OpenAI released GPT-5.5 today, codenamed “Spud.” It’s faster, more independent, and better at coding and research than its predecessor. Here’s what the new model actually does and why it matters for everyday users.
SpaceX just struck a deal to potentially buy Cursor, the popular AI code editor, for $60 billion. Here’s what Cursor does, why a rocket company wants it, and what it means for AI-assisted coding.
OpenAI introduces workspace agents in ChatGPT — shared AI assistants that teams can build once and use together. They connect to tools like Slack and Linear, run in the background, and handle multi-step tasks without constant supervision. Available now for Business, Enterprise, and Edu plans.
OpenAI released ChatGPT Images 2.0 with dramatically improved text rendering, 2K resolution, flexible aspect ratios, and a new Thinking mode that reasons before it draws. Here’s what changed and why it matters.
Vercel, the hosting platform behind many popular AI apps and websites, confirmed a security breach on April 19. Attackers got in through a compromised AI tool called Context.ai, accessed employee credentials, and reached internal systems. Here’s what happened and why it matters.
Google AI Studio received its biggest overhaul yet: a unified Playground, Build Mode with the Antigravity coding agent, new Gemma 4 models, Gemini 3.1 Flash TTS, and Veo 3.1 Lite video generation. Here’s what changed and what it means for beginners.
Google is reportedly in talks with Marvell Technology to develop two new AI chips, sending MRVL shares to new highs. Here’s what it means for the custom silicon race and why it matters for the future of AI computing.
One in three American adults now asks AI chatbots for health advice. A growing stack of research says those answers are wrong about half the time.
Anthropic’s new Claude Design tool turns text prompts into prototypes and slide decks. Figma’s stock dropped 7% within hours — but the real story is an 80% decline since last August.
Anthropic’s Opus 4.7 shipped with a reworked effort system, adaptive thinking, and a new default configuration inside Claude Code. Here’s how to actually use all of it.
Anthropic’s Mythos model finds real vulnerabilities, but the “too dangerous to release” framing is doing more marketing work than safety work.
A Verisk study finds 99% of insurers have encountered AI-altered claim documents, while 55% of Gen Z consumers say they’d consider faking one.
OpenAI ships a massive Codex update with computer use, plugins, automations, and a built-in browser — but the real story is the superapp strategy behind it.
Anthropic ships Opus 4.7 with serious coding and vision gains, and openly admits its more powerful Mythos model is too dangerous to release broadly.
Canva’s AI 2.0 replaces the template-first workflow with agentic orchestration, a proprietary design model, and persistent memory — its biggest architectural overhaul since 2013.
Jack Dorsey and Sequoia’s Roelof Botha argue middle management exists to route information — and Block just cut 40% of its workforce betting AI can do that job better.
Anthropic shipped eight new ways to work with Claude in six weeks — from free memory to Claude Code auto mode — and the pattern is a land grab, not a roadmap.
Lovable’s AI app builder now generates working payment flows directly from chat, with Paddle, Stripe, or Shopify handling the backend and tax compliance built in.
Google rolled out Skills in Chrome on April 14, letting Gemini users save working prompts and trigger them with a single click — a feature Anthropic’s Claude in Chrome already offered, minus the paywall.
Google’s new Gemini 3.1 Flash TTS undercuts ElevenLabs on price while shipping 200+ audio tags, native multi-speaker dialogue, and support for over 70 languages.
Anthropic is expected to ship Claude Opus 4.7 within days as OpenAI’s codename-Spud model finishes pretraining — two rival labs colliding on release week.
Anthropic’s April 14 redesign of Claude Code’s desktop app puts multi-session parallel workflows at the center — and introduces Routines, a research-preview feature that runs automated tasks on cloud infrastructure even when your Mac is offline.
The EU rejected Meta’s per-query fee workaround for WhatsApp AI access, calling it functionally identical to the ban it replaced — and is now pushing toward a forced restoration of rival AI assistants.
Adobe’s Firefly AI Assistant promises to orchestrate workflows across Photoshop, Premiere, Lightroom, and more from a single chat interface — a bet on unifying Creative Cloud rather than replacing it.
OpenAI’s Agents SDK update separates the control plane from the execution layer — a deliberate architectural choice that answers the enterprise question of where data actually goes when an AI agent touches it.
A federal judge ruled that a fraud defendant’s AI chatbot conversations don’t qualify for attorney-client privilege — and lawyers across the US are now warning clients to treat their AI chats as potentially discoverable evidence.
Google’s native Swift Gemini app arrives on Mac today with screen sharing and keyboard shortcuts — a strategic move ahead of Apple’s expected Gemini-powered Siri redesign at WWDC.
