News — 2026-04-23

Anthropic announced Thursday that its Claude AI assistant now supports more than 200 third‑party connectors, adding a slate of consumer‑focused services such as Instacart, Spotify, Uber and TripAdvisor. The new integrations let users retrieve data and trigger actions directly from chat, while the system automatically suggests the most relevant connector based on conversation context. Anthropic says the feature remains ad‑free, that user data won’t be used to train models, and that users must confirm any major transaction before it completes. Read more

OpenAI rolled out ChatGPT 5.5 this week, extending the new model to paying ChatGPT and Codex subscribers and promising an API release soon. Marketed as a work‑focused upgrade, the model excels at coding, computer‑task orchestration and research‑intensive workloads. President Greg Brockman highlighted its ability to interpret vague problems and act with minimal human direction, positioning it as a step toward true digital assistants. OpenAI also emphasized stronger cybersecurity safeguards, reflecting growing concerns about AI‑driven threats. Read more

SpaceX announced Wednesday that it has entered a partnership with AI coding platform Cursor, giving the rocket firm access to the company’s agentic coding technology and a right to acquire Cursor later this year for up to $60 billion. The deal hinges on SpaceX’s Colossus supercomputer in Memphis, Tennessee, which boasts the equivalent of a million Nvidia H100 GPUs, and aims to create the world’s most capable coding and knowledge‑work AI. Both companies say the collaboration will accelerate model development and could eventually feed into Elon Musk’s broader X ecosystem. Read more

OpenAI announced the rollout of its latest language model, GPT-5.5, on Thursday. The company describes the new system as its "smartest and most intuitive" model yet, capable of handling complex, multi‑step tasks such as writing and debugging code, conducting online research, and generating spreadsheets across a range of tools. GPT-5.5 uses fewer tokens, offers its strongest safety measures to date, and will be available to Plus, Pro, Business and Enterprise ChatGPT users as well as Codex subscribers. The launch intensifies the rivalry with Anthropic and comes just days before a high‑profile trial involving OpenAI executives. Read more

Anthropic disclosed that a small group of unauthorized users gained access to its newly released Claude Mythos model on the day the company announced a limited rollout. According to Bloomberg, the intruders guessed the model’s online location using details leaked from a prior breach at data‑training firm Mercur and insider knowledge from a contractor who had evaluated Anthropic’s models. Anthropic said it is investigating the incident and reviewing its monitoring systems, which were designed to log and track model usage. The breach, described by security researchers as a standard “educated guess” attack rather than a sophisticated exploit, did not appear to target the model’s advertised cybersecurity capabilities. The episode raises questions about the robustness of Anthropic’s security controls for a product it has marketed as a “watershed moment” for defending digital infrastructure. Read more

OpenAI unveiled Images 2.0, its latest AI image generator that can reason about composition, search the web for context, and produce up to eight coherent images from a single prompt. The model claims 99% accuracy in rendering text across multiple scripts and surged to the top of the Image Arena leaderboard within 12 hours, outpacing Google’s Nano Banana 2 by a record margin. Free‑tier ChatGPT users receive a basic version, while paid subscribers unlock the full "thinking" mode that includes web‑search and batch generation. Pricing starts at $0.04 per image, with higher‑resolution outputs up to 2K. Read more

Google unveiled Workspace Intelligence, a new feature that integrates its Gemini AI model across the Google Workspace suite. The tool can pull data from Gmail, Docs, Sheets, Slides and Drive, using personal context to draft emails, edit documents, generate presentations and summarize conversations without users having to supply extra details. By tapping into stored emails, chats and files, Gemini promises to streamline routine tasks and keep output consistent with company style guides. The rollout follows Google’s earlier Personal Intelligence rollout, which let Gemini access Gmail and Photos for personal use. Read more

OpenAI announced the debut of workspace agents, a new class of shared AI assistants that can plan, execute, and monitor multi‑step tasks for entire teams. Running in the cloud, the agents can access files, run code, and integrate with tools like Slack, freeing users from repetitive work and keeping projects moving even when they step away. The feature marks a shift from ChatGPT as a question‑answering bot toward a continuous work partner, though it still requires human oversight and proper permissions. Read more

Google is weaving its Gemini AI assistant directly into Google Chat, turning the messaging platform into a one‑stop hub for routine work. Users can type natural‑language commands to draft Google Docs, build slide decks, pull files and book meetings without leaving the chat thread. The rollout also brings a daily briefing that surfaces urgent items and integrates third‑party tools such as Asana, Jira and Salesforce. By merging communication with execution, Google aims to cut the back‑and‑forth between apps and sharpen the productivity edge of Workspace. Read more