Latest in AI

A group of House Democrats led by Rep. Greg Casar of Texas asked House Speaker Mike Johnson on August 10, 2026 to bring the CEOs of OpenAI, Anthropic, and other major AI companies before Congress to testify under oath about a series of incidents in which the companies' own models hacked into outside organizations during safety testing, CNBC reported from a copy of the letter shared with the network. The lawmakers pointed to the breaches, all disclosed by the companies themselves over the past…

Why organizations that redesign customer journeys before implementing AI will outperform those that don't Over the past year, AI has produced a series of high-profile customer experience failures. Pizza Hut is being sued for $100 million after a franchisee alleged an AI-powered delivery optimization system increased delays and hurt customer satisfaction rather than improving it. Air Canada was held responsible after its automated AI agent provided customers with incorrect information about…

London-based Edgify has raised $9 million in Series A+ funding to expand its edge artificial intelligence platform, bringing the company’s total funding to $25 million. The round was backed by Rank Ventures and Mangrove Capital Partners. Edgify plans to use the capital to accelerate deployments across physical retail while developing its platform into a broader infrastructure layer for managing AI models across distributed devices. The company has initially found traction in grocery retail…

HD Hyundai Heavy Industries has signed a contract worth $673.8 million to supply 1,000 megawatts of power generation systems to U.S. data centers, the company announced on August 9, 2026. The buyer is Corban Energy Group, a U.S. energy infrastructure developer, and the equipment will feed data centers operated by what HD Hyundai describes as a major U.S. technology company. The systems are built around the company's 9.6-megawatt HiMSEN engine, a medium-speed engine it positions for continuous…

Firebird opened its first AI factory in Hrazdan, Armenia, on August 8, 2026, and used the ceremony to lay out the rest of the map: a second market in Kazakhstan with 125 megawatts of secured capacity, an intended investment from NVIDIA, and a pipeline the company says will reach 2 gigawatts of AI infrastructure by the end of 2028, per its announcement. The Armenia site is the operational debut. Firebird plans to scale it beyond 70,000 NVIDIA Rubin and Blackwell GPUs and 300 megawatts of…

OpenAI has acquired NextSlide, a startup that built software for turning raw material into finished presentations, with the NextSlide team now working on ChatGPT. Founder Ahmed Beshry announced the move in a note on the startup's website. He did not disclose financial terms. NextSlide's product converted prompts, notes, documents, or research into a polished, editable presentation, the pitch being that producing a deck should not depend on design skills or hours of manual layout work. Beshry…

Pokee AI released Pokee-Isaac 28B, a 28B text-only foundation model with a 10M-token context window built to run inside the customer boundary. It scores 93.3% on RULER at 10M tokens, where every baseline in its comparison panel returns 0.0 beyond 2M, and leads BFCL v4 at 70.94 while placing second on Terminal-Bench 2.1. Prefill reaches 137,200 tokens/s at full context on a single B200, with decode flat near 335 tokens/s. Weights are not published; deployment is licensed into VPC, on-premises, or on-device, with list pricing at $0.15/$1.00 per million tokens.

The post Pokee AI Releases Pokee-Isaac 28B: A 10M-Token Context Agentic Model Built to Run Inside the Customer Boundary appeared first on MarkTechPost.

Firebird, a U.S.-based AI cloud company, opened the CIS region's largest AI factory in Armenia on August 8, 2026: a facility in Hrazdan built on NVIDIA accelerated computing and Dell PowerEdge servers, with Armenian Prime Minister Nikol Pashinyan, Kazakhstan's Deputy Prime Minister Zhaslan Madiyev, and U.S. chargé d'affaires David Allen attending the opening ceremony. The buildout plan is large by any regional standard: Firebird plans to deploy more than 70,000 NVIDIA Rubin and Blackwell GPUs…

Dive into the advanced visualization capabilities of the Reflex XY Python library. This tutorial guides you through building high-performance, interactive charts—from handling million-point datasets and real-time streaming to creating custom mark plugins and exporting publication-ready visuals.

The post Designing Scalable Interactive Visualizations with Reflex XY: Composition, Million-Point Rendering, Streaming, Custom Marks, and Export appeared first on MarkTechPost.

Mistral AI has released Shieldstral 1.0 3B, an open-weights, policy-adaptive multimodal safety classifier that frames content moderation as a single yes/no question instead of a fixed harm taxonomy. Operators supply the policy as a plain-language query at inference time and get back a calibrated safety score from one forward pass — no retraining required to re-target the model. Built on Ministral-3-3B-Base-2512 with a Pixtral vision encoder and trained on roughly 54.1M samples, it reports 84.9% average F1 on text safety (matching GPT-OSS-Safeguard-20B), 83.8% on multimodal safety, and 91.3% on Mistral's adaptability benchmark — while fitting in 16GB of VRAM under an Apache 2.0 license.

The post Mistral AI Releases Shieldstral 1.0 3B: An Open-Weights Policy-Adaptive Multimodal Safety Classifier Matching Models 7× Its Size appeared first on MarkTechPost.

xAI shipped Grok Imagine Image 2.0 on August 7, 2026, making the new model generally available as the Quality Mode on grok.com/imagine and inside the company's iOS and Android apps, according to the company's announcement. The release lands as xAI lists its models under the SpaceXAI name, and it arrives with region-level editing tools, multi-image reference inputs, and a claim to the number-two spot on both major image arenas. What Image 2.0 adds The release is built around editing as a…

I’ve spent way too much time doing tasks that feel like they should be someone else’s job, whether it's formatting reports, organizing information, or repeating the same steps in different apps. At some point, you start thinking, “Maybe I should just hire a VA.” That’s why Claude Cowork’s Record a Skill feature caught my attention. Instead of writing out a giant list of instructions for an AI, you can record yourself doing a task, explain what you’re doing as you go, and have Claude turn your…

Tencent Cloud has open-sourced TencentDB Agent Memory v2.0, a team-level memory hub that turns conversations, documents and code into four governed, reusable assets — Chat Memory, Skill, LLM-Wiki and Code-Graph. It is MIT-licensed, self-hosted via Docker, and integrates with Claude Code, OpenClaw, Hermes and CodeBuddy. The differentiator is not retrieval but governance: ACL-based visibility decides which agent gets which asset, and which version is valid.

The post Tencent Cloud Open-Sources TencentDB Agent Memory v2.0: A Team-Level Memory Hub for AI Coding Agents appeared first on MarkTechPost.

In this tutorial, we build an advanced multimodal retrieval-augmented generation pipeline with NVIDIA NeMo Retriever. We begin by configuring a Python 3.12 environment, installing the required packages, and performing offline PDF text extraction without relying on a GPU or external API key. We then extend the workflow with hosted NVIDIA NIM endpoints to detect page […]

The post Building a Multimodal RAG Pipeline with NVIDIA NeMo Retriever, Hosted NIMs, LanceDB, Reranking, and Grounded Generation appeared first on MarkTechPost.

NVIDIA Labs has open-sourced NOOA (NVIDIA Object-Oriented Agents), a model-agnostic Python framework for building AI agents. Agent development today is split across prompt templates, tool schemas, callback code, and workflow graphs. NOOA collapses all of it into one Python class. Methods are the actions the model can take. Fields are agent state. Docstrings are prompts. […]

The post NVIDIA AI Releases NOOA: An Object-Oriented Python Framework That Turns an AI Agent Into a Single Python Class appeared first on MarkTechPost.

Microsoft has brought its fourth Indian cloud region online, declaring general availability of the India South Central region in Hyderabad, Telangana on August 6, 2026. The three-availability-zone build is the largest hyperscale region Microsoft operates in the country and, by the company's count, gives it the biggest hyperscale cloud footprint of any provider in India. The region is the first concrete capacity delivery against the US$17.5 billion India commitment Microsoft announced on…

OpenAI said on August 7, 2026 that internal evaluations of Astra, one of its upcoming models, show such strong agentic coding and cybersecurity performance that the company can no longer rule out the Critical cybersecurity capability level defined in its own Preparedness Framework. It is the first time OpenAI has attached that label's possibility to a specific model, and it triggered immediate containment steps: stricter security controls, a pause on some internal Astra work, and plans to bring…

AI agent reliability startup Lemma has raised $2.3 million in pre-seed funding to build monitoring infrastructure designed to catch a particularly difficult class of problem: AI agents that appear to have completed a task successfully while quietly producing the wrong result. The round includes participation from Matrix, Y Combinator, Liquid 2 Ventures, Vermilion Cliffs Ventures, Irregular Expressions, Cervin Ventures, Comma Capital, Position Ventures, and Eight Capital, alongside angel…

For the last two years, the enterprise AI conversation has been dominated by one question: how much more productive can we make the developer? The metrics have been familiar: lines of code generated, hours saved, user stories completed faster and pull requests merged sooner. These are useful indicators, but they tell only part of the story. In large enterprises, software delivery is rarely constrained by coding alone. It is constrained by context, governance, architecture, security, domain…

Anthropic on August 7, 2026 released an update to the biology safeguards on Claude Fable 5 that the company says cut biology-related "fallbacks" by about 85% in testing across its product surfaces — the automatic handoffs that route a user's query to a less capable model when the system judges a request touches safeguarded biology territory. In its announcement, Anthropic said users should now see far fewer fallbacks on everyday health and educational questions, such as interpreting lab…