Chinese AI • Open Source • Security • Incidents. Signal, not noise.
RSAC researchers hacked Apple Intelligence using the Neural Exect method and Unicode manipulation. The post Apple Intelligence AI Guardrails Bypassed in New Attack appeared first on SecurityWeek.
Thursday. Another week, another batch of things that probably should've been caught sooner but weren't. This one's got some range — old vulnerabilities getting new life, a few "why was that even possible" moments, attackers leaning on platforms and tools you'd normally trust without...
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Is fake grass a bad idea? The AstroTurf wars are far from over.  In...
As AI tools become more accessible, employees are adopting them without formal approval from IT and security teams. While these tools may boost productivity, automate tasks, or fill gaps in existing workflows, they also operate outside the visibility of security teams, bypassing controls...
Threat actors have been exploiting a previously unknown zero-day vulnerability in Adobe Reader using maliciously crafted PDF documents since at least December 2025. The finding, detailed by EXPMON's Haifei Li, has been described as a highly-sophisticated PDF exploit. The artifact...
An apparent hack-for-hire campaign likely orchestrated by a threat actor with suspected ties to the Indian government targeted journalists, activists, and government officials across the Middle East and North Africa (MENA), according to findings from Access Now, Lookout, and...
Apr 9, 10:15 UTCResolved - This incident has been resolved. Thank you for your patience and understanding as we addressed this issue. A detailed root cause analysis will be shared as soon as it is available.Apr 9, 10:15 UTCMonitoring - The degradation has been mitigated. We are monitoring to ensure...
A rare warm spell in January melted enough snow to uncover Cornell University’s newest athletic field, built for field hockey. Months before, it was a meadow teeming with birds and bugs; now it’s more than an acre of synthetic turf roughly the color of the felt on a pool table, almost digital in...
When I started digging into desalination technology for a new story, I couldn’t help but obsess over the numbers. I’d known on some level that desalination—pulling salt out of seawater to produce fresh water—was an increasingly important technology, especially in water-stressed regions including...
Apr 9, 08:53 UTCResolved - This incident has been resolved.Apr 9, 08:07 UTCIdentified - The issue has been identified and a fix is being implemented.
Reputable researcher Haifei Li has come across what appears to be a PDF designed to exploit an unpatched vulnerability. The post Adobe Reader Zero-Day Exploited for Months: Researcher appeared first on SecurityWeek.
In December 2025, hackers stole names and passport numbers from the European travel company’s network. The post 300,000 People Impacted by Eurail Data Breach appeared first on SecurityWeek.
MiniMax’s MMX-CLI enables AI agents to run full multimodal workflows autonomously with optimized, machine-friendly design.
ByteDance’s new Seeduplex model brings real-time, full-duplex voice interaction to Doubao, making AI calls more natural and responsive.
Manycore Tech kicks off its Hong Kong IPO with strong cornerstone backing, poised to become the first listed firm among Hangzhou’s “Six Little Dragons.”
Alibaba and Unitree are reportedly teaming up on a strategic global push, with AliExpress set to play a central role.
neueHCT’s single-chip urban NOA solution wins major automaker deals, signaling a shift toward affordable, full-scenario intelligent driving.
A hacker transferred more than 50 bitcoin from the Bitcoin ATM operator’s wallets after stealing credentials. The post $3.6 Million Stolen in Bitcoin Depot Hack appeared first on SecurityWeek.
Apr 9, 04:57 UTCResolved - This incident has been resolved. Thank you for your patience and understanding as we addressed this issue. A detailed root cause analysis will be shared as soon as it is available.Apr 9, 04:57 UTCMonitoring - The degradation has been mitigated. We are monitoring to ensure...
arXiv:2604.05136v1 Announce Type: new Abstract: Fuzzy Cognitive Maps constitute a neuro-symbolic paradigm for modeling complex dynamic systems, widely adopted for their inherent interpretability and recurrent inference capabilities. However, the standard FCM formulation, characterized by scalar...
arXiv:2604.05142v1 Announce Type: new Abstract: As artificial intelligence systems (AIs) become increasingly produced by recursive self-improvement, a form of evolution may emerge, in which the traits of AI systems are shaped by the success of earlier AIs in designing and propagating their...
arXiv:2604.05157v1 Announce Type: new Abstract: Computer-Use Agents (CUAs) leverage large language models to execute GUI operations on desktop environments, yet they generate actions without evaluating action quality, leading to irreversible errors that cascade through subsequent steps. We propose...
arXiv:2604.05162v1 Announce Type: new Abstract: Reconfigurable Intelligent Surfaces (RIS) are pivotal for next-generation smart radio environments, yet their practical deployment is severely bottlenecked by the intractable computational overhead of Channel State Information (CSI) estimation. To...
arXiv:2604.05165v1 Announce Type: new Abstract: Reconfigurable Intelligent Surfaces (RIS) has a potential to engineer smart radio environments for next-generation millimeter-wave (mmWave) networks. However, the prohibitive computational overhead of Channel State Information (CSI) estimation and the...
arXiv:2604.06227v1 Announce Type: new Abstract: Accurate short-term forecasting of agricultural commodity prices is critical for food security planning and smallholder income stabilisation in developing economies, yet machine-learning-ready datasets for this purpose remain scarce in South Asia....
arXiv:2604.06228v1 Announce Type: new Abstract: We introduce probabilistic language tries (PLTs), a unified representation that makes explicit the prefix structure implicitly defined by any generative model over sequences. By assigning to each outgoing edge the conditional probability of the...
arXiv:2604.06253v1 Announce Type: new Abstract: Cross-lingual code generation is critical in enterprise environments where multiple programming languages coexist. However, fine-tuning large language models (LLMs) individually for each language is computationally prohibitive. This paper investigates...
arXiv:2604.06256v1 Announce Type: new Abstract: Training dynamics during grokking concentrate along a small number of dominant update directions -- the spectral edge -- which reliably distinguishes grokking from non-grokking regimes. We show that standard mechanistic interpretability tools (head...
arXiv:2604.06260v1 Announce Type: new Abstract: Test-time scaling investigates whether a fixed diffusion language model (DLM) can generate better outputs when given more inference compute, without additional training. However, naive best-of-$K$ sampling is fundamentally limited because it...
arXiv:2604.06171v1 Announce Type: new Abstract: Communications networks now form the backbone of our digital world, with fast and reliable connectivity. However, even with appropriate redundancy and failover mechanisms, it is difficult to guarantee "five 9s" (99.999 %) reliability, requiring rapid...
arXiv:2604.06192v1 Announce Type: new Abstract: Recent work uses entropy-based signals at multiple representation levels to study reasoning in large language models, but the field remains largely empirical. A central unresolved puzzle is why internal entropy dynamics, defined under the predictive...
arXiv:2604.06193v1 Announce Type: new Abstract: Depression is underdiagnosed in primary care, yet timely identification remains critical. Recorded clinical encounters, increasingly common with digital scribing technologies, present an opportunity to detect depression from naturalistic dialogue. We...
arXiv:2604.06195v1 Announce Type: new Abstract: Large language models often produce unsupported claims. We frame this as a misclassification error at the output boundary, where internally generated completions are emitted as if they were grounded in evidence. This motivates a composite intervention...
arXiv:2604.06196v1 Announce Type: new Abstract: Three-way logical question answering (QA) assigns $True/False/Unknown$ to a hypothesis $H$ given a premise set $S$. While modern large language models (LLMs) can be accurate on isolated examples, we identify two recurring failure modes in 3-way logic...
Hackers vowed to revive its efforts against America when the time was right — demonstrating how digital warfare has become ingrained in military conflict. The post Shaky Ceasefire Unlikely to Stop Cyberattacks From Iran-Linked Hackers for Long appeared first on SecurityWeek.
Apr 8, 23:57 UTCResolved - Some users experienced an error when trying to access Claude.ai on desktop and the web from 23:22–23:50 UTC. This issue has now been resolved.
a quiet day lets us reflect on MSL finally shipping!
Apr 8, 19:40 UTCResolved - The issue affecting API workspace creation has been resolved. Impact occurred from approximately 4:15 PM to 7:30 PM UTC.Apr 8, 19:28 UTCMonitoring - A fix has been deployed for the issue affecting API workspace creation and we are monitoring for recovery.Apr 8, 19:22...
Apr 8, 19:30 UTCResolved - Between 19:47 and 21:57UTC on April 8, 2026, Cloudflare identified and mitigated a network performance issue specifically affecting our Miami and Atlanta regions. Users may have experienced increased latency or intermittent connectivity due to degraded routing paths in...
Cybersecurity researchers have flagged a new variant ofmalware called Chaosthat'scapable of hitting misconfigured cloud deployments, marking an expansion of the botnet's targeting infrastructure. "Chaos malware is increasingly targeting misconfigured cloud deployments, expanding...
Apr 8, 17:50 UTCResolved - From 17:25–17:44 UTC, we saw elevated errors on attempts to log into Claude.ai (mobile, web, and desktop) and platform.claude.com, as well as /login attempts via Claude Code. This issue is now resolved.Apr 8, 17:46 UTCInvestigating - We are currently investigating an...
Status: ResolvedAll impacted services have now fully recovered.Affected components Login (Operational) Login (Operational)
We present GenLCA, a diffusion-based generative model for generating and editing photorealistic full-body avatars from text and image inputs. The generated avatars are faithful to the inputs, while supporting high-fidelity facial and full-body animations. The core idea is a novel paradigm that...
Diffusion models have driven remarkable advancements in fashion image generation, yet prior works usually treat garment generation and virtual dressing as separate problems, limiting their flexibility in real-world fashion workflows. Moreover, fashion image synthesis under multi-source...
We evolved for a linear world. If you walk for an hour, you cover a certain distance. Walk for two hours and you cover double that distance. This intuition served us well on the savannah. But it catastrophically fails when confronting AI and the core exponential trends at its heart. From the time I...
Large language models often fail on tasks they seem to already understand. In our experiments, this appears to be less about missing knowledge and more about certain internal circuits not being strongly activated during inference. We explore Selective Neuron Amplification, which increases the...
Neural networks deliver impressive predictive performance across a variety of tasks, but they are often opaque in their decision-making processes. Despite a growing interest in mechanistic interpretability, tools for systematically exploring the representations learned by neural networks in...
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology. Desalination plants in the Middle East are increasingly vulnerable  As the conflict in Iran has escalated, a crucial resource is under fire: the...
Psychological stress is clinically relevant in cardio-oncology, yet it is typically assessed only through patient-reported outcome measures (PROMs) and is rarely integrated into continuous cardiotoxicity surveillance. We estimate perceived stress in an elderly, multicenter breast cancer cohort...
Apr 8, 09:01 UTCMonitoring - A fix has been implemented and we are monitoring the results.Apr 8, 07:59 UTCIdentified - The issue has been identified and a fix is being implemented.Apr 8, 07:06 UTCUpdate - We are continuing to investigate this issue.Apr 8, 06:23 UTCInvestigating - We are currently...
Apr 8, 08:04 UTCIdentified - The issue has been identified and a fix is being implemented.Apr 8, 07:28 UTCInvestigating - Cloudflare is investigating issues with operations on the Cloudflare One Dashboard.Customers using the Cloudflare One dashboard are impacted. Customers may face issues when...
Apr 8, 06:45 UTCResolved - Cloudflare investigated increased HTTP 5XX errors in Sau Paulo between 0645: - 06:55 UTC. Issue has been resolved
Anthropic steps up the offensive vs OpenAI's upcoming IPO woes
Apr 7, 20:08 UTCMonitoring - Current status: We've confirmed that our service has recovered following the routing of affected traffic to healthy infrastructure, and we're continuing to monitor as the health of Snowsight continues to improve. Other services are fully recovered, but some users who...
We shed light on OpenAI's first Dark Factory for the first time.
Hackers linked to Russia's military intelligence units are using known flaws in older Internet routers to mass harvest authentication tokens from Microsoft Office users, security experts warned today. The spying campaign allowed state-backed Russian hackers to quietly siphon authentication tokens...
Apr 7, 16:35 UTCIdentified - The cause of this issue has been identified and a fix is being implemented.Apr 7, 15:55 UTCInvestigating - Cloudflare Analytics processing is running behind. This affects timely delivery of customer data. This also affects querying data, and some users may see errors...
Apr 7, 14:27 UTCResolved - This incident has been resolved.Apr 7, 14:20 UTCMonitoring - A fix has been implemented and we are monitoring the results.Apr 7, 13:39 UTCIdentified - The issue has been identified and a fix is being implemented.Apr 7, 13:34 UTCMonitoring - A fix has been implemented and...
New advisory warns cyber threat group APT28 have exploited vulnerable edge devices to support malicious operations.
Russian cyber actor APT28 exploit vulnerable routers to hijack DNS, enabling adversary‑in‑the‑middle attacks and theft of passwords and authentication tokens.
Status: ResolvedAll impacted services have now fully recovered.Affected components Image Generation (Operational) Deep Research (Operational) Connectors/Apps (Operational) Login (Operational) File uploads (Operational) Search (Operational) Conversations (Operational) Compliance API (Operational)...
a quiet day lets us give due respect to the enormously successful Gemma 4 launch
Status: ResolvedAll impacted services have now fully recovered.Affected components Responses (Operational)
Status: ResolvedAll impacted services have now fully recovered.Affected components Conversations (Operational)