Latest AI News

Anthropic has made Claude Cowork generally available on all paid plans, adding enterprise controls for organization-wide deployment. New features include role-based access controls with SCIM support, group spend limits, usage analytics via dashboard and Analytics API tracking DAU/WAU/MAU, and expanded OpenTelemetry support compatible with Splunk and Cribl pipelines. A new Zoom MCP connector brings meeting summaries and action items into Cowork workflows. Admins can also restrict per-tool connector permissions org-wide from the admin console.

Perplexity has expanded its Plaid integration to let users securely link bank accounts, credit cards, mortgages, auto loans, and student loans alongside previously supported brokerage accounts. The read-only connection covers 12,000+ financial institutions including Chase, Fidelity, and Schwab, with user data never touching Perplexity's servers. Using Perplexity Computer, users can analyze spending, calculate net worth, and build custom dashboards or debt payoff plans via freeform questions. The feature is live on desktop for US and Canada users, with advanced tools limited to Pro and Max subscribers.

Upwork has launched a ChatGPT app that lets businesses describe project needs and find and hire talent from the platform's 18 million professionals without leaving the chatbot. Users can draft job posts inside ChatGPT, then move to the Upwork Marketplace to handle compliance, payments, and contracts, with Upwork's AI agent Uma available for additional help. The integration supports OpenAI's push to make ChatGPT an all-in-one hub, which already serves over 900 million weekly users.

OpenAI is updating its ChatGPT subscription tiers to better accommodate growing demand for Codex, its AI-powered coding tool. The company is introducing a new Pro tier priced at $100 per month, which offers five times more Codex usage than the existing Plus plan. The Pro tier is designed for longer, high-effort Codex sessions, making it well-suited for power users who rely on extended coding workflows. The update signals OpenAI's push to serve more intensive developer use cases within ChatGPT.

A D.C. federal appeals court denied Anthropic's request to pause the Pentagon's supply-chain risk designation, dealing a setback to the AI company despite a separate San Francisco court granting a preliminary injunction last month blocking bans on Claude. The split rulings mean Anthropic still faces ongoing legal battles. The Pentagon can still exclude Anthropic from new contracts under a separate statute, though it is expected to continue using Anthropic products for the next six months.

Meta has published an updated Advanced AI Scaling Framework and previewed a Safety & Preparedness Report for its Muse Spark model. The framework broadens risk evaluations to include chemical, biological, cybersecurity, and loss-of-control risks, and applies across open, API, and closed deployments. For Muse Spark, Meta tested the model before and after applying safeguards against thousands of adversarial scenarios, confirming it lacks autonomous capability sufficient to pose control risks and shows strong ideological balance in responses.

Google Colab has added two new features to its Gemini AI integration: Learn Mode and Custom Instructions. Learn Mode turns Gemini into a personal coding tutor, providing step-by-step guidance and concept explanations instead of simply generating copy-paste code. Custom Instructions let users tailor Gemini's behavior to specific coding styles, class syllabi, or preferred libraries, and are stored at the notebook level so they travel with shared notebooks, giving collaborators the same personalized AI experience.

Factory.ai has launched a native desktop app for its AI Droids, available now on macOS and Windows across all plans. The app supports multi-agent sessions, letting users run several Droids simultaneously with separate context and history. Key features include Droid Computers with persistent storage, BYO machine registration, local model support via Ollama or vLLM, computer use capabilities, and VS Code integration. Enterprise data shows users running both CLI and desktop interfaces complete 4.6x more sessions than CLI-only users.

Anthropic has redesigned its Claude Managed Agents hosted service by decoupling three core components: the harness (brain), sandbox (hands), and session log. Previously coupled in a single container, the architecture caused debugging difficulties and security risks. The new design lets each component fail or be replaced independently. A harness crash is recovered via wake(sessionId), while sandboxes become disposable cattle. The restructure cut median time-to-first-token by 60% and p95 latency by over 90%, and eliminated credential exposure in sandboxes.

A hacker using the name FlamingChina allegedly breached China's National Supercomputing Center in Tianjin, claiming to have stolen over 10 petabytes of sensitive data including classified defense documents, missile schematics, and aerospace research. The attacker reportedly gained access via a compromised VPN and used a botnet to extract data over six months undetected. Cybersecurity experts at SentinelOne who reviewed samples called them credible. Full dataset access is priced at hundreds of thousands of dollars in cryptocurrency.

OpenAI has released a Child Safety Blueprint, a policy framework aimed at combating AI-enabled child sexual exploitation. The blueprint focuses on three priorities: modernizing laws to address AI-generated CSAM, improving provider reporting and coordination with law enforcement, and embedding safety-by-design measures into AI systems. Developed with partners including NCMEC, Thorn, and the Attorney General Alliance—co-chaired by North Carolina AG Jeff Jackson and Utah AG Derek Brown—the framework targets prevention before harm occurs.

Cursor has launched a new feature allowing users to run AI coding agents on any machine and control them remotely from a phone or other device. Developers can now kick off agents from their phone to run on a remote development box, enabling asynchronous, location-independent workflows. This means coding tasks can be delegated and executed on powerful remote machines without requiring the user to be physically present at their workstation.

HeyGen has launched Avatar V, a new AI avatar tool that captures a user's identity in just 15 seconds and maintains character consistency across all generated videos. Users can then modify the look, outfit, and setting to produce unlimited personalized video versions while preserving their core identity. The feature addresses a longstanding challenge in AI video generation: keeping a consistent character appearance across different scenes and styles.

CapCut has launched Dreamina Seedance 2.0 in the United States, its AI video generation tool, available across the CapCut app, desktop, and web platforms. Every user receives one free trial of the new model. New users can also claim 90% off their first month of CapCut Pro for a limited time. The rollout marks a significant push into the competitive AI video generation market as CapCut expands its creative AI offerings to American audiences.

The CIA used a classified tool called Ghost Murmur to locate a downed F-15 weapons systems officer hiding in a mountain crevice in Iran. Developed by Lockheed Martin's Skunk Works division, the technology uses long-range quantum magnetometry — sensors built around microscopic defects in synthetic diamonds — paired with AI to detect a human heartbeat's electromagnetic signature from up to 40 miles away. It was the tool's first field deployment, referenced publicly by President Trump and CIA Director John Ratcliffe at a White House briefing.

Anthropic has published a system card for Claude Mythos Preview, its most capable frontier model to date, which shows striking benchmark improvements over Claude Opus 4.6. Due to powerful cybersecurity capabilities that could enable sophisticated offensive exploits, Anthropic decided against general release. Instead, access is restricted to partner organizations maintaining critical software infrastructure under Project Glasswing, limited to defensive cybersecurity uses. The card notes Mythos Preview is Anthropic's best-aligned model yet, though rare misaligned actions remain concerning given its high capability level.

Anthropic has launched Project Glasswing, a cybersecurity initiative uniting AWS, Apple, Google, Microsoft, Nvidia, Cisco, CrowdStrike, JPMorganChase, and others to defend critical software using Claude Mythos Preview, an unreleased frontier model. Mythos Preview has autonomously found thousands of zero-day vulnerabilities across every major OS and browser, including a year-old OpenBSD flaw and a year-old FFmpeg bug. Anthropic is committing $100M in usage credits and $4M in donations to open-source security organizations.

Anthropic has launched auto mode for Claude Code, a new permission system that uses model-based classifiers to replace manual approval prompts. Since users accept 93% of prompts anyway, auto mode automates safe decisions while blocking dangerous ones like credential harvesting, scope escalation, and safety-check bypasses. It uses a two-layer defense: a prompt-injection probe on inputs and a Sonnet 4.6 transcript classifier on outputs, with a fast single-token filter followed by chain-of-thought reasoning only when needed. Configuration is available via Claude Code docs.

Spotify is expanding its AI-powered Prompted Playlists feature to include podcasts, allowing premium users in the U.S., Canada, U.K., Ireland, Australia, and Sweden to generate podcast playlists using natural language prompts in English. Originally tested in New Zealand for music in late 2025, the beta feature lets users describe what they want, such as highly rated true crime series, and choose update frequencies of daily or weekly. Each added episode includes a short note explaining why it was recommended.

X has launched a brand new photo editor built directly into its post composer, bringing long-requested features including drawing and text tools. Two additions are exclusive to X: an AI-powered feature called Edit with Words, powered by Grok, which lets users modify photos using written descriptions, and a redaction blur tool that allows users to obscure sensitive parts of an image before posting. The update was announced by X team member Nikita Bier.

OpenAI has launched Paper Review, a new AI workflow within its Prism platform, designed to evaluate technical and scientific papers. Announced by OpenAI CPO Kevin Weil, the tool aims to improve scientific rigor, correctness, and reproducibility. The feature is positioned as the opposite of AI-generated low-quality content, using AI instead to strengthen the quality and reliability of scientific research submissions and peer review processes.

HappyHorse-1.0 is a new video generation model that has debuted at the top of the leaderboard in both text-to-video and image-to-video categories. Early testing by Justine Moore highlights its strength in generating multi-shot videos and following detailed directional prompts, setting it apart from competing models. Its simultaneous lead across two major benchmarks marks a notable debut in the competitive AI video generation space.

The accessible source details point to this update: World Labs Rolls Out Marble 1.1 and Marble 1.1-Plus Model Updates. Because the full article could not be reliably extracted or rewritten, this TLDR stays conservative and is based on headline-level information plus limited source context from World Labs on X.

Intel has joined Elon Musk's Terafab AI chip project alongside SpaceX and Tesla, aiming to produce processors for robotics and data centers. The partnership targets 1 terawatt per year of compute capacity, with two advanced chip factories planned in Austin, Texas — one for cars and humanoid robots, another for space-based AI data centers. Intel shares jumped nearly 3% on the news. CEO Lip-Bu Tan met Musk at Intel's campus, calling Terafab a step change in semiconductor manufacturing.

Runway has integrated Seedance 2.0 into its platform, bringing multi-shot video generation capabilities to its users. The model accepts text, image, video, or audio as inputs and supports full sound design and dialogue, enabling creators to produce complete video sequences within a single workflow. Seedance 2.0 is currently available exclusively to Runway Unlimited plan subscribers and Enterprise accounts, though access remains restricted to users located outside the United States.

GitHub has launched Rubber Duck in experimental mode for GitHub Copilot CLI, a cross-model review agent that uses a second AI from a different model family to critique coding plans and implementations. When Claude models serve as the orchestrator, Rubber Duck runs on GPT-5.4. Benchmarks on SWE-Bench Pro show Claude Sonnet 4.6 paired with Rubber Duck closes 74.7% of the performance gap between Sonnet and Opus. Access it via the /experimental slash command in Copilot CLI.

Anthropic has signed a new agreement with Google and Broadcom for multiple gigawatts of next-generation TPU compute capacity expected to come online starting in 2027. The deal will power Anthropic's frontier Claude models amid explosive growth: run-rate revenue has surpassed $30 billion, up from $9 billion at end of 2025, and business customers spending over $1 million annually have doubled to more than 1,000 in under two months. Most new compute will be sited in the United States.

OpenAI has formally written to California Attorney General Rob Bonta and Delaware AG Kathy Jennings, urging investigations into what it calls anti-competitive behavior by co-founder Elon Musk. The move comes ahead of an April 2026 trial between the two parties and signals a strategic shift from defensive litigation to offensive regulatory pressure. OpenAI alleges Musk's control of xAI, his Grok chatbot on X, and his Tesla board seat create unfair competitive advantages against rivals like ChatGPT.

Google has quietly launched Google AI Edge Eloquent, a free offline-first AI dictation app for iOS, competing with apps like Wispr Flow and SuperWhisper. Powered by on-device Gemma-based speech recognition models, it transcribes speech in real time while automatically removing filler words like 'um' and 'ah.' Users can reformat text with options like Formal, Short, or Key Points. An optional cloud mode uses Gemini models for cleanup. The app can also import custom keywords from Gmail. An Android version with system-wide keyboard access is referenced in the App Store description.

Meta plans to open-source versions of its next AI models, developed under Alexandr Wang, who joined the company via a $15 billion Scale AI deal. While Meta remains the largest U.S. company to allow modification of frontier models, it will adopt a hybrid strategy, keeping its largest models proprietary. Wang aims to democratize AI access and counter Anthropic and OpenAI's enterprise focus by targeting consumers through WhatsApp, Facebook, and Instagram. The new model family is designed to help Meta catch up after Llama 4 fell behind rivals.

OpenAI has launched the OpenAI Safety Fellowship, a pilot program for external researchers, engineers, and practitioners to conduct safety and alignment research on advanced AI systems. The program runs September 14, 2026 through February 5, 2027, with workspace available in Berkeley at Constellation alongside remote options. Priority areas include safety evaluation, ethics, robustness, agentic oversight, and misuse domains. Fellows receive a monthly stipend, compute support, API credits, and mentorship. Applications close May 3, with successful applicants notified by July 25.

"Ben Sigman and actress Milla Jovovich have released MemPalace, an AI memory system built using Anthropic's Claude that claims to have achieved the first perfect score on LongMemEval, the standard benchmark for AI memory evaluation. The system reportedly outperforms every competing product, both free and paid. Sigman describes MemPalace as fundamentally different in architecture from existing memory solutions, though the post was cut off before elaborating on its technical approach."

OpenAI's new image model GPT-Image-2 has reportedly leaked and is available on the Arena platform under three aliases: maskingtape-alpha, gaffertape-alpha, and packingtape-alpha. Early impressions suggest the model demonstrates strong world knowledge and improved text rendering capabilities, potentially surpassing previous models. The leak was flagged by developer Pieter Levels on X, who noted the model may outperform existing options, though no official confirmation from OpenAI has been provided.

OpenAI is seeing significant C-suite reshuffling. Fidji Simo, CEO of AGI deployment, is taking medical leave for several weeks due to a neuroimmune condition, with President Greg Brockman stepping in to lead product and the super app effort. CMO Kate Rouch is also stepping down for health reasons, with Gary Briggs temporarily replacing her. COO Brad Lightcap is transitioning to a special projects role reporting to Sam Altman, with CRO Denise Dresser absorbing much of his responsibilities.

Starting tomorrow at 12pm PT, Anthropic's Claude subscriptions will no longer cover usage on third-party tools such as OpenClaw. Users can still access these tools through their Claude login by purchasing extra usage bundles, which are now available at a discount, or by using a Claude API key directly. The change was announced by Boris Cherny and affects how third-party integrations are billed under existing Claude subscription plans.

A leaked OpenAI cap table circulating on social media reveals enormous paper gains for early investors and partners. Microsoft reportedly turned its $13 billion initial investment into $215 billion in value, representing an 18x return. Ashton Kutcher's VC fund achieved a 43x return, qualifying it as a fund-returner. OpenAI's non-profit arm holds an estimated $220 billion in gains at a zero cost basis. None of these figures have been independently verified by outside sources.

Pika Labs has launched a beta video chat skill compatible with any AI agent, powered by its new real-time model PikaStream1.0. The skill adds a face and voice to agent interactions while preserving memory and personality across conversations, enabling more human-like real-time communication. This marks a significant step toward embodied AI agents, as any agent can now integrate the skill to support live video chat, moving beyond text-based interactions toward a more expressive and continuous conversational experience.

Nous Research has released Hermes Agent v0.7.0, featuring a major update that transforms memory into an extensible plugin system. Users can now swap in any memory backend or build their own custom solution. Built-in memory works out of the box, and six third-party memory providers are available immediately. Users can select a provider using the command hermes memory setup. The full changelog is available on the Nous Research GitHub.

Anthropic's Claude has added Microsoft 365 connectors across all subscription plans, allowing users to link Outlook, OneDrive, and SharePoint directly to their conversations. The integration lets Claude access emails, documents, and files from these services, enabling more context-aware assistance without switching between apps. The feature is available to all Claude plan tiers, making it accessible to both free and paid users.