ZeroHour

Search: “streaming”

264 stories

How Does mHC Use Its Residual Streams? Selective Routing and Near-Identity Mixing

Analysis of DeepSeek-V4-Flash shows four-stream mHC residual blocks use only about two streams effectively, with late-layer mixing providing little benefit.

The study examines the four-stream residual pathway of DeepSeek-V4-Flash, finding typical attention or FFN sites effectively use about two streams and that residual mixing is modest, occurring primarily in early layers. Replacing late mixers with identity increases C4 perplexity by only 1.9% while replacing early mixers raises it by 41%. Retaining the three largest routing weights per token increases perplexity by at most 2.7%, showing the model uses only part of the flexibility afforded by the four-stream design.

arXiv cs.AI / cs.LG / cs.CL · 12d agoAI research

NVIDIA Brings Real-Time AI to Broadcast, Sports and Global Streaming at IBC

NVIDIA expanded its AI for Media suite at IBC 2026, adding NIM microservices for synthetic video detection, body pose, frame generation, upscaling and HDR.

At IBC 2026 in Amsterdam, NVIDIA announced a major expansion of NVIDIA AI for Media, a collection of GPU-accelerated SDKs, NIM microservices and blueprints for broadcast and streaming workflows. The Synthetic Video Detector (SVD) NIM microservice reaches 99.3% accuracy on text-to-video and 97.7% on image-to-video content, while Video Frame Generation boosts frame rates 2x-4x and Video Super Resolution adds 10-bit support; TrueHDR converts SDR to HDR at up to roughly 2,000 nits. Partners including Dalet, TwelveLabs, Wowza, Vizrt and Ross Video are integrating the new services into verification, compliance and live-production workflows.

NVIDIA Blog · 7d agoAI industry

Omni-Streaming Thinking

Omni-Streaming Thinking fixes premature cross-modal commitment in streaming omni-modal models via pending claims verified against modality-specific evidence, beating baselines by over 10%.

The paper identifies 'premature cross-modal commitment', where streaming models keep relaying early visual interpretations even after audio contradicts them. OST generates evidence-linked pending claims with future verification intervals, stores audio and visual evidence separately, and refutes claims when contradictory evidence appears. Built on a frozen Qwen3-Omni-30B-A3B-Instruct backbone with lightweight adaptation, it outperforms open baselines by more than 10% relative on five streaming and audio-visual benchmarks. On the new OST-DiagBench it reaches d-prime 2.95 versus at most 1.38 for open baselines, while reducing vision-induced auditory hallucinations.

Hugging Face daily papers · 2d agoAI research1

ReactVAU: A Slow-Fast Decoupled Framework for Streaming Video Anomaly Understanding

ReactVAU is a slow-fast streaming framework for real-time video anomaly understanding that reserves heavyweight MLLM reasoning for suspicious events, improving efficiency.

ReactVAU addresses causal streaming video anomaly understanding with three components: a lightweight Fast Detection Module using Spatial Grid Folding, Anomaly-Aware Persistent Memory that protects critical visual cues from temporal decay, and a Slow Reasoning Module activated only on suspicious events. This design minimizes heavyweight MLLM invocations during long normal intervals. Experiments show competitive anomaly detection and causal reasoning under strict streaming constraints with significantly enhanced computational efficiency.

Hugging Face daily papers · 9d agoAI research

ShallowStream: Index Shallow then Answer Deep for Streaming Video Understanding

ShallowStream builds streaming-video retrieval indexes from shallow MLLM layers, cutting per-frame prefill latency by up to 52.1x.

ShallowStream is a framework for streaming video understanding with multimodal LLMs that uses the model's shallow layers to simultaneously encode frames and maintain an always-on lightweight retrieval index via shallow-layer KV caches, avoiding full-depth prefill for every incoming frame. At query time, shallow-layer attention scores plus a diversity-aware selection strategy retrieve relevant context frames. It reports performance on par with the strongest existing streaming methods while reducing per-frame prefill latency by up to 52.1x and 10-second end-to-end latency by up to 11.9x, with code released on GitHub.

Hugging Face daily papers · 14d agoAI research

microsoft/VibeVoice-ASR-Streaming-7B — new model trending #27 on Hugging Face

Microsoft released VibeVoice-ASR-Streaming-7B, an open streaming ASR model with speaker attribution, custom hotwords, and support for 10 languages under MIT license.

Microsoft Research released VibeVoice-ASR-Streaming-7B on Hugging Face, a unified streaming speech recognition model that continuously transcribes who said what as speech arrives. The 7B model supports customized hotwords for domain-specific terms and 10 languages including Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, Russian, and Spanish. Code is available at github.com/microsoft/VibeVoice with a live demo, and a technical report is on arXiv (2609.02812). The model is licensed under MIT.

Hugging Face trending models · 14d agoModel release1

Real-SWE: Benchmarking AI models on private, real-world, enterprise codebases

Real-SWE benchmark tests coding agents on licensed private enterprise codebases; top model Fable 5.1 resolves only 38.8% of tasks.

Real-SWE is a new benchmark evaluating frontier AI coding agents on tasks drawn from private production codebases licensed from real companies, spanning billing, tax calculation, and cross-service migrations. Fable 5.1 with Claude Code leads at 38.8% resolution rate (pass@1 over eight runs), followed by GPT-6 Astra Codex CLI at 33.8% and Gemini 3.8 Flash Gemini CLI at 31.2%. Tasks use native harnesses and realistic tooling including Docker, Kubernetes, PostgreSQL, Redis, and Linear; median reference solutions edit 11 files versus 6 for DeepSWE and FrontierCode.

Retrospectively Reverse-Engineering Apple's Neural Engine

A developer reverse-engineers Apple's M1 Neural Engine architecture, mapping compute cores, MAC datapaths, and schedulers to explain the NPU's decline as transformers displaced CNN workloads.

A developer who previously maintained a reverse-engineered Linux driver for Apple's Neural Engine (ANE) published a retrospective deep dive mapping the M1 ANE's full internal architecture: compute, datapath, scheduler, memory, and execution model. The M1 ANE has 16 compute cores with 128 FP16 (or 256 INT8) MAC lanes each, totaling 2048 parallel MAC lanes, using 32-bit Q16.16 fixed-point accumulation with FP16 readout and an accumulator that saturates at 2^15. The author argues the ANE's dataflow was architected around the predictable reuse patterns of 2017-era CNN workloads (dating to the A11 Bionic), which autoregressive transformer decode broke, limiting its usefulness for general ML. With Apple's M5 folding ANE cores into GPU cores to tout LLM performance, the post frames this as the beginning of the end for the standalone NPU.

Event-Native Symbolic-Temporal Spike Encoding Framework for Heterogeneous Cyber Streams

An event-native spike encoding framework lets recurrent spiking neural networks run intrusion detection directly on heterogeneous packet and CAN bus streams.

The paper maps heterogeneous cyber events—categorical identifiers, local frequency context, and inter-event timing—directly into sparse spike-compatible inputs for spiking neural networks. This avoids flow aggregation and fixed windows that add buffering latency and obscure temporal structure in traditional IDS pipelines. Compact recurrent SNNs under edge-oriented neuromorphic hardware constraints achieve a hybrid anomaly metric of 0.987 on packet-level Network IDS and 0.980 on message-level CAN IDS.

arXiv cs.CR · 2d agoResearch

AlayaVista: Streaming World Modeling from Panoramic States to Perspective Video

AlayaVista is a camera-controllable streaming video world model that decouples panoramic scene evolution from perspective synthesis, trained on a 1,318-hour 4K dataset.

AlayaVista builds a 360-degree scene prior from a single perspective image, evolves it as a camera-conditioned panoramic latent state, and maps it to perspective video via a latent viewport renderer plus a perspective refiner. Chunk-autoregressive generation and few-step distillation enable efficient streaming. The authors introduce MUGEN, a real-world panoramic video dataset with 1,318 hours of at-least-4K video and rich semantic and geometric annotations.

Hugging Face daily papers · 3d agoAI research

Top 10 Best Browser Isolation Solutions in 2026

A 2026 market overview ranks ten remote browser isolation tools, with Menlo Security as the pure-play reference as SSE vendors bundle isolation.

The article compares ten remote browser isolation (RBI) options, including Menlo Security, Zscaler, Cloudflare, Palo Alto Networks, Broadcom (Symantec), Forcepoint, Skyhigh Security, Ericom (Cradlepoint), Authentic8, and Garrison. It argues that RBI has become a bundled policy action inside SSE platforms from Zscaler, Cloudflare, Palo Alto, Broadcom, Forcepoint, and Skyhigh, compressing standalone pricing and driving consolidation such as Ericom's isolation moving under Cradlepoint (Ericsson). Enterprise browsers like Island and Chrome Enterprise Premium are reshaping the RBI-versus-browser decision for managed users, while selective policy-driven isolation of risky categories is described as the prevailing 2026 architecture. The piece is a buyer's guide with vendor positioning, not an incident or vulnerability report.

Cyber Security News · 5d agoIndustry

BreezeBlue/Breeze-TTS-2 — new model trending #19 on Hugging Face

BreezeBlue open-weights Breeze TTS 2, a bilingual text-to-speech model it ranks #1 among open-weight models on the Artificial Analysis TTS leaderboard.

BreezeBlue released open weights and Apache 2.0-licensed PyTorch inference code for Breeze TTS 2 on 2026-08-25. The text-to-speech model supports English and Chinese, voice cloning, reference-free voice design, voice direction, and inline vocal events like (laugh) and (sigh). Reported performance includes #1 open-weight ranking on the Artificial Analysis Elo leaderboard, under 40 ms time-to-first-audio, a 0.32 real-time factor on an NVIDIA H100, and about 7.7 GiB GPU memory for eager inference.

Hugging Face trending models · 22d agoModel release

12 Best Enterprise Browsers Compared (2026): Features & Pricing

2026 comparison of twelve enterprise browsers ranks Island and Palo Alto Talon as purpose-built leaders, with Chrome Enterprise and Edge free or bundled.

Guide compares twelve enterprise browser options across three models: purpose-built secure browsers (Island, Talon, Surf), layered controls on existing browsers (Chrome Enterprise, Edge for Business, LayerX, Seraphic), and streamed/isolated browsers (Kasm). Island and Palo Alto's Prisma Access Browser lead the purpose-built category for BYOD and contractor DLP. It also notes Mammoth Cyber has ceased operations.

GBHackers · 1d agoTools

12 Best Browser Isolation Solutions Compared (2026): Features & Pricing

2026 comparison ranks Zscaler, Cloudflare, Menlo Security, Garrison (Everfox), Authentic8 and Kasm among twelve remote browser isolation solutions.

Guide compares twelve RBI products across four architectures: pixel streaming, DOM/vector reconstruction, platform-embedded SSE isolation, and self-hosted containers. Zscaler and Cloudflare lead RBI delivered inside SSE platforms, while Menlo Security leads isolate-everything efficacy and Garrison (Everfox) provides hardware-grade isolation for government use. Most offerings price per user per month.

GBHackers · 1d agoTools

Kimi K3 (2.8T) at 1 token/s on a MacBook Pro, streamed from four SSDs

A GitHub project streams Moonshot's 2.8T-parameter Kimi K3 from four SSDs to run at 1 token/s on a MacBook Pro.

A GitHub repository (argonautlabsai/deltafin) demonstrates running Moonshot AI's Kimi K3, a 2.8-trillion-parameter model, on a MacBook Pro by streaming its weights from four SSDs at 1 token per second. The project drew 55 points and 22 comments on Hacker News, highlighting consumer-hardware inference for extremely large models.

Risky Bulletin: White House lets private companies carry out offensive cyber ops

A White House memo directs DHS to create a program letting vetted private companies conduct US-government-directed offensive cyber operations against cybercrime.

A presidential memo tasks the DHS National Coordination Center with building a program, under DOJ and DHS oversight, through which private-sector companies can conduct offensive cyber operations against large-scale cybercrime organizations. Requirements include secure facilities, vetted personnel, a $1 million escrow for damages, and written approvals co-signed by DHS and DOJ executive directors. The program must launch within 60 days, around October 11, expanding a March executive order targeting scam compounds, ransomware, and other large-scale cybercrime.

Risky Business News · Aug 14, 2026Policy & legal

The Fraud Ecosystem: A Transition From Known Marketplaces to a Fragmented Environment

Rapid7 analyzes how fraud marketplaces are fragmenting into specialized shops after larger marketplaces were dismantled, aided by new MITRE F3 framework

Rapid7 reports a shift from large known fraud marketplaces to a fragmented environment of smaller specialized storefronts such as Xleet, Blackpass, Infodig, and Styx, operating across dark web channels, Telegram, and P2P options. These Fraud-as-a-Service shops sell stolen accounts, PII, synthetic identity generation, infrastructure, and money laundering support, supporting schemes like business email compromise. MITRE's Fraud Fighting Framework (F3), introduced in early 2026, aims to help security teams prioritize monitoring of fraud TTPs, particularly account takeover techniques. Fraud damages are anticipated to approach hundreds of billions of USD.

Rapid7 Blog · 5d agoPhishing & fraud

Data access: the hidden cost of security vendor lock-in

Elastic compares SIEM data egress cost, latency, and fidelity across CrowdStrike, Microsoft, Google, and Splunk, arguing vendors engineer lock-in.

Elastic Security Labs published an opinion piece comparing how major SIEM and security vendors handle data egress, based on each vendor's public documentation as of September 2026. It rates CrowdStrike Falcon Data Replicator and Palo Alto Networks XSIAM Event Forwarding as restricted (paid add-ons with batch delays), Microsoft as partially open, Splunk as open, and Elastic as open with no export license. The piece argues frictionless ingestion paired with licensed or delayed egress is an intentional lock-in business model, and cites CrowdStrike's 2026 Global Threat Report eCrime breakout time of 29 minutes to argue real-time telemetry access is now essential.

Elastic Security Labs · 12d agoIndustry

[AINews] Fal’s H3 Max Live breaks the infinite videogen barrier

Fal post-trained MiniMax H3 into a 'Max' variant with 35x-faster inference, enabling faster-than-realtime AI video generation and infinite streams.

Fal post-trained MiniMax's H3 model into a 'Max' variant and optimized it for its in-house inference engine, achieving roughly 35x the speed of the official endpoint. The optimization enables faster-than-realtime video generation, demonstrated by an infinite interactive AI-generated stream productized by levels.io. The roundup also notes Meta Muse Code's general availability with an SDK, open DeepSeek-V4-Flash-Vision-Exp weights, GLM-5.3-Flash's strong agentic cost/performance rankings, and Tencent's 770B-parameter Hy4 Preview MoE with 49B active parameters.

Latent Space · 15d agoAI industry

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.

ANY.RUN · 10h agoTools

PhysStream: Streaming Physics-Grounded Video Generation with Structured Scene Memory and Fine-Grained Motion Control

PhysStream enables mid-generation interactive control of physics-grounded video via structured scene memory and velocity-increment signals, reducing motion distribution distance 33%.

PhysStream is an autoregressive physics-grounded image-to-video model that maintains structured scene memory—positional maps and object tracking maps derived online from previously generated frames—and accepts fine-grained motion control via sparse velocity-increment signals encoding physical quantities. Training runs in two stages: a bidirectional model finetuned with motion-control conditioning, then a causal autoregressive model with structured scene memory. It supports interactive mid-generation control over multi-object tabletop rigid-body scenes, reducing motion distribution distance (FVMD) by 33% and trajectory error by 12% over the strongest baselines. Human evaluators preferred it in over 85% of in-the-wild comparisons.

Hugging Face daily papersupdated · 23h agofirst · 1d agoAI research 2 sources

Cyble Introduces Major Upgrade to its Executive Monitoring Module

Cyble upgraded Executive Monitoring in Cyble Vision, unifying impersonation, exposure, and mention findings with AI scoring and multi-channel alerting.

Cyble released a major upgrade to Executive Monitoring within its Cyble Vision platform, consolidating Mentions, Impersonations, Exposures, and a new Surface Mentions source (blog and news coverage) into a single findings stream. The release adds AI-generated verdicts and recommendations, per-executive Risk and Reputation scoring for board reporting, and richer onboarding with aliases, reference images, and addresses to cut same-name false positives. Alerting now supports data residency, bulk import, delivery via email, WhatsApp, and SMS, an Access API, and branded executive PDF reports; the upgrade is live with no customer action required.

Cyble · 6d agoTools

Edge0/Edge0-35B-A3B-preview — new model trending #30 on Hugging Face

Edge0 released a 35B sparse MoE model running in under 3 GiB of memory at 15 tok/s via SSD expert offload and int4 quantization.

Edge0-35b-a3b-preview is a 35B-parameter MoE (256 experts, 4 active per token) built on Qwen3.5-MoE 35B-A3B, shipped as a 4-bit checkpoint with LoRA and prerouter adapters under Apache 2.0. The edge0 framework streams expert weights from SSD on demand, bounding peak active memory at 2.9 GiB and achieving 14.9-17.7 tok/s decode on a Mac mini M4 Pro (MLX backend). Recover-LoRA distillation keeps the int4 model within 3.9 points of its fp16 base (79.2 vs 83.2 average on OpenCompass benchmarks including AIME 2026, HumanEval, GPQA-Diamond, MMLU-Pro, and IFBench).

Hugging Face trending models · 8d agoModel release

Omni Interaction Agent Technical Report

Researchers release Gander, an end-to-end omni interaction model with full-duplex streaming across video, speech, and text plus agentic capabilities.

Gander is an end-to-end model unifying omni perception, realtime interaction, and agentic capabilities in a single framework, accepting continuously streaming video, speech, and text. It uses a Cerebellum-Brain architecture where the Cerebellum handles realtime conversation and the Brain handles reasoning and agentic tasks, built on a streaming Thinker-Talker design with chunk-level token streams. Internal human evaluations report spoken dialogue on par with SOTA open source models and competitive omni interaction; the models, code, and data are released publicly.

Hugging Face daily papers · 8d agoModel release

Twitch is Mining Peoples' Streams to Train Amazon's AI

Twitch will train Amazon's AI models on streamers' content by default, requiring users to manually opt out in security settings.

Twitch announced a setting that lets streamers opt out of having their channel content train Amazon's generative AI models, confirming about a week of rumors. The toggle, labeled 'Allow your channel content to train generative AI content models at Amazon,' is opt-out rather than opt-in and sits at the bottom of the security settings page. Critics note many streamers may never learn of the change or find the buried setting, sweeping their content into Amazon's training data by default.

404 Media · Aug 13, 2026AI industry