NerdNews

Google Cloud has announced its post-quantum migration roadmap, targeting its first major risk domain for completion by the end of 2027. The roadmap includes interim deadlines for mitigating store-now-decrypt-later risk, hardening digital signatures, and rebuilding key management. Google Cloud API endpoints now offer quantum-safe key exchange, and application and proxy load balancers support hybrid key exchange for TLS 1.3.

OpenAI introduces Ultrafast, a new mode that makes GPT-5.6 Sol work at 14x the speed, delivering up to 750 output tokens per second. This high-octane version can be deployed across various corporate workflows, including incident response, customer service, and financial market analysis.

IBM and OpenAI have partnered to bring AI models and tools to enterprise customers. The deal will see IBM establish a dedicated OpenAI practice and integrate OpenAI's models into its AI platform. The partnership aims to expand OpenAI's enterprise business and help IBM accelerate its AI growth.

Microsoft is merging its consumer and business Copilot apps, discontinuing unsuccessful AI features like Group Chats and AI-generated podcasts. The goal is to create a simpler, more cohesive experience for users.

The US government will allow vetted private companies to launch offensive cyber operations against international criminal gangs and hackers. This move aims to combat cybercrime and threats targeting Americans, such as ransomware attacks and financial scams. Participating companies must deposit $1 million in escrow and comply with government rules, with operations conducted under federal supervision.

Microsoft is merging its Copilot AI into a single app, making it easier for users to access and utilize the AI features. The new app will combine the commercial and consumer ends of the Copilot experience, providing a simpler and more cohesive experience. The merger will also result in the deletion of some features, including the Podcasts feature.

AI agent security is an identity problem that starts as a secrets and credential problem. Many AI agents operate using static API keys, tokens, and other reusable credentials, bypassing traditional identity controls and creating a governance blind spot. To address this, it's essential to discover and attribute agent credentials, prevent new secret exposure, and migrate to scoped, short-lived access.

AI can identify threats, but it can't own security decisions. Probabilistic security is not enough, and explainable security controls are needed. Zero Trust for Code approach evaluates software behavior before execution and enforces consistent policy decisions.

Writer introduces Palmyra X6, a new AI model aimed at reducing deployment costs for its users. The model, combined with upgrades to the company's harness infrastructure, is expected to cut costs by up to 50% for basic tasks. This move is seen as a response to the growing concern over the high costs of AI deployments and the need for more efficient solutions.

X has open sourced its ranking algorithm, allowing users to see if they've been 'shadowbanned'. The algorithm is available on GitHub, and a new transparency tool will let users download their aggregate stats as a JSON file to see if any labels have been applied to their account or posts.

Percona launched a suite of consulting and support offerings to help enterprises solve performance and AI-readiness challenges. The services include Quickstart, Performance Optimization, and AI Readiness packages, which provide fast and measurable results within 30 days.

OpenAI has replaced its chief revenue officer with Dali Rajic, formerly of Wiz, as the company undergoes a broader executive shake-up. This change comes as OpenAI focuses on enterprise deployment and aims to increase its revenue, with products reaching over 1 billion weekly active users and 2 million businesses.

The use of AI coding assistants is increasing, but it also introduces a new challenge: the ingestion of unvetted dependencies. This can lead to slopsquatting and supply chain attacks. To secure the development pipeline, it's essential to govern the ingestion of dependencies at the point of selection, rather than relying on post-commit scans. ActiveState's Secure Open Source Library and Curated Catalog can help by providing pre-vetted, continuously remediated open source packages.

The White House has issued a memo allowing private security firms to conduct cyberattacks against overseas-based criminal organizations, with oversight from the Departments of Justice and Homeland Security. The program aims to combat transnational cyber-enabled crime, including ransomware and phishing campaigns.

Anthropic, the AI startup behind Claude, is expected to go public with a valuation of $2 trillion or more, driven by rapid revenue growth and booming demand for its advanced AI models and tools. The company's annualized revenue is projected to reach $100-120 billion by the end of 2026, with investors expecting a valuation of 30 times revenue, making it a potential $3 trillion company.

Rivian CEO RJ Scaringe is set to speak at TechCrunch Disrupt 2026, where he will detail the company's journey and lessons learned from designing, manufacturing, and delivering hardware at scale. Scaringe will discuss Rivian's investments in EVs, robots, and autonomy, including its roadmap to full Level 4 autonomy by 2028 and the launch of its R2 SUV.

Form Energy, a battery startup, has raised $750 million to expand its manufacturing capacity for 100-hour batteries, which can help bridge the power supply gap in renewable power generation. The company's unique iron-air battery chemistry stores massive amounts of electricity for less, and it has landed big customers like Google and Xcel Energy.

Anthropic, the developer of Claude, is reportedly setting a $2 trillion target for its October IPO. The company's investors are confident in its growth, citing projections of $100 billion to $120 billion in revenue by 2026. However, the company faces challenges such as high costs and regulatory hurdles.

Apple is in talks with publishers to pay for their content to power Siri's current news and information. The proposed compensation model would pay publishers when their content is used, rather than a fixed licensing fee. This move aims to enhance Siri's capabilities, expected to roll out later this year.

Trezor, a hardware wallet manufacturer, disclosed a data breach affecting nearly 14,000 customers after its shipping provider, ShipMonk, was hacked. The breach exposed customers' full names, shipping addresses, email addresses, and phone numbers. The company warned affected customers to be wary of phishing attempts.