Latest in AI

Apple said on October 2, 2026, that it will introduce additional controls around Full Disk Access in macOS, stating in a post on its Developer News site that some developers are using the permission in ways that could put users at risk and that the risks tied to that level of access will grow as AI agents become more capable and autonomous. What Apple Announced The post, titled "Updates to Full Disk Access in macOS," said Full Disk Access largely bypasses the controls that back Apple's…

NVIDIA announced a new 64GB configuration of DGX Spark — from Acer, ASUS, Dell, Gigabyte, HP and MSI — its GB10-powered desktop AI system. It gives developers a way to start with one system for local models and agents, then cluster two 64GB units for 128GB of memory across the cluster and more compute when […]

The post NVIDIA Announces DGX Spark 64GB: A 1-PetaFLOP Grace Blackwell Desktop for Local AI Agents, Fine-Tuning, and Inference appeared first on MarkTechPost.

Brian Hartzband, President of US Operations, GMEX Robotics, leads the company’s U.S. expansion, with a focus on building its commercial footprint, developing strategic partnerships, and bringing its physical AI and robotics platforms into real-world deployments. Prior to joining GMEX Robotics, Hartzband spent nearly four years at TEKsystems across technical recruiting and account management roles serving government and higher education clients. He previously co-founded and helped lead…

OpenAI published a customer case study on October 2, 2026, describing how Chatham Financial, a capital markets advisory firm, uses OpenAI's Codex and GPT-5.6 models across its advisory workflows, including a trade validation application the firm says reduced review time from approximately 30 minutes to under four. The case study lists Chatham as a small and medium-sized business in the finance industry in North America, with ChatGPT, Codex, and the OpenAI API as the products in use. The firm…

Anthropic on October 2, 2026 launched Claude Frontier Academy, a training initiative backed by a $100 million commitment that aims to train 10,000 Frontier Deployed Engineers by the end of 2027. Engineers from Accenture, Bain, Capgemini, Commonwealth Bank of Australia, Deloitte, McKinsey, Morgan Stanley, Novo Nordisk and other organizations make up the first cohorts. Anthropic described the Academy as the first of its kind from an AI company, built from its own work deploying Claude inside…

Fredrik Nilsson, Vice President, Americas, Axis Communications, oversees the company’s operations across North and South America, leading teams focused on developing and delivering network-based security solutions for partners and end users. A longtime Axis executive, Nilsson has played a significant role in the security industry’s transition from analog closed-circuit television to IP-based network video. Since taking responsibility for the Americas in 2003, he has helped grow regional revenue…

A mechanical engineer can know exactly what a part needs to do and still spend hours translating that intent into sketches, dimensions, and constraints. Hestus is betting that AI can take on more of that repetitive work—and has raised a $7.4 million seed round to build a purpose-built model for mechanical engineering. The San Mateo company announced the financing on October 2, with Insight Partners leading the round and participation from Chalfen Ventures, Rock Yard Ventures, Garage Capital…

The Allen Institute for AI (Ai2) said on October 2, 2026 that it is open-sourcing AstaBrief 8B, a model that turns a research question and retrieved literature excerpts into a cited report, releasing the model weights and training data as the system goes live as Fast mode in Asta, its agentic platform for scientific work. Fast Mode in Asta's Report Generation AstaBrief is available in Asta's Generate a report feature as Fast mode, running alongside the existing Claude-powered Thinking mode…

Content-based row matching, 6 per-value verdicts and a null rule make OmniExtractBench an extraction benchmark anyone can audit.

The post Datalab Introduces OmniExtractBench to Fix Bias and Opacity in Extraction Benchmarks appeared first on MarkTechPost.

Over the past few years, and counting, the AI infrastructure market has spent a great amount of time and money (close to $7 trillion of data-centre investment by 2030, by McKinsey’s estimate) learning how to build large data centres wherever power is available. “Wherever” being the key word because the location wasn’t very important. A training cluster simply needs access to sufficient power. But location is starting to become increasingly relevant nowadays as AI’s usage grows beyond chat…

AI agents are getting remarkably good at completing tasks. Give an agent a goal, access to the right tools, enough context, and a well defined workflow, and it can research information, analyze documents, make decisions, update systems, and coordinate with other agents. That is the exciting part of Agentic AI. It is also the part we tend to see in demos. Enterprise operations look different. A loan application changes halfway through underwriting. A customer provides new information after…

There is a popular vision of the digital future in which users verify themselves only once, digital credentials move freely between services, and KYC (Know Your Customer) fades into the background. It is not difficult to see why this is impossible in today’s hostile digital environment. Such an approach would create not only unnecessary transfers, but also a much broader exposure surface for identity data. In reality, KYC is unavoidable, so it has to become faster, safer, more private, more…

Tesla reported third-quarter 2026 production of 464,391 vehicles and deliveries of 486,532 vehicles on October 2, 2026, along with 13.7 GWh of energy storage products deployed, in a press release furnished to the U.S. Securities and Exchange Commission as an exhibit to a Form 8-K. Q3 2026 Model-Line Breakdown Model 3 and Model Y vehicles accounted for 457,387 units produced and 478,237 delivered in the quarter, according to the release, with 1% of those deliveries subject to operating lease…

World models learn an internal representation that predicts how an environment may evolve when an agent or other actor takes an action. This guide explains the mechanism, trade-offs, evaluation, and controls that matter in practice.

AKASA on October 2, 2026 announced the launch of an autonomous AI platform for the mid-cycle of the healthcare revenue cycle, expanding the South San Francisco company from AI-powered prebill review into autonomous inpatient medical coding and clinical documentation integrity (CDI). AKASA, which describes itself as a leader in generative AI for the healthcare revenue cycle, said its customers represent more than $180 billion in aggregate net patient revenue and roughly 10% of U.S. inpatient…

Amazon on October 2, 2026 announced Built Together, a community investment framework that will add more than $1 billion over five years to the company's existing spending in US communities where it builds and operates data centers. The commitment came in an About Amazon post written by Amazon Web Services CEO Matt Garman, which also introduced a codified Amazon Data Center Commitment covering how the company builds and operates those facilities. Garman opened by likening the AI-driven data…

The question isn't whether an AI-built form can look ready for production. It's whether the system receiving its data will agree. A date picker can display perfectly and still submit a locale-dependent string when the API expects an ISO date. A checkbox can say yes or no while the database expects a Boolean. The demo passes, the screenshot looks clean, and the failure waits downstream. What Is the Form Actually Promising? Form design is usually reviewed as an interface problem. Can people…

AWS's Strands Agents team released Strands Decider 2B, an Apache-2.0 decision model built on Qwen3.5-2B-Base. It returns choices, yes/no probabilities and scores with calibrated confidence in one forward pass, never text. It runs at a 115 ms median on an RTX 3090 and scores 0.723 on the JevBench public set, which makes it a fast local option for routing, tool selection and guardrails in AI agents.

The post AWS Strands Labs Releases Strands Decider 2B: An Open Source Decision Model That Picks Options in About 115 ms appeared first on MarkTechPost.

Cloudflare has released Clef (27B) and Clef-flash (9B), open-weight decision models that return typed probabilities instead of text. They are Jev-API compatible, accept images, and run on Workers AI at 209.3 ms and 38.8 ms median latency.

The post Cloudflare Releases Clef and Clef-flash: Open-Weight Decision Models That Return Typed Probabilities Instead of Text appeared first on MarkTechPost.

In this comprehensive coding guide, we explore Google Research's Kauldron—a JAX training library optimized for research velocity and modularity. Learn how konfig turns experiments into plain dictionaries, kontext wires components via string paths, and ktyping enforces runtime shape checks.

The post A Coding Guide to Google Research’s Kauldron: Configs That Are Plain Data, Components Wired by String, and a JAX Trainer You Can Read End to End appeared first on MarkTechPost.