Live intelligence

Latest AI updates

1,735 updates

Tutorial

OmniParseBench: an open benchmark for document parsers, rebuilt and audited

This tutorial walks through Datalab's OmniParseBench, an open document-parsing benchmark comprising 16,288 pass/fail tests across 2,937 pages covering scanned forms, handwriting, arXiv math and dense tables. Using the released verdicts for 27 parsers, the author recomputes the leaderboard, inspects how the scorer's unit tests reach a verdict, and tests ranking certainty and per-document-type slices. It reports that the top two parsers are statistically tied, that the leader changes across most s

MarkTechPostOct 23, 2026Score 70

OmniParseBenchDatalabdocument AI benchmarking

Products

Microsoft Research unveils Flint-interactive, a language for AI-generated interactive charts

Microsoft Research introduced Flint-interactive, a language built on Flint that lets AI agents and developers specify interactive data visualizations using high-level semantic events and update operators. Its compiler handles interaction logic and visual dependencies, translating clicks, drags and selections into chart updates, which the team says improves reliability and reduces generation cost versus low-level visualization code. It supports chat-based chart generation via Flint-MCP and MCP Ap

Microsoft ResearchOct 23, 2026Score 73

Microsoft ResearchFlintdata visualization

Industry

AI-Discovered PHP Vulnerabilities Broadly Exploited, Amazon Warns

Amazon says vulnerabilities in a widely used programming language, discovered with AI tools, are now being broadly exploited in the wild. The company identified a hacking attempt four days before the flaws were publicly disclosed. The flaws have prompted governments and companies to rush to secure systems. The language reportedly underpins much of the internet.

Ars Technica AIOct 22, 2026Score 75

AI securityvulnerability researchAmazon

Industry

AI model helped identify critical PHP vulnerability affecting much of the web

Two critical vulnerabilities in PHP, a language used by a large share of websites, led to temporary shutdowns of sites including DeviantArt, 4Chan and the Internet Archive. The flaw reportedly exists in PHP versions spanning roughly 20 years, including unsupported releases. Security firm Wiz described it as among the most severe vulnerabilities discovered in software history. One item in the coverage states that GPT-5.6 Sol was used to identify the critical PHP flaw.

The Verge AIOct 22, 2026Score 70

PHPcybersecurityAI for vulnerability discovery

Products

NVIDIA Minerva Brings AI-Driven Observability to AI Factories

NVIDIA launched Minerva, an AIOps platform that combines telemetry from GPUs, networking, storage, schedulers and workloads to provide end-to-end observability for AI factories. It uses supervised and unsupervised ML to build a dynamic knowledge graph, detect anomalies and predict failures across large GPU clusters. Minerva includes a native AI assistant and supports external agentic AI tools for natural-language investigation and remediation. Partners including IREN, IBM, Jio, Lambda and Tesla

NVIDIA BlogOct 22, 2026Score 72

NVIDIAAI infrastructure managementAIOps

Industry

DeepSeek reportedly raises $15B days after US restrictions

DeepSeek has reportedly raised $15 billion, described as one of China's largest funding rounds, just days after the US ordered federal agencies to remove its technology over national security concerns. The report ties the funding to a period of heightened US restrictions on the company. Additional terms, investors, and details of the round were not disclosed in the source.

TechCrunch AIOct 22, 2026Score 70

DeepSeekAI fundingUS-China tech policy

Products

1Password lets AI agents sign in without exposing passwords

1Password announced a feature called 1Password for Agents that lets AI agents sign into user accounts without seeing or handling the passwords. Users must approve each sign-in request rather than giving agents open access to all connected accounts. CEO Aashish Tripathi framed it as replacing the old model of handing over a credential and hoping for the best. The feature aims to make agent-driven tasks that require service logins less cumbersome.

The Verge AIOct 22, 2026Score 69

1PasswordAI agent securitycredential management

Industry

Symantec finds self-directed AI agent hijacking exposed language models for cyberattacks

Security researchers at Symantec report an offensive AI agent that takes over unprotected language models exposed on the web and uses them to attack servers, steal databases and install cryptominers without human input. The agent can reportedly run up to 100,000 steps autonomously, and the owner of the hijacked model bears the resulting cost. The finding highlights the security risks of leaving model endpoints open to abuse.

The DecoderOct 22, 2026Score 78

AI agentscybersecurityLLM infrastructure

Research

OpenAI agents analyze millions of mouse brain cells in 32 minutes

OpenAI worked with PerturbAI and Scripps Research to deploy thousands of AI agents that analyzed biological data from millions of mouse cells. The reported analysis took 32 minutes, a task the article says would take humans months. The work focuses on mouse brain data.

The DecoderOct 22, 2026Score 73

OpenAIPerturbAIScripps Research

Products

Anthropic launches Claude Duet, an always-on proactive AI agent

Anthropic has released Claude Duet, an AI agent that performs tasks on its own even while the user is offline. The app runs on Claude Opus 5.5 and integrates with services including Gmail, Slack and Stripe through more than 6,000 connectors. It requests approval before sensitive actions such as payments. Anthropic positions Duet as a response to similar proactive agent products from xAI, Meta and OpenAI.

The DecoderOct 22, 2026Score 73

AnthropicClaudeAI agents

Models Editor selected

IBM Releases Granite Speech 5.0 2B, an Apache 2.0 Multilingual Speech Model

IBM released Granite Speech 5.0 2B, a 2B-parameter open speech-language model for transcription and translation in 6 languages under Apache 2.0. It combines a 16-block Conformer speech encoder with a window q-former and a fine-tuned granite-4.0-1b-base backbone, and adds language-ID prompting, keyword biasing up to 2,000 terms, and requested punctuation and capitalization. IBM reports 5.68% average English WER on the Open ASR Leaderboard, ranking second among open-weight models behind Qwen3-ASR-

Why it matters: Developers get a concrete, benchmarked open-weight ASR and translation release with deployment options and licensing details they can evaluate directly.

MarkTechPostOct 22, 2026Score 83

IBM Granitespeech recognitionopen-weight models

Industry

European banks build proprietary AI on open weight models and NVIDIA infrastructure

At GTC Berlin, European financial services organizations detailed production AI systems built on open weight models and NVIDIA accelerated computing. Revolut is rebuilding its stack around a single foundation model, PRAGMA, running in production for an initial 80 million customers, while Starling Bank, Lloyds and SBS described sovereign and transaction-focused intelligence platforms. Other deployments include bunq's agentic assistant Finn, which the company says matches leading closed models at

NVIDIA BlogOct 22, 2026Score 78

NVIDIAEuropean bankingopen models

Products Editor selected

NVIDIA Expands Cosmos Platform for World Model Training and Deployment

NVIDIA announced an expansion of its Cosmos platform for building world models used in physical AI systems, adding training tools, simulation checkpoints and Cosmos Data Factory blueprints. The release introduces multimodal multisensory data curation, support for adapting open models such as GR00T, Alpamayo, QWEN and Pi0.5, and updated Cosmos 3 NIM microservices. NVIDIA also described Cosmos 3 Nano variants for simulation-based policy testing and a Metropolis VSS 3.3 blueprint. The company cited

Why it matters: It gives AI developers a concrete, detailed look at new open tooling and deployment options for building physical AI world models.

NVIDIA BlogOct 22, 2026Score 81

NVIDIAworld modelsphysical AI

Industry

NVIDIA Details European Retail Deployments of Open Models and Physical AI

NVIDIA outlined how European retailers and brands are deploying its open models, inference software and digital twin tooling across commerce and logistics. Shopify post-trained NVIDIA Nemotron 3.5 Lightning for a Sidekick subagent, reporting better accuracy on an internal benchmark alongside 90% lower serving costs and 7x faster responses. Otto Group extended an NVIDIA Omniverse-based warehouse simulation effort to its largest fulfillment center in Germany, and L'Oréal is using NVIDIA ALCHEMI an

NVIDIA BlogOct 22, 2026Score 75

NVIDIAEurope retail AIopen-weight models

Products

NVIDIA Earth-2 Adds StormScope EU Model, Open Weather AI Adopted Across Europe

NVIDIA introduced StormScope EU, a regional short-term forecasting model trained on European weather and climate data that predicts cloud conditions across 16 Meteosat channels at 10-minute intervals using EUMETSAT observations. At GTC Berlin, NVIDIA demoed a live pipeline built with Earthmover and Earth2Studio that serves forecasts as map tiles as new satellite data arrives. The Earth-2 open model family is being applied by AXA for cyclone risk simulation, Vaisala Xweather for lightning-enhance

NVIDIA BlogOct 22, 2026Score 76

NVIDIA Earth-2AI weather and climate modelingStormScope

Research Editor selected

Perplexity and NVIDIA Researchers Introduce JOLT: Joint Teacher-Student RL for Post-Training

Perplexity and NVIDIA researchers propose JOLT (Joint On-policy Learning and Teaching), a post-training method where a single model is trained as both teacher and student, with the teacher receiving privileged information and also being optimized. JOLT+ adds the student's own outcome reward to the distillation signal. On MATH, JOLT matched GRPO accuracy while using 6.5x fewer processed training tokens according to the authors' fitted curves, and JOLT+ led GRPO across the five tested settings. Th

Why it matters: It offers a concrete post-training technique with claimed compute savings and benchmark comparisons that RL practitioners can evaluate.

MarkTechPostOct 22, 2026Score 83

PerplexityNVIDIAReinforcement Learning

Tutorial

Independent CPU Benchmark Pits Cactus Whistle Against Whisper Base for Speech-to-Text

A MarkTechPost coding guide benchmarks Cactus Compute's 16.9 MB Whistle speech-to-text model against Whisper tiny and base on seeded LibriSpeech test-clean and test-other samples, using speaker-cluster bootstraps with Holm correction. The author reports a lead over greedy Whisper base on test-other and no significant gap against beam search, with a speed advantage that narrows on longer clips. The tutorial also documents a 30-second single-pass window where longer audio is silently truncated, pl

MarkTechPostOct 22, 2026Score 70

Cactus ComputeOpenAI Whisperspeech recognition

Tutorial

NVIDIA Nsight Python Tutorial: Profiling GPU Kernels In-Process from Jupyter

MarkTechPost walks through NVIDIA's open-source nsight-python library, which attaches Nsight Compute to a running Python process so developers can annotate code regions and get hardware metrics back as a pandas DataFrame. The tutorial covers run-count and confidence intervals, cache and clock control, a fp32/fp16 matrix-size sweep with a derived TFLOP/s metric, and hardware counters for distinguishing compute-bound from memory-bound kernels. It also demonstrates the upcoming 1.1 features: an exp

MarkTechPostOct 22, 2026Score 68

NVIDIANsight PythonGPU performance profiling