Tech News
← Home  ·  All topics

Agentic

214 GoKawiil briefs on this topic · RSS for agentic

Norwest closes $4B fund, its largest, to expand AI investments

Norwest, the Wells Fargo-backed venture firm, announced Thursday it has closed its 18th fund at $4 billion, the largest in its 65-year history. The firm said it will deploy the capital across the US, India and Israel, targeting AI startups in consumer, healthcare and tech, including physical AI, voice AI, agentic cybersecurity and aerospace.

Jane Street adds new interview rounds testing design skills and AI-assisted coding

Jane Street says it is trialing new software engineering interview formats for campus candidates, starting this summer. These rounds involve less hands-on code writing and more focus on software design, code quality verification, and using LLMs to produce code, while most rounds remain traditional coding interviews done without AI assistance.

Coinbase's Armstrong addresses Binance hack fallout and shifting crypto politics

In his first TV interview since Binance was hacked, Coinbase CEO Brian Armstrong discussed cybersecurity vulnerabilities in crypto, the prolonged market slump, and the political landscape following midterm elections that gave Democrats control of Congress. He noted Coinbase blocked transfers to compromised wallets and offered direct assistance to Binance during the breach, and called for greater industry collaboration in such moments.

Google releases Modern Web Guidance 1.0 for coding agents, with Firefox and Edge on board

Google has launched version 1.0 of Modern Web Guidance, a toolkit of skills that feeds web platform expertise and browser compatibility data directly into AI coding agents. The release adds a codebase-scanning modernization workflow, an expanded database of verified guides, stricter human expert and accessibility review, and a public evaluation dashboard. Firefox and Edge are now joining Chrome as contributors to the project, building on existing Baseline compatibility work.

United Airlines to block AI agents from booking flights

United Airlines said it will update its policies to prevent AI agents from booking flights through its website, third-party travel agencies, and flight aggregators. The airline cited risks like unauthorized purchases, account changes, or misuse of MileagePlus miles, as well as regulatory requirements that only humans can confirm certain safety checks during check-in.

Google DeepMind formally verifies core cryptographic guarantees of Signal protocol

Google DeepMind researchers used an AI system called AlphaProof Code to mathematically prove several key cryptographic guarantees underlying the Signal protocol, completing the project in about four weeks. The analysis did not uncover new vulnerabilities and covered only portions of the protocol, not the full Signal mobile app.

Open-source pen-test tool Artex adopted by threat actors for automated attacks

Developer Autumn-27 released Artex, an open-source multiagent AI penetration-testing framework written in Go, on GitHub in July. By late summer and fall, multiple threat actors were using the tool — which can brute-force passwords, dump databases and abuse APIs — to automate attacks at high speed; Microsoft observed an AI agent go from scanning GitHub for secrets to exfiltrating data from an Azure Key Vault in five minutes. The creator deleted the original repository in late October citing abuse, but forks remain active.

Google Cloud launches Gemini for Healthcare in preview

Google Cloud has announced Gemini for Healthcare, an agentic AI platform built on its Gemini model to automate administrative and operational tasks in medical settings. The system connects to electronic health records, billing platforms, Medicare/Medicaid data and clinical guidelines to generate multi-step workflows, with human oversight built in. Early customers include Blue Cross Blue Shield of Massachusetts, Hartford HealthCare, HCA Healthcare, Highmark Health, Northwell Health, Optum and Stanford School of Medicine.

AMD executives dispute Nvidia and Arm 'agentic AI' CPU design claims

AMD's chief Zen architect Mike Clark and corporate VP Madhu Rangarajan said in separate sessions that the company did not design its upcoming Zen 6 (Venice) EPYC CPUs specifically for agentic AI, since that workload trend only emerged around March this year. They contrasted this with Nvidia and Arm, who have described their newest data center CPUs as 'purpose-built' or 'designed for' agentic AI despite similarly long development cycles.

Adobe adds planning mode and always-on agents to Firefly AI Assistant

At its Max conference in Miami, Adobe announced an expanded Firefly AI Assistant that can execute multi-step tasks across Creative Cloud apps. A new planning mode lets users review and approve the agent's proposed workflow before it runs, choose which model orchestrates tasks, and set brand guidelines like colors for consistency. Adobe also previewed always-on agents for recurring work, such as sorting incoming footage, and a new natural-language editing agent coming to the Premiere mobile app in private preview.

Adobe Unveils Firefly Video Editor and Design Tool at Adobe Max

At Adobe Max, Adobe announced updates to its Firefly AI suite focused on video editing, including an upgraded video editor with a built-in AI assistant, ideation panel, and storyboarding tools. The company also introduced Firefly Design, a graphic design tool covering print, motion graphics and 3D, alongside new AI image tools for relighting, retouching and editing objects.

Adobe previews MCP agent integrations and always-on bots for Creative Cloud

At its Max conference in Miami, Adobe unveiled local Model Context Protocol integrations letting third-party AI agents from OpenAI, Google and Anthropic directly edit native files in Photoshop, Illustrator, Premiere and After Effects, while keeping layers and sequences manually editable. The company also showed 'always-on' agents that can run recurring tasks like asset organization or quality checks in the background, pending user approval of outcomes. Adobe said the Creative Cloud local MCP integrations will enter private beta in early 2027.