60
60
60
30
42
In 2022, more than 40% of zero-day exploits used in the wild were variations of previous issues
60
55
57
60
55
60
30
55
Parallels Desktop Flaw Lets Non-Admin Mac Users Gain Root, but Intel Macs Can't Install Fixnew
JFrog disclosed CVE-2026-90894, a 7.8-rated local privilege escalation in Parallels Desktop for Mac, patched only in version 27, which Intel Macs cannot install.
JFrog researcher Yuval Moravchick disclosed CVE-2026-90894 (dubbed ParaShells, CVSS 7.8), a local privilege escalation in Parallels Desktop for Mac that lets non-admin users run code as root. The root-level prl_disp_service listens on a world-writable socket, and argument injection into a tar command via QProcess::splitCommand and the --use-compress-program option yields code execution as root, demonstrated on Parallels Desktop 26.4.0 build 57513 on Apple silicon. The fix appears in version 27.0.0, but Parallels Desktop 27 requires Apple silicon, leaving Intel Macs on the 26.x line with no build JFrog describes as fixed. No exploitation in the wild has been reported and Parallels has not published a statement.
52
Prior Labs Releases TabPFN-3.5: A Tabular Foundation Model That Beats the Winning Otto Kaggle Solution With Default Settings
Prior Labs releases TabPFN-3.5, a 220M-parameter open-weights tabular foundation model that beats the 2015 Otto Kaggle winning score with default settings.
Prior Labs released TabPFN-3.5, a tabular foundation model that predicts in a single forward pass without per-dataset training or tuning. The base model grew from 53M to 220M parameters with a single multitask checkpoint, learned Fourier features, and in-context ECDF rank encodings. It scores 0.375 on the 2015 Otto Kaggle private leaderboard versus the winning 0.382 and claims first place on seven tabular benchmarks including TabArena and BeyondArena. Open weights cover the base, Fast (84M), and Thinking variants, but production use requires the Prior Labs API or a commercial license.
55
Right Tool, Right Job: Native-Language Evaluation, Tokenizer Sensitivity, and Methodological Findings from a French-Only BabyLM
French BabyLM entry METRON-FR (125M GPT-2, 92.47M words) shows tokenizer artifacts dominate child-scale zero-shot evaluation; proposes standard diagnostics.
METRON-FR is a 125M-parameter GPT-2 pretrained on 92.47M French words, submitted to the BabyLM 2026 Strict track, scoring 85.97% on the native Quebec-French QFrBLiMP benchmark and 62.80% on the BabyLM-weighted leaderboard. A cross-lingual GLUE protocol combining French task-data translation with rank-16 LoRA shows relational tasks gain while world-knowledge tasks regress. Bilingual Lexicon Induction reaches p@1 of 68.84%, 18x above chance, and ablations show single-token zero-shot scoring is dominated by tokenizer and template artifacts at child scale.
12
Microsoft Confirms Remote Desktop Services Might Stop Working Following Sept. 2026 Security Update
Microsoft's September 2026 Patch Tuesday updates (KB5124008) can break Remote Desktop Services, causing RDP failures and freezes across Windows clients and servers.
Microsoft confirmed its September 2026 security updates, including KB5124008 for Windows 11 24H2/25H2, introduced a reliability regression where RDP connections fail after several minutes and servers hang at the 'Please wait for the Remote Desktop Configuration' screen. MMC, RDS Licensing Diagnoser, File Explorer, and the Windows Update settings page may also stop responding. The issue affects Windows 10/11 clients and Windows Server 2012 through 2025; Microsoft marked it Mitigated on September 11 and is developing a permanent fix, with VM restart offered as a temporary workaround.
50
Debian 13.7 ships the fixes behind 92 security advisories, updates 106 packages
Debian 13.7 'trixie' point release bundles 92 security advisories and 106 package updates, including kernel, glibc, u-boot and qemu fixes.
Debian shipped version 13.7 of 'trixie', folding in 92 previously published security advisories and corrections to 106 source packages, including six Linux kernel advisories (DSA-6381, DSA-6393, DSA-6405, DSA-6415, DSA-6466, DSA-6477). glibc fixes a buffer overflow (CVE-2026-5928) and buffer underflow (CVE-2026-5450), with 17 packages rebuilt against the updated library; qemu carries 25 CVEs including a secure boot bypass (CVE-2026-16288), imagemagick 24, wolfssl 15 and perl 13. Boot-chain fixes include a u-boot FIT image verification bypass (CVE-2026-46728), a BOOTP/DHCP buffer overread (CVE-2024-42040), and corrected intermediate certificate verification in sbsigntool. The installer was rebuilt with kernel ABI 6.12.107+deb13, and existing systems receive the fixes through normal package mirror updates.
30
[AINews] DeepSeek v4.1-Flash: 763B-P8B-D16B novel causal Encoder–Decoder architecture with vision marks the Return of the Whale
DeepSeek released V4.1-Flash, an open-weight 763B-parameter model with a novel causal encoder-decoder architecture, 1M context, vision input, and MIT license.
DeepSeek launched V4.1-Flash, an open-weight MIT-licensed model using a novel causal encoder-decoder architecture with 763B total parameters and asymmetric active parameters: 8B for prefill and 16B for decode. It supports 1M-token context and text+image input, priced at $0.30 per 1M input and $1.20 per 1M output tokens with a 50% off-peak discount. Artificial Analysis scored it 40 on its Intelligence Index, above DeepSeek V4 Pro 0813, and Vals ranked it the #1 open-weight model ahead of Kimi K3. Baseten shipped day-0 support and Ollama began rolling it out to paid subscribers.
80
Anthropic Adds Plugin Evals to Claude Code: 6 Grader Types, a No-Plugin Baseline, and a CI Gate for Skills
Anthropic ships a plugin evals workflow for Claude Code with six grader types, a no-plugin baseline arm, and a CI gate via threshold and cost flags.
Anthropic published a plugin evals workflow for Claude Code, exposed via the "claude plugin eval" command on v2.1.269+. Six grader types exist: regex, tool_used, tool_order, and file_exists are free transcript checks, while llm and baseline invoke a billed judge model. Every case runs with and without the plugin, and the delta (Δ) isolates the plugin's contribution; a Δ near zero with a failing tool_used:Skill grader indicates the skill never triggers. CI gating uses --threshold 0.8, --max-cost-usd, --trust-plugin, and --no-publish flags, with results written to a report.html under evals/results/.
42
Metasploit Wrap Up: This One Goes to Sixteen!
Metasploit adds 16 modules including 10 exploits, five covering CISA KEV vulnerabilities in Cisco, SonicWall, PaperCut, JetBrains and Langflow.
Rapid7's weekly Metasploit update ships 16 new modules, 10 of them exploit modules, with five targeting CISA KEV entries. New exploits cover Cisco Secure Firewall Management Center auth bypass (CVE-2026-20079), a SonicWall SMA1000 SSRF-to-root RCE chain (CVE-2026-83548/CVE-2026-83549), JetBrains TeamCity deserialization RCE (CVE-2026-63077), PaperCut NG/MF chain (CVE-2026-81578/CVE-2026-82078), and Langflow authenticated RCE (CVE-2026-19295). The SonicWall and PaperCut chains were reported as actively exploited zero-days, and a new ESC8 relay module exploits CVE-2026-20929 against AD CS Web Enrollment.
48
PaperCut Replaces Emergency Patches With Fixes for Two Actively Exploited Flaws
PaperCut shipped maintenance releases replacing emergency patches for two actively exploited flaws abused in AI-assisted attacks on 395 organizations.
PaperCut released PaperCut NG/MF versions 26.0.5, 25.0.13 and 24.1.10, superseding Emergency Patch Releases 1-3 for CVE-2026-81578 and CVE-2026-82078, which enable authentication bypass and arbitrary code execution on susceptible instances. GreyNoise and Blackpoint Cyber reported a suspected Russian-speaking actor weaponizing both flaws against at least 395 organizations in 48 countries, concentrated in the U.S. education sector. The campaign used hundreds of AI agents powered by OpenAI's Codex harness and a DeepSeek model, originating from IP 45.142.193.132, and avoided organizations in Russia, China, Hong Kong, Thailand and Iran.
78
WordPress Blocks High-Risk Plugin Releases With New AI-Powered Automated Security Review
WordPress.org now runs AI-powered automated security reviews on every plugin release, automatically blocking high-risk updates before distribution to millions of sites.
WordPress launched an automated security review that combines multiple AI models and Jetpack Scan during a six-hour cooldown to score each plugin release; updates above the blocking threshold are automatically held back from the WordPress.org update API. The change follows a July 28 incident where a backdoor added to a plugin with roughly 20,000 active installations was detected during cooldown and never delivered; the Plugins Team removed it 26 minutes after a Wordfence notification. Blocked developers receive an email with findings and are advised to publish a corrected version rather than await manual appeal.
58
MInTRL: Off-policy Intervention can boost On-policy RL
MInTRL injects sparse judge corrections into on-policy RL rollouts, expanding exploration beyond on-policy sampling while preserving learnability on math and code benchmarks.
Minimal Intervention Reinforcement Learning periodically has a judge-intervention policy replace erroneous suffixes of the current policy's output with short corrections, then returns control, keeping trajectories largely on-policy. Training uses a sequence-level advantage-regression objective that removes the need for importance sampling. Across math and code benchmarks it consistently beats standard on-policy and off-policy baselines, remains effective with self-intervention, and performs best at moderate intervention intensity.
40
Cognition launches new SWE-2 model, Rivaling Fable 5.1 and GPT-Astra
Cognition released SWE-2, a coding model post-trained from Kimi K3 that scores 50.0% on FrontierCode 1.1 Main, near Fable 5.1 at 64% lower cost.
Cognition introduced SWE-2, its most advanced coding model, post-trained from the 2.8T-parameter Kimi K3 base model. It achieves 50.0% on FrontierCode 1.1 Main, 73.0% on DeepSWE 1.1, and 92.8% on Terminal-Bench 2.1, beating Grok 4.6 and SWE-1.7 while matching Fable 5.1 and GPT-5.6 Sol at a fraction of the price. The company says it scaled reinforcement learning to the multi-trillion-parameter regime for the first time, using Pareto-informed cost penalties that train all reasoning-effort levels in a single run, tripled RL environments, and NVFP4/FP8 quantization-aware training. SWE-2 is available today in Devin Desktop and CLI, with rollout on Devin Web and Fusion.
72
Training a 3.8B LLM to 0.384 CORE for $998 – Hugo Vergnes
Independent developer Hugo Vergnes trained a 3.8B-parameter Llama-style model to 0.384 CORE on 65B tokens for $998 in 43 hours on rented B200s.
Hugo Vergnes trained little-lm, a 3.848B-parameter decoder-only LLM, on 65.3B tokens in 43 hours for $998 using rented NVIDIA B200s, scoring 0.384 on the CORE benchmark and beating nanochat d32 (0.310) at similar cost. The Llama-style architecture uses RMSNorm, RoPE, GQA with 24 query and 8 KV heads, relu-squared MLPs, QK-norm, and ResFormer-style value embeddings that account for 19% of parameters. Key wins included the Muon optimizer for matrix parameters, a trapezoidal learning-rate schedule with linear cooldown, FP8 training plus vocabulary padding for roughly 33% throughput gains, and the ClimMix dataset over FineWeb-Edu. The project, inspired by Karpathy's nanochat, was built as a config-driven YAML framework for small LLM training.
35
CVE-2026-34486: Apache Tomcat Auth Bypass Vulnerability
SentinelOne details CVE-2026-34486, an Apache Tomcat EncryptInterceptor bypass exposing unencrypted cluster communications, with patches and workarounds available.
CVE-2026-34486 stems from an incomplete fix for CVE-2026-29146, letting network-positioned attackers bypass the EncryptInterceptor and read plaintext Tomcat cluster traffic, including session and application data between nodes. Affected versions 9.0.116, 10.1.53 and 11.0.20 are fixed in 9.0.117, 10.1.54 and 11.0.21. Workarounds include network-level encryption such as IPsec or VPN tunnels and isolating cluster communications to secured network segments.
45
Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails
Research shows on-policy expert correction, not imitation fine-tuning, lets weaker agent models catch up under evolved harnesses.
Researchers study how to combine automated agent-harness evolution with lightweight fine-tuning across seven enterprise agent tasks. Naively training weaker models (Qwen3-Coder, Gemma 4) on expert trajectories under an evolved harness regressed performance by 4 to 30 points on all tasks. They propose an on-policy correction pipeline, automated by a meta-level MLE agent, where an expert rewrites only the failing turn of the weaker model's rollout, preserving model-harness fit.
35
When Does Scale-Invariant Optimization Become Unstable? An Exact Schedule Law with Weight Decay
Researchers derive an exact law linking learning-rate schedules and weight decay in normalized networks, pinpointing when scale-invariant optimization destabilizes.
The paper shows that normalization makes large parts of neural networks scale-invariant, creating a hidden feedback loop where learning-rate schedules and weight decay interact through the parameter norm to control the effective optimizer step. An exact discrete-time law with a single scalar quantity separates contraction- and expansion-dominated effective learning-rate regimes, and the balance point is intrinsically unstable, so constant learning rate with weight decay produces recurrent behavior instead of a stable equilibrium. A unified homogeneous-optimizer framework explains why adaptive methods stabilize more weakly under normalization. The law is validated with high precision on MLPs, CNNs, and GPT-2 across MNIST, CIFAR, WikiText, and OpenWebText, with code released on GitHub.
30
Switching Password Managers in 2026
An Apple password-management engineer (writing personally) explains switching password managers via iOS direct app-to-app export, crediting FIDO Credential Exchange for passkey portability.
Ricky Mondello, an Apple engineer in password management and authentication (writing personally, not for Apple), demonstrates moving 100 items from 1Password to Apple Passwords via the iOS system export, which transfers passwords, passkeys, verification codes, and notes directly between apps with Face ID confirmation and no intermediate files. He credits the FIDO Alliance's Credential Exchange format (first draft May 2024, now an open spec) and iOS 26/macOS 26 for enabling phishing-resistant credential transfer between apps like 1Password, Bitwarden, Dashlane, DuckDuckGo, and Devolutions. The post is advice rather than news, recommending bulk transfer on the existing device, switching AutoFill to the new app, and treating the new manager as the source of truth.
14
SAP September 2026 Security Update Fixes 4 Critical Vulnerabilities and 15 Other Flaws
SAP's September 2026 Patch Tuesday ships 19 security notes, including four criticals led by a CVSS 10.0 memory-corruption flaw in Extended Passport Processing.
SAP released 19 new Security Notes plus one update, fixing four critical vulnerabilities and 15 additional flaws. The most urgent is CVE-2026-44756 (CVSS 10.0), memory corruption in Extended Passport (EPP) Processing affecting many SAP Kernel and Web Dispatcher releases. Other criticals include CVE-2026-58240 (NetWeaver Message Server missing authentication, 9.8), CVE-2026-76969 (credential disclosure in CAP sap/cds-mtxs, 9.4), and CVE-2026-66768 (improper access control in SAP GUI for Java, 9.0). High-severity fixes cover XXE in Integration Suite Trading Partner Management, deserialization in NetWeaver Business Client, and memory corruption in NetWeaver AS for ABAP.
65
Marigold V2: Revisiting Diffusion Transformers for Monocular Depth Estimation
Marigold V2 adapts diffusion transformers for monocular depth estimation, improving AbsRel 16-26% over the previous best on KITTI and ETH3D.
Huawei's Bayer lab revisits the Marigold approach to repurpose image generation and editing models built on the diffusion transformer (DiT) architecture into monocular depth estimators. The recipes target single-step inference from pretrained multi-step flow-matching models, with remedies including alignment to ground-truth semantic features and a two-stage fine-tuning protocol using a Sinkhorn-based loss. The resulting model produces crisper depth maps that generalize out-of-distribution and also achieves state-of-the-art results on surface normals estimation and intrinsic image decomposition.
42
Co-Evolving Harnesses and Models: On-Policy Correction Helps Weaker Models Catch Up Where Imitation Fails
Research shows imitation of expert trajectories breaks weaker models' harness fit, while on-policy expert correction preserves gains across seven enterprise agent tasks.
The paper studies combining automated agent-harness evolution with lightweight fine-tuning across seven enterprise agent tasks using Qwen3-Coder and Gemma 4. Training weaker models on complete expert trajectories under an evolved harness regressed performance by 4-30 points on all tasks, disrupting model-harness fit. The authors propose an on-policy expert-correction pipeline, automated by a meta-level MLE agent, that rewrites only failing turns and preserves the model's planning style.
48
Latent-MoE: Domain-Aware Mixture-of-Experts for PDEs with Multi-Regime Physics
Latent-MoE adds domain-aware mixture-of-experts routing to PINNs, improving accuracy over an order of magnitude on multi-regime physics PDEs.
The paper shows standard coordinate networks used in physics-informed neural networks have translation-variant NTKs causing long-range gradient conflicts on PDEs with spatially varying physics. MoE architectures with centered compact-support routers produce a uniformly banded NTK that localizes learning. Latent-MoE interleaves domain-aware MoE blocks in a shared backbone, outperforming FB-PINNs and X-PINNs by over an order of magnitude on multi-stage time-variable physics benchmarks.
22
Scalability Analysis of Distributed Kolmogorov-Arnold Network Training on High-Performance Computing Systems
An empirical study shows distributed Kolmogorov-Arnold Network training reaches 74.7% parallel efficiency at 8 A100 GPUs, with overheads driven by All-Reduce choices.
The study evaluates data-parallel Kolmogorov-Arnold Network (KAN) training on the FinisTerrae III supercomputer using up to 8 NVIDIA A100 GPUs across 4 nodes with PyTorch Distributed Data Parallel. Strong scaling yields 5.97x speedup and 74.7% parallel efficiency at 8 GPUs, comparable to conventional deep learning workloads, while communication overhead ranges from 1.3% to 6.1%, driven mainly by All-Reduce algorithm selection and inter-node latency rather than KAN's edge-wise gradient structure. Weak scaling shows an initial single-to-multi-GPU throughput drop followed by stability, and the parameter-to-memory ratio improves with model size even as training time scales unfavorably. The authors provide GPU topology and model-size deployment guidelines for KAN training.
26
Attackers Hijack MikroTik Routers Through Internet-Exposed SSH Without Authentication
CERT Polska warns attackers exploit internet-exposed MikroTik SSH for unauthenticated full admin control; attacks observed since at least September 2.
CERT Polska warned on September 5 that attackers exploit MikroTik routers with internet-reachable SSH to gain full administrative control without authentication, calling the two-flaw combination 'MikroTrick'. Successful attacks date to at least September 2; no victim count or attacker identity was disclosed. MikroTik shipped fixed RouterOS releases: 6.49.21, 7.23.4 (with 7.23.5 fixing an IPv6 DHCP regression), 7.24.2, and 7.25beta3. CERT recommends immediate updates, checking for unauthorized configuration changes, and restricting SSH, WWW/WWW-SSL, and bandwidth-test services on unpatched devices.
75
Jackrong/Qwopus3.8-27B-Flash-GGUF — new model trending #26 on Hugging Face
Community fine-tune Qwopus3.8-27B-Flash, built on Qwen3.8-27B, cuts agent reasoning latency with 12.8% faster decoding and 80.7% MTP acceptance.
Jackrong released Qwopus3.8-27B-Flash, a fine-tune of Qwen3.8-27B optimized for long-running agent workloads, reporting 12.8% faster decoding and 80.7% multi-token-prediction acceptance. Training used roughly 1.5 million teacher-scored SFT examples filtered to the top 10%, followed by reinforcement training with NVIDIA NeMo-RL and GSPO. The author notes an explicit trade-off: MMLU-Pro mixed-set scores are lower than the base model, and a known bug can produce incorrect Python indentation. Author-provided benchmarks have not been independently verified.
35
What LLM Trading Agents Actually Do in Production: A Six-Month, Population-Scale Record from Two Fleets
Six-month record of 7.5M LLM trading agent invocations shows volatility-blind sizing, minimal upside capture, and no directional edge across two fleets.
The study records autonomous LLM trading agents in production across DX Terminal Pro (3,505 user-funded vaults trading real ETH in Base memecoin markets) and the DXAP fleet (500-599 agents on Hyperliquid perpetuals), spanning roughly six months, 7.5M single-model invocations and about 300K onchain actions. A risk slider explains leverage (+0.425 per level), median leverage is 5.0x in every volatility sextile, and one posture-slider cell holds 62% of liquidations. Agents capture little upside: 43.2% of positions saw +300 bps favorable excursion within 24h yet 49.3% of those closed negative, while the DXAP fleet trails a matched retail benchmark (41% vs 50% roundtrip win rate). A paired-replay league of frontier models finds decision quality statistically indistinguishable at this horizon.
30
How Developers Prevent Production Risk at the Source
Wiz argues shift-left code-stage security cuts remediation costs and ships Wiz Code, WizOS, and WizCLI to catch flaws pre-deployment.
A Wiz blog post argues that fixing vulnerabilities at the code stage costs a fraction of production remediation, citing the node:20-slim base image which ships 14 known CVEs, three of them critical. It claims AI agents can now analyze open-source commits and generate working exploits in minutes, shrinking the window between a public fix and automated attacks. The post promotes WizOS near-zero-CVE base images, WizCLI and MCP-server checks inside developer and AI agent sessions, and Wiz Code scanning across pull requests, repositories, and pipelines covering SAST, SCA, secrets, and IaC misconfigurations.
15
Attackers plant remote access tools on compromised PaperCut servers
Attackers chained two PaperCut NG/MF zero-days for unauthenticated access, installing SimpleHelp and AnyDesk remote access tools on compromised servers.
An ongoing campaign exploits chained zero-days CVE-2026-81578 (improper access control) and CVE-2026-82078 (unsafe dynamic class loading) in internet-facing PaperCut NG and MF Application Servers, enabling authentication bypass and arbitrary Java bytecode execution. Post-compromise activity includes user and domain enumeration, payload download from sendit.sh, and silent installation of SimpleHelp and AnyDesk for redundant remote access; Defused observed CVE-2026-81578/CVE-2026-82078 exploit activity in honeypots since August 29, including data theft via Derby database dumps. Emergency patches were released August 28 and August 30, but 47% of roughly 2,500 PaperCut installs tracked by Huntress run v23 or older, for which no patch is available.
78