Latest updates from the world of artificial intelligence
Meta’s new terminal coding agent splits a task across parallel sub-agents in isolated worktrees, and pitches itself against Claude Code and Codex on cost rather than capability.
TechCrunch
Read more →Microsoft’s agentic security system enters public preview inside Defender today, with red and blue agents live — and the green agents that actually remediate held back behind human approval. The constraint, not the capability, is the story.
Microsoft Security
Read more →QM is MIT-licensed, self-hosted, and driven by Pi, OpenCode, Codex or Claude Code interchangeably. What separates it from the crowded harness field is that it was built for a company rather than a developer — scoped memory, per-person sandboxes, an org-level security posture. And going multi-tenant forced YC to publish a threat model that says plainly what its defaults cannot do, including that classifier approval “cannot guarantee prompt-injection resistance.”
GitHub — yc-software/qm
Read more →Three additions change what Anthropic's coding agent produces and where it runs: sessions can be published as live, shareable pages that call MCP connectors with each viewer's own credentials; Claude Code can build an iOS app and verify it in the simulator without leaving the session; and the desktop app is now in beta for Ubuntu and Debian.
Anthropic
Read more →Spotlighted during OpenAI's global Build Week, ChatGPT Work is a GPT-5.6-powered agent that gathers context from your connected apps and files and produces finished documents, spreadsheets, presentations, and dashboards — working in the background across web, phone, and desktop, with Codex now built into the ChatGPT desktop app.
OpenAI
Read more →
Anthropic is giving verified US K-12 teachers a full year of free premium Claude, plus a teaching-skills library and curricula mapped to all 50 states' standards. It's built with the AFT, Teach For America, and the Gates Foundation — and it keeps student data out of training.
Anthropic
Read more →
GPT-Live replaces ChatGPT's turn-based voice with a full-duplex model that listens and speaks at once — it says "mhmm," handles interruptions, stays quiet when you think, and quietly hands hard questions to a frontier model in the background.
OpenAI
Read more →Anthropic launched Claude Science on June 30 — a beta research workbench that unifies literature review, HPC compute, and reproducible, publication-ready figures. It ships with 60+ scientific skills, connectors to databases like PDB and UniProt, an NVIDIA BioNeMo integration, and a reviewer agent that checks its own citations.
Anthropic
Read more →Mark Zuckerberg has ordered Meta to build a standalone prediction-market app, codenamed “Arena,” where users bet play money on real-world events — with Llama generating the questions, recommending markets, and resolving outcomes in near real-time.
NPR
Read more →Anthropic launched Claude Tag on June 23 — tag @Claude in a Slack channel and it breaks the task into stages, works asynchronously with a shared team identity, and posts results back in the thread. Running on Opus 4.8 with an optional proactive "ambient" mode, it ships as a research preview for Enterprise and Team customers. Anthropic says 65% of its own product team's code now ships through it.
TechCrunch
Read more →A practical, step-by-step guide to Robinhood Agentic Trading: connect Claude, Claude Code, or ChatGPT to the Robinhood Trading MCP, fund a ring-fenced agentic wallet, set approval guardrails, and let your AI agent analyze and trade equities — safely.
Robinhood
Read more →Tencent has started a limited test of Xiaowei, an AI assistant woven directly into WeChat — running on its WeLM model with DeepSeek for some queries, and aiming to turn the 1.4-billion-user super-app into a voice-and-text concierge.
The Next Web
Read more →
Google has opened $99.99 pre-orders for the Gemini-powered Google Home Speaker — its first new smart speaker since 2020 — promising natural multi-step conversations, 360-degree sound, and a June 25 ship date.
A practical, illustrated walkthrough of Claude Code routines — saved prompts that run themselves in Anthropic’s cloud on a schedule, an API call, or a GitHub event. How to create one with /schedule or the web UI, write a prompt that behaves, pick a cadence, and worked examples you can copy.
Anthropic
Read more →
As of June 18, Gemini CLI stops serving consumer requests. Google steers developers to the Go-built, multi-agent Antigravity CLI — keeping Skills and Hooks, but dropping open source.
Google Developers Blog
Read more →At State of Unreal on June 17, Tim Sweeney revealed Unreal Engine 6 — a unified engine folding UE5 and the Unreal Editor for Fortnite into one, built around the Verse language, open USD/glTF standards, and AI-assisted creation via MCP with Claude, Gemini and others. Early Access is targeted for late 2027.
Epic Games
Read more →Rolling out June 17, the June 2026 Pixel Drop adds Gemini Omni for AI video, Lyria for AI music, and a Pixel-first Screen Reactions tool — and ships stable Android 17 with Bubbles floating-window multitasking and Voice Translate on the Pixel 10a.
Claude Code’s desktop app now packs five permission modes into one Ctrl+Shift+M menu — Ask permissions, Accept edits, Plan, Auto, and the newly added Bypass permissions. Here’s what each one does and when to reach for it.
Claude Code Docs
Read more →Oracle customers will soon be able to apply their existing Universal Credits toward OpenAI models and the Codex coding agent through OCI — folding frontier AI into the cloud commitments enterprises have already signed.
OpenAI
Read more →
At its Payments Forum in San Francisco, Visa and OpenAI agreed to embed Visa's rails into ChatGPT — so AI agents can discover products, check out, and pay across any Visa-accepting merchant, with spending caps and approvals built in.
SiliconANGLE
Read more →Google dropped its mid-tier Gemini subscription from $7.99 to $4.99 a month and doubled the storage to 400GB. The discounting that started in India has now reached the US — and analysts are calling it the start of AI's commoditization era.
TechCrunch
Read more →Google's new Gemini 3.5 Live Translate streams near real-time speech-to-speech translation in 70+ languages while preserving the speaker's tone, pace, and pitch — rolling out across the Translate app, Google Meet, AI Studio, and the Gemini Live API.
At Tim Cook's final WWDC keynote as CEO, Apple confirmed a rebuilt Siri running on a custom 1.2-trillion-parameter Gemini model — and a new Extensions system that lets users pick Gemini, ChatGPT, or Anthropic's Claude to power Apple Intelligence.
BuildFastWithAI
Read more →Agent Skills are folders of instructions, scripts, and resources that Claude loads on demand to specialize at a task. Here’s how they work, how to build one, and why they’ve become a cross-vendor open standard.
Anthropic
Read more →xAI's Connectors plug Grok into the apps people already work in — GitHub, Microsoft 365, Google Workspace, Notion, Linear — plus any MCP server. The sharper move is Grok Skills: instead of building a rival format, xAI reads Anthropic's Claude Code conventions directly, loading CLAUDE.md and .claude skill packs with zero configuration.
xAI
Read more →OpenAI is recasting Codex as an agent for every knowledge worker, not just engineers. A June 2 update adds six role-specific plugins bundling 62 apps and 110 skills, a prompt-to-app builder called Codex Sites, and Annotations that now reach documents, spreadsheets and slides. Codex is at 5 million weekly users, with non-developers its fastest-growing cohort.
OpenAI / VentureBeat
Read more →OpenAI's new Lockdown Mode strips ChatGPT down to a more defensive core — switching off live web browsing, agent mode, deep research, web images and file downloads — to make it harder for hidden prompt-injection attacks to exfiltrate sensitive data. It's rolling out beyond enterprise to Business and personal accounts, but OpenAI is candid that it doesn't stop injections outright.
OpenAI / TechCrunch
Read more →
An internal Meta memo lays out an aggressive wearables roadmap: an always-listening AI pendant built on its Limitless acquisition, four new smart-glasses models in 2026, and an enterprise 'Wearables for Work' service — all aimed at selling 10 million devices in the back half of the year.
Engadget
Read more →In a busy few days, xAI previewed Grok Imagine 1.5, a single-image-to-video model that debuted at No. 1 on the Artificial Analysis arena with audio baked into every clip — and named Grok the default engine behind the voices on Vapi’s 2.5-million-agent platform.
xAI
Read more →OpenAI is rolling out “Dreaming,” a rebuilt ChatGPT memory that synthesizes context from past chats automatically and keeps it current — no more telling the assistant what to remember. It starts with US Plus and Pro users after a roughly 5x cut in compute cost.
OpenAI
Read more →Seen the violet “ultracode” in Claude Code and wondered what “Ultra Code Mode” is? Here is the focused how-it-works guide: what the setting does, how to turn it on with /effort ultracode or the ultracode keyword, what happens under the hood, and when not to use it.
Claude Code Docs
Read more →
Build 2026’s headline was MAI-Thinking-1 — Microsoft’s first from-scratch reasoning model, claiming Opus 4.6-level coding at 35B active parameters — alongside MAI-Code, Image, Voice and Transcribe, the Microsoft IQ context layer, a GitHub Copilot desktop app, and an agent-governance stack.
Microsoft
Read more →Jensen Huang’s GTC Taipei keynote put Vera Rubin into full production with HBM4 from Samsung, SK Hynix and Micron, unveiled the agent-first Vera CPU (1.8x vs x86), RTX Spark AI PCs landing this fall, and Cosmos 3 for physical AI.
CRN Asia
Read more →A BitsMinds analysis. The frontier fight has moved from models to the agents wrapped around them. We line up the three coding agents developers actually argue about in 2026 — Anthropic’s terminal-native Claude Code, OpenAI’s everywhere-at-once Codex, and Google’s agent-first Antigravity IDE — across form factor, autonomy, verification, model choice and price. The short version: Claude Code owns deep terminal autonomy, Codex wins ubiquity, and Antigravity is the open, free, multi-agent cockpit. Here is the full scorecard.
BitsMinds Analysis
Read more →
Starting June 1, every GitHub Copilot plan moves from fixed premium request quotas to metered GitHub AI Credits — each worth $0.01 and drawn down by token usage. Code completions stay unlimited, monthly subscribers migrate automatically, and annual subscribers stay on the old model until renewal. GitHub says the change is a response to agentic coding making the flat-rate model "no longer sustainable."
The GitHub Blog
Read more →Bloomberg’s Mark Gurman has published renderings and details of iOS 27 ahead of Apple’s June 8 WWDC, centered on a long-promised Siri overhaul. The rebuilt assistant is said to run on a custom Google Gemini model — reported at roughly 1.2 trillion parameters — hosted on Apple’s Private Cloud Compute, living inside the Dynamic Island and gaining a standalone, ChatGPT-style app with chat history, voice and document upload.
Bloomberg
Read more →Meta is putting a price tag on AI for the first time. As it rolls out paid tiers across Instagram, Facebook and WhatsApp, the company is folding everything into a new “Meta One” subscription brand — including two AI plans, Meta One Plus at $7.99 a month and Meta One Premium at $19.99 a month, with deeper reasoning and more image and video generation. The AI tiers begin testing next month in Singapore, Guatemala and Bolivia; the free Meta AI stays free.
CNBC / TechCrunch / Bloomberg
Read more →On May 27, Robinhood opened its platform to autonomous AI agents via Model Context Protocol, letting Claude, ChatGPT and other LLMs read portfolios, analyze risk and execute equity trades out of a sandboxed wallet. Options, crypto, futures and prediction markets are next; an agentic Gold credit card ships alongside.
TechCrunch / CNBC
Read more →Luxury phone-maker Vertu unveiled the Alphafold on May 28 — a foldable Android starting at $6,880 (and topping out near $46,800) whose headline feature is Hermes Agent, an enterprise AI that connects to ERP, CRM and 80+ apps and routes prompts across Claude, GPT, Gemini and open-source models. The first 115-unit batch ships this week.
TechCrunch
Read more →Microsoft has made computer-using agents in Copilot Studio generally available across every commercial region, becoming the first major hyperscaler to put production-grade "computer use" into enterprises' hands. The agents operate desktop and web apps directly through the interface — clicking, typing and navigating like a person — using OpenAI's CUA and Anthropic's Claude Sonnet 4.5, wrapped in Azure Key Vault, Microsoft Purview auditing and human-in-the-loop review.
Microsoft
Read more →In its May 21 release, OpenAI gave Codex permission to keep clicking, typing and navigating macOS apps even after the screen is off. The new "Locked Use" mode — triggered remotely from the ChatGPT mobile app — turns a Mac sitting on your desk into a 24/7 agent runtime, with short-lived authorization and a hard relock the moment you touch the keyboard.
OpenAI Developers / MacRumors
Read more →
A new genai.apple.com subdomain spotted on May 23 — sitting dormant but reserved — is the clearest signal yet that WWDC 2026 will be the keynote where Apple finally ships the Siri overhaul it promised in 2024, reportedly powered by a custom 1.2-trillion-parameter Gemini model from Google.
9to5Mac / MacRumors
Read more →
At Google Marketing Live 2026, Google unveiled Ask Advisor — one Gemini agent that spans Ads, Analytics, Merchant Center, and the Marketing Platform — alongside conversational ad formats built for AI Mode search.
Cursor shipped its third-generation in-house coding agent at /usr/bin/bash.50/.50 per million tokens — about a tenth of Opus 4.7 — and benchmarks put it shoulder-to-shoulder with Anthropic's flagship.
TechTimes
Read more →
At I/O 2026 Google unveiled Intelligent Eyewear — Gemini-powered audio and display glasses built on Android XR, co-designed with Samsung, Warby Parker and Gentle Monster, with the audio frames shipping first this fall.
Google Blog
Read more →
Pics, unveiled at Google I/O 2026, lets Workspace users generate and edit marketing graphics, mockups and social posts from text prompts — with every element re-editable like a comment in Google Docs.
TechCrunch
Read more →
At I/O 2026, Google replaced its iconic blue-link search with a Gemini-powered intelligent box, generative UI, and background 'information agents' that monitor the web for you.
TechCrunch
Read more →
Pichai opened I/O with Gemini 3.5 Flash shipping immediately, Gemini Omni for video, a 24/7 Spark agent in beta next week, Android XR audio glasses for fall, and a $5B Blackstone TPU venture.
9to5Google
Read more →
Spark runs in the cloud while you sleep, plans events, drives spreadsheets, and chases follow-ups — a clear shot across the bow of ChatGPT Pulse and Claude Cowork.
Analytics Insight
Read more →
Anthropic's new Cowork plugin for marketers ships /campaign-plan, /brand-review, /seo-audit, and email-sequence commands with connectors to Canva, HubSpot, Amplitude, and more.
Anthropic
Read more →
12 role-specific plugins, a cross-app Microsoft 365 agent, and live deployments at Freshfields and Quinn Emanuel — Anthropic is making legal its biggest Cowork beachhead.
Anthropic
Read more →Built by ex-Tesla and Netflix engineers, Osaurus routes between local Llama, DeepSeek V4, and Gemma 4 and cloud GPT-5.5, Claude, and Gemini in one sandboxed macOS app — with 112K downloads so far.
TechCrunch
Read more →Grok Build pairs a new Grok 4.3 model with a 2 million token context window and the ability to spawn up to 8 parallel subagents — the most aggressive answer yet to Anthropic's lead in agentic coding tools.
Bloomberg
Read more →
OpenAI is letting developers monitor and steer Codex coding sessions from iPhone, iPad, and Android, turning the phone into a remote control for an agent that runs back on a desktop or hosted runtime.
TechCrunch
Read more →
ChatGPT Pro users in the U.S. can now wire up Plaid and 12,000+ institutions, surfacing portfolios, spending, and long-range planning inside the chat.
TechCrunch
Read more →
Claude for Small Business ships 15 ready-to-run agentic workflows for payroll, invoice chasing, and lead triage, plus a 10-city free training tour kicking off in Chicago.
TechCrunch
Read more →
Higgsfield's new agent orchestrates Claude Opus 4.7, GPT-5.5 Pro, Gemini 3.1 Pro, and video models like Kling 3.0 through a Hermes brain with three layers of memory and 40+ built-in tools.
Higgsfield
Read more →
Amazon is sunsetting its standalone Rufus chatbot and rolling its product expertise into Alexa for Shopping, a personalized agent that tracks prices, schedules reorders, and even buys items off Amazon through Shop Direct.
Amazon
Read more →Sam Altman pitches enterprises switching from Claude with 60 days of complimentary Codex access. Hours later, Anthropic boosts Claude Code weekly limits by 50% through July 13 — for free.
Business Insider / OpenAI / Anthropic
Read more →Self-serve campaigns now start at a $50,000 minimum (down from $250,000), with conversion measurement and CPC bidding aimed at making ChatGPT a real Google Ads competitor.
OpenAI
Read more →
Google introduced Gemini Intelligence, a sweep of agentic Android features — cross-app task automation, a Chrome auto-browse tool, a Rambler voice-to-text rewriter and natural-language Create My Widget — debuting on the Samsung Galaxy S26 and Pixel 10 this summer ahead of Apple's AI reboot.
AWS is the first cloud provider to host the full Claude Platform — APIs, console, Managed Agents, Skills and all — natively inside a customer's AWS account, with consolidated billing, IAM access controls and same-day availability for new Anthropic releases across 17 regions.
AWS
Read more →
Palo Alto-based Akool rebuilt its full stack for a 10-20x speedup, pushing generative video clips to 1-3 seconds and live streaming below 30ms per frame.
PR Newswire
Read more →
A new grok-imagine-image-quality endpoint targets product visualization, brand-safe ad creative and consistent character workflows, with image outputs designed to feed directly into Grok video generation.
xAI
Read more →
A bespoke build of GPT-5.5 with relaxed restrictions on offensive-security tasks is rolling out in limited preview to thousands of verified cybersecurity professionals, OpenAI's answer to Anthropic's Mythos cyber preview from a month earlier.
Help Net Security
Read more →
A privacy researcher revealed that Chrome has been quietly downloading a 4GB Gemini Nano model to billions of devices without explicit consent, reigniting debate over default-on AI features at browser scale.
9to5Google
Read more →
Google shut down Project Mariner on May 4 without a public announcement, ending its 17-month run as a standalone browser-controlling AI and folding the technology into the Gemini API and the new Gemini Agent.
Android Authority
Read more →
Anthropic launches a 10-agent suite for banks and insurers, full Microsoft 365 integration, a Moody's data partnership, and Claude Opus 4.7 — with JPMorgan, Goldman, Citi, and Visa as launch customers.
Fortune
Read more →OpenAI replaced GPT-5.3 Instant with GPT-5.5 Instant as ChatGPT's default model, delivering a 16-point jump on AIME 2025 math, sharper multimodal reasoning, and lower hallucination rates in legal, medical, and financial queries.
TechCrunch
Read more →
OpenAI has launched ChatGPT for Clinicians, a free GPT-5.4-powered tool for verified U.S. physicians, nurse practitioners, PAs, and pharmacists, with safety scores that beat human doctors on a new HealthBench Professional benchmark.
OpenAI
Read more →Amazon Web Services puts OpenAI's frontier models, the Codex coding agent, and a new managed-agents service inside Bedrock — and ships an Amazon Quick desktop app for AI work outside the browser.
AWS Blog
Read more →
A new Cloudflare-Stripe protocol lets AI coding agents provision accounts, register domains, and ship apps to production autonomously, with a $100-per-month default spending cap per provider.
Cloudflare Blog
Read more →
Microsoft's control plane for AI agents went GA on May 1 at $15 per user per month, adding multicloud discovery, shadow-AI detection on Windows endpoints, and runtime blocking of malicious agent behavior.
Microsoft Security Blog
Read more →
Microsoft launches a Word-native AI agent purpose-built for legal work — playbook-driven contract review, deterministic redlining, and tracked changes that preserve document structure rather than rewriting it.
Microsoft / Thurrott
Read more →Powered by Opus 4.7 and the Mythos-fueled Project Glasswing, Claude Security reasons over entire codebases like a human researcher — tracing data flows instead of matching patterns.
SiliconANGLE
Read more →
Claude can now drive a coalition of creative apps — Adobe Creative Cloud, Blender, Ableton Live, Autodesk Fusion, SketchUp, Affinity, Resolume, and Splice — through MCP-based connectors.
9to5Mac
Read more →
A new Create button inside the Gemini tab on Google TV lets viewers swap outfits in photos, animate stills, and generate clips by voice — turning the television into the first big-screen home for generative AI.
TechCrunch
Read more →Google opened a public beta of Ask YouTube on April 28, turning the world's largest video platform into a conversational answer engine that returns clips, segments, and step-by-step explanations instead of a list of links.
TechCrunch
Read more →
The Sora consumer app went dark on April 26 and the API will follow in September, ending a 17-month experiment that reportedly cost OpenAI roughly $1 million per day to keep running.
OpenAI Help Center
Read more →
Amazon rolled out a conversational audio shopping assistant on April 28 that answers product questions in real time, deepening its push to embed AI agents directly into the buying flow.
TechCrunch
Read more →Founded by ex-Google and Meta chip veterans, Majestic Labs has revealed a server purpose-built to break AI's memory wall — packing up to 128TB of unified memory to run multi-trillion-parameter models in a single box.
BusinessWire
Read more →
Sundar Pichai revealed at Cloud Next 2026 that three-quarters of Google's new code is now AI-generated, up from 30% a year ago — with engineers shifting from authors to reviewers.
Google Blog / Cloud Next 2026
Read more →
Anthropic has unveiled Claude Design, a new research preview product that lets founders, product managers, and anyone without design expertise generate prototypes, slides, and polished one-pagers from a text description — powered by Claude Opus 4.7.
TechCrunch
Read more →
NVIDIA's new open reference architecture automates training data generation for physical AI systems, dramatically cutting the cost and time to train robots, vision agents, and self-driving vehicles.
NVIDIA Newsroom
Read more →NousResearch shipped Hermes Agent v0.8, an autonomous server-side agent with a built-in learning loop that builds persistent memory and self-generated skills over time.
NousResearch / AIToolly
Read more →
Anthropic debuts Claude Managed Agents, a fully managed cloud service that lets developers build and deploy autonomous AI agents with secure sandboxing, persistent sessions, and built-in tooling — no infrastructure required.
SiliconANGLE
Read more →Rather than a public release, Anthropic is deploying Claude Mythos Preview exclusively to 40+ vetted tech companies — including Apple, Google, and Microsoft — to bolster critical software security before attackers can exploit the model's power.
The Hacker News
Read more →
Google has released LiteRT-LM, a production-grade open-source framework enabling developers to run large language models directly on smartphones and IoT devices without cloud dependency.
AIToolly
Read more →
Suno V4 delivers a dramatic quality leap — full songs up to 4 minutes long at studio quality.
Suno Blog
Read more →
Midjourney V7 delivers unprecedented photorealistic quality and more precise control over image generation.
Midjourney
Read more →