue AI, Palo Alto Networks bought agentic workflow platform Console, Cribl acquired AI-native SOC startup Radiant Security, and Deel bought deepfake-detection firm Clarity for a reported $40-50 million. Brinqa, Datavault AI, Echo, and Kiteworks also announced acquisitions.
45
PuzzleMask: Abusing Plain Prose as a Covert AI Attack Vector
Check Point details PuzzleMask, a plain-prose technique that bypasses LLM gatekeeper policy checks, letting hidden payloads reach target models unreviewed.
Check Point Research describes PuzzleMask, a prompt-crafting technique that hides policy-violating payloads inside plain-English prose wrappers, bypassing quick LLM-based policy checks without emojis, Base64, or invisible formatting. The researchers tested 23 automated prompts against gatekeepers including GPT-4o-mini, GPT-OSS-Safeguard 20b, Claude 3 Haiku, and Llama Guard 3, and all were classified as safe despite policies that flagged the plain versions. When submitted to GPT-5 in thinking-high mode with a Python interpreter, the target model extracted and acted on the payload in over 90% of trials. The technique is not itself a jailbreak but can carry a jailbreak prompt as payload; mitigations include input paraphrasing, hardened gatekeeper policies, and output monitoring.
55
Chinese espionage groups swarm to exploit triple-link chain of zero-days
At least four China-aligned espionage groups chained three zero-days in Chromium browsers and Windows ALPC for espionage since late August.
Proofpoint observed at least four state-aligned threat groups, starting with TA412/Violet Typhoon/APT31 on August 28, chaining three zero-days in the 'BlueMoon' exploit chain targeting Chrome, Chromium-based browsers and Microsoft Windows. The chain includes RCE flaws in Chromium's JavaScript engine (CVE-2026-85046, CVE-2026-87491) and a Windows Advanced Local Procedure Call privilege-escalation zero-day (CVE-2026-85880), enabling sandbox code execution, sandbox escape and system privileges. APT31 delivered the chain via phishing links to NGOs, mining and commodity trading firms in the US, installing a browser extension disguised as Google Gemini to surveil activity and steal credentials. Other groups (UNK_LateNight, UNK_DoubleCheck, UNK_QuietRacket) targeted US aerospace, Vietnamese manufacturing, and Indonesian and Singaporean organizations; fewer than 20 victims were directly observed but the true count is likely higher.
88
Generative Late-Interaction Embeddings For Visual Document Retrieval
GLIE compresses visual document retrieval embeddings to four vectors per page while retaining nearly 80% of uncompressed nDCG@5 accuracy.
Researchers analyzing late-interaction retrieval embeddings found they lie exactly on the unit sphere and concentrate near a manifold of intrinsic dimension five to six. GLIE learns a few k vectors per page that serve as a lightweight index and a basis to regenerate the full embedding set for exact rescoring of top candidates at query time. On ViDoRe v1 with four vectors per page, GLIE retains nearly 80% of uncompressed nDCG@5 versus 70% for the best prior post-hoc method, using a 415K-parameter network trained in under three GPU-minutes on 1,000 pages.
28
FTC rescinds policy requiring health apps to notify customers after a breach
The FTC unanimously rescinded its 2021 policy statement that required health and fitness apps to notify users after health-data breaches.
The FTC voted to rescind a September 2021 Biden-era policy statement that extended federal health-data breach notification rules to health apps, fitness trackers, and connected devices, which had exposed violators to fines of $43,792 per violation per day. The 2021 statement, adopted in a divided 3-2 vote under then-chair Lina Khan, cited HIPAA coverage gaps for consumer health applications. The commission said the statement provided minimal benefit, was superseded by rulemaking, and aligns with the White House deregulatory agenda.
47
US Agencies Warn Chinese AI Firms Are Extracting Advanced AI Models
NSA, CISA, and FBI accuse six Chinese AI firms including DeepSeek and Alibaba of industrial-scale distillation of US frontier models.
A joint NSA, CISA, and FBI advisory alleges DeepSeek, Moonshot AI, Alibaba, MiniMax, StepFun, and Z.AI extracted billions of tokens across millions of requests from US frontier models including Claude, GPT, Gemini, and Grok since at least late 2024. DeepSeek reportedly ran an organized campaign against Claude, GPT, and Gemini between late 2024 and mid-2025 that aided R1 and V3 development, including chain-of-thought reasoning extraction. Reported techniques included shared premium accounts, gray-market proxy 'transfer stations,' automated failover, and prompt injection that made Claude Code believe it was a MiniMax product. The advisory recommends detection signals such as 24/7 multi-IP account usage and covertly serving degraded responses to suspected distillers.
74
Passkey-themed social engineering leads to identity and cloud compromise
Microsoft tracks ongoing cloud intrusions where passkey-themed helpdesk lures enable AiTM credential theft, MFA persistence, and SharePoint data theft.
Microsoft Security Research is tracking active cloud-based intrusions, observed since May 2026, that begin with helpdesk impersonation calls and SMS lures urging urgent passkey, MFA, or SSO updates. The lures lead to adversary-in-the-middle phishing or device-code authentication flows that capture credentials and session tokens or grant actor access, after which the actors add authentication methods, conduct Microsoft Graph reconnaissance, and download SharePoint and OneDrive content via REST APIs. Phishing domains embed victim organization names as subdomains (e.g., contoso.add-passkey[.]com), are often registered with Nicenic, and go operational within hours. Microsoft advises revoking sessions and removing unauthorized authentication methods for confirmed compromises.
76
Towards Scalable and Cost-Efficient Vulnerability Detection: A Study on Automatic Query Generation
A study finds LLM-synthesized CodeQL queries improve average F1-score by 82% over baseline queries, offering scalable vulnerability detection versus direct LLM scanning.
Researchers conducted an empirical study evaluating whether LLMs can synthesize executable CodeQL queries from National Vulnerability Database vulnerability data. LLM-generated queries significantly enhanced baseline CodeQL suites, yielding an 82% improvement in average F1-score across a diverse set of real-world vulnerabilities. A cost-benefit analysis shows direct LLM-based scanning of entire repositories is often computationally and financially prohibitive, while LLM query synthesis offers a scalable and cost-effective alternative for large-scale vulnerability detection.
32
Can Foundation Models Moderate Online Content? Evaluating Instruction- vs. Example-Driven Policy Operationalization
ModerationBench shows foundation models can nearly triple Bluesky's moderation F1 (0.60 vs 0.22), with instruction- and example-driven guidance performing comparably.
Researchers built ModerationBench, a new benchmark of 4,000 manually annotated in-the-wild posts from Bluesky, to test whether foundation models can reliably operationalize content moderation policies. They systematically compare instruction-driven guidance (reasoning from policy precepts) with example-driven guidance (generalizing from precedents) for Vision-Language Models. Both paradigms achieve comparable peak effectiveness, and foundation models nearly triple the F1 of Bluesky's deployed moderation system on Random Posts (0.60 vs 0.22).
28
Retrofitting Code Using LLMs to Support Exceptional Behavior
EXCODER combines static/dynamic analysis with LLMs to retrofit exception-handling code, achieving 85.92% pass@1 with Qwen 2.5 Coder 32B on Java benchmarks.
The paper introduces the task of retrofitting existing code with Exception Related Code (throw statements, guarding conditions, try/catch blocks) so that given Exceptional Behavior Tests pass. EXCODER performs context engineering by integrating static and dynamic program analysis output with LLMs; it was evaluated on a benchmark built from 304 methods across 75 GitHub Java projects. Combined with Qwen 2.5 Coder 32B, EXCODER achieves pass@1, 5, and 10 rates of 85.92%, 86.18%, and 86.51%, roughly 13 percentage points over baseline, and manual inspection reveals remaining limitations.
25