Tech Roundup: Google Gemini Intelligence, Meta Threads Refresh & More

[A recurring feature on the latest in Science & Technology.]

• OpenAI introduces AI-generated pets to the Codex app, its agentic coding tool, as "optional animated companions" that serve as a floating overlay to tell users what Codex is working on and notify when Codex completes a task or whether it needs their input.
• WhatsApp tests an independent cloud backup system that gives users more direct control over their chat histories, by allowing the backups to be stored on its own servers, instead of third-party cloud services like Google Drive and Apple iCloud.
• Amazon opens up its global logistics network to all businesses; debuts Supply Chain Services, which lets companies use its logistics network to move, store and deliver everything from raw materials to final products, pitting it directly against UPS and FedEx.
• Meta's Instagram tests optional 'AI creator' labels, as it encourages accounts that frequently post Gen AI content to use the feature; brings DMs to Threads on desktop web and tests a Grok-like feature that allows users to mention @meta.ai in Threads posts and replies to seek additional context. (The new Meta AI account cannot be blocked, however.)
• Apple officially brings end-to-end encryption for RCS messages between Apple and Android devices with iOS 26.5.
• Pinterest reports monthly active users up 11% YoY to 631 million.
• X CEO Elon Musk agrees to pay $1.5 million to settle SEC allegations that he cheated Twitter shareholders in 2022 by failing to disclose the 5% stake he had in the company.
• Google, Microsoft and xAI join OpenAI and Anthropic to give the U.S. government early access to their artificial intelligence models for evaluating them prior to public release.
• Meta says it will expand Instagram teen account safeguards to 27 E.U. countries, and plans to roll them out on Facebook in the U.S., followed by the U.K. and E.U. in June 2026; comes as the company says it's is using AI on Facebook and Instagram to detect under-13 users by analysing bone structure, height and visual cues, but emphasises it's "not facial recognition."
• Meta and CEO Mark Zuckerberg get sued in a fresh lawsuit that accuses the tech company of illegally copying millions of books, articles and other works to train its AI systems; says "Zuckerberg himself personally authorized and actively encouraged the infringement."
• Google updates Home with upgraded Gemini voice assistant and new camera controls, along with a refreshed user interface that allows for faster navigation and smoother scrubbing through videos.
• OpenAI rolls out GPT-5.5 Instant to all ChatGPT users as the new default model, replacing GPT-5.3; says it has "significant improvements in factuality across the board," reduces hallucinations in sensitive areas like law, medicine, and finance and better at personalisation.
• Etsy launches its app within ChatGPT, offering a new way for shoppers to explore its catalogue of over 100 million listings.
• Bose integrates Amazon Alexa+ into its speakers, including Lifestyle Ultra Speaker and Lifestyle Ultra Soundbar, marking the first time the upgraded smart assistant has been bundled into a non-Amazon-made device.
• Google updates Chrome for Android to allow users to share their approximate location with websites, instead of giving them access to their precise location.
• Samsung adds short-form video from media publishers to Samsung News; available for Galaxy phones and tablets released in 2023 and later and running Android 13 or higher.
• Apple blocks AI coding apps on the App Store, sparking complaints among AI companies that the company's App Store rules are outdated and draconian.
• Pennsylvania sues Character.AI, alleging that one of its characters posed as a psychiatrist, as the U.S. state seeks to prevent chatbots from impersonating doctors.
• Canadian adult entertainment company Aylo says it will restore Pornhub access in the U.K., but only for users who have verified their age via iOS 26.4's device-level age verification system. (The development comes as a new Internet Matters survey suggests the U.K.'s Online Safety Act age checks are easy for many children to bypass using workarounds include fake birthdays, borrowed IDs, video game characters and even drawing on a fake moustache.)
• Anthropic unveils 10 new AI agents for the financial sector, including for drafting pitch decks, reviewing financial statements and escalating compliance cases.
• The U.S. Federal Trade Commission (FTC) plans to ban data broker Kochava and its subsidiary CDS from selling Americans' location data as part of a settlement over charges brought in 2022.
• Apple reaches a $250 million settlement in a U.S. federal court to resolve a false advertising class action lawsuit over the launch of a "personalised" Siri in 2024.
• Microsoft Xbox says it "will begin winding down Copilot on mobile and will stop development of Copilot on console" as part of a strategic shift.
• Google quietly shut down Project Mariner, its Chrome-browsing AI agent for completing tasks on users' behalf, after highlighting it onstage at I/O 2025.
• A study of Google Chrome's browser extension architecture, known as Manifest v3 (MV3), has found "no statistically significant reduction in ad-blocking or anti-tracking effectiveness for MV3 ad blockers compared to their MV2 counterparts, and in some cases, MV3 instances even exhibit slight improvements in blocking trackers," contradicting earlier concerns that MV3 was designed to render ad blockers ineffective.
• Google updates its AI Search features to make it easier for users to find information from sources they know and trust; introduces "a preview of perspectives" from firsthand sources like social media, Reddit and other web forums.
• Perplexity launches Perplexity Computer for Professional Finance, allowing finance teams to bring licensed data from providers like Morningstar, PitchBook, Daloopa, and Carbon Arc into Computer.
• Anthropic details "model spec midtraining", which adds a stage between pretraining and fine-tuning to improve generalisation from alignment training; updates Claude Managed Agents with "dreaming," a scheduled process that reviews recent work and updates memory.
• Snap ends its $400 million deal with Perplexity, with the latter saying "the original implementation was not the right fit for each company's product goals and have resolved the matter amicably on confidential terms."
• Anthropic signs a deal with SpaceX to use "all of the compute capacity" at Colossus 1, giving it access to over 300MW of new capacity within the month amid surging demand; doubles Claude Code's five-hour rate limits for paid plans and removes peak hours limit reduction for Pro and Max plans. (CEO Elon Musk says SpaceX reserves "the right to reclaim the compute" from Anthropic if its "AI engages in actions that harm humanity.")
• Google releases Multi-Token Prediction drafters for its Gemma 4 models, which use a form of speculative decoding to guess future tokens for faster inference without any degradation in output quality or reasoning logic; DeepMind partners with EVE Online to study intelligence in complex, dynamic, player-driven systems.
• Samsung halts all home appliance sales in mainland China amid fierce competition from local brands, as pivot to AI accelerates.
• OpenAI takes steps to improve "openness, access, retention, and children's privacy" after Canadian privacy regulators identify several concerns in the way the company developed and deployed ChatGPT, including "overcollection of personal information; lack of valid consent and transparency; factual inaccuracies involving personal information; issues related to individuals' ability to access, correct and delete their personal information; and a lack of accountability for the personal information under OpenAI's control."
• TikTok pulls back on testing an AI feature that added text summaries to videos after finding that it's prone to hallucinations; to be limit its functionality to identifying products in videos.
• Google rebrands the Fitbit app to Google Health as it turns it into a unified portal for health and fitness data; to come integrated with the Gemini-powered Google Health Coach if users pay for Google Health Premium (previously Fitbit Premium) subscription; debuts a new fitness band called Fitbit Air. (As part of this consolidation, Google plans to shut down Fit later this year, at which time users will have to migrate their data to Google Health.)
• Spotify lets AI agents like OpenClaw and Claude Code generate audio summaries and personal podcasts using a command-line tool called Save to Spotify; expands its interactive AI DJ feature beyond English and Spanish to support French, German, Italian and Brazilian Portuguese, and celebrates its 20th anniversary with a Wrapped-like experience that features "never-before-shared data" going back to when users first joined Spotify and their entire music listening history.
• AWS unveils Amazon Bedrock AgentCore Payments and partners with Coinbase and Stripe to enable AI agents to execute transactions using stablecoins.
• OpenAI debuts a Codex plugin for Google Chrome to allow users to test web apps, collect context from across open tabs and use Chrome DevTools in parallel while the user performs other tasks; launches GPT‑Realtime‑2, GPT‑Realtime‑Translate and GPT-Realtime-Whisper models to create realistic vocal simulation, provide real-time translation and transcription services.
• Online dating platform Bumble plans to replace swipes to something the company says is "revolutionary for the category" (not exactly clear, is it?); to start rolling out in the fourth quarter of this year.
• OpenAI introduces an optional Trusted Contact safeguard for to alert a trusted third party if mentions of self-harm are expressed within a conversation; comes as the company faces a wave of lawsuits from the families of people who have committed suicide after talking with its chatbot.
• OpenAI begins rolling out GPT-5.5-Cyber, a security-focused variant of the model, in a limited preview capacity to vetted cybersecurity teams.
• Anthropic researchers detail "natural language autoencoders," which convert LLM activations, the numbers encoding a model's thoughts, into natural language text.
• E.U. legislators reach a deal to postpone restrictions on high-risk AI until December 2027 and to exempt the use of AI in industrial applications from the AI Act.
• Google plans to discontinue the ability to back up media to Google Photos using the Google Drive desktop app effective June 15, 2026; recommends users to set up the Back up folders functionality directly via the Google Photos web app.
• xAI becomes SpaceXAI following SpaceX's acquisition of the company; releases Grok Voice mode for Apple CarPlay, allowing CarPlay users to ask the chatbot questions and make requests directly from their vehicle dashboard, handsfree.
• Google updates Help me write for Gmail with personalisation enhancements that connects Google Drive and Gmail to insert relevant information into the email draft and create personalised email drafts that match the tone and style of previously written emails.
• Apple and Meta oppose a Canadian bill, named Bill C-22, that the companies say could force them to create backdoor access to encrypted user data.
• The E.U. warns that VPNs are being used to bypass online age-verification systems, calling their use "a loophole in the legislation that needs closing"; comes as Utah enforces law requiring adult websites to implement strict age checks on anyone physically located in the U.S. state, even if they use a VPN to mask their location.
• Amazon follows Netflix and Disney by adding a TikTok-like Clips feed to Prime Video.
• Google releases an updated version of Snapseed for Android with a new pro camera, nearly a year after pushing out a similar update for iOS.
• The Indian government has blocked 652 mobile applications on account of concerns relating to data security and other malpractices, according to the Internet Freedom Foundation.
• Iran's 70+ day internet blackout becomes "the longest recorded national internet shutdown in a connected society," as businesses warn of layoffs and closures.
• The U.S. FCC passes an anti-robocall proposal requiring telecoms, including VoIP providers, to verify user identities before activating service, raising privacy fears.
• Google pitches Google Cloud Fraud Defense as an evolution of reCAPTCHA, aiming to offer bot, account and transaction protection across the agentic web; appears to require Google Play Services for passing next-gen reCAPTCHA on Android, effectively locking out de-Googled Android phones.
• TikTok plans to roll out TikTok Ad-Free, a £3.99-per-month subscription for U.K. accounts aged 18 or older in the coming months, after testing the option in 2023.
• Meta's WhatsApp being rolling out its paid WhatsApp Plus subscription to iOS for €2.49 per month in Europe and $29 in Mexico, following beta testing of the new personalisation-focused tier amongst a small group of users; comes with premium sticker packs, custom themes, 14 alternate app icons, increased pinned-chat limit from 3 to 20 and 10 new ringtones.
• Discord launches Nitro Rewards, giving Nitro subscribers access to offers from gaming services like Xbox Game Pass and hardware like Logitech G at no extra cost.
• PayPal-owned Venmo implements a major privacy measure, setting new users' posts to "friends only" by default during onboarding.
• OpenAI launches the OpenAI Deployment Company with a $4 billion investment to help organizations build and deploy AI systems.
• The European Commission says it is in ongoing discussions with OpenAI and Anthropic to access their latest AI models, as OpenAI proactively offers access; plans to take action against "addictive design" on TikTok and Instagram, including "endless scrolling."
• General Motors agrees to pay $12.75 million to settle a California data privacy lawsuit that accused the automaker of selling driver location and driver data to data brokers and insurance companies.
• Microsoft tests a new speed boost feature called Low Latency Profile in Windows 11 to improve app launch times by ramping up CPU frequency in short bursts.
• Thinking Machines Lab details interaction models, which can think and respond in real time, letting users and AI interact continuously for better collaboration.
• Digg relaunches as an aggregator of AI news and social media commentary, with plans to expand to other topics, after shutting down its previous reboot in March.
• Apple releases iOS 26.5 with official support for end-to-end encryption for RCS messaging in beta with supported carriers across Android and iOS.
• Microsoft expands interoperability by allowing Microsoft Teams users on Android devices to join third-party meetings via the Session Initiation Protocol (SIP).
• Apple acquires Patchflyer, the company behind Color.io, a web-based colour grading tool popular with photographers and filmmakers.
• Amazon launches Amazon Now, its 30-minute delivery service, in dozens of U.S. cities including Atlanta, Dallas-Fort Worth, Philadelphia, and Seattle, after pilots.
• The U.S. state of Texas sues Netflix for allegedly not obtaining user consent before collecting and sharing subscriber data with advertisers and data brokers like Experian and Acxiom following the debut of an ad-supported streaming plan in 2022; accuses it of building an "behaviour-surveillance program" and "intentional engineering to track and log users' viewing habits, preferences, devices, household networks, application usage, and other sensitive behavioural data."
• Meta updates its Meta AI app with live AI features, letting users point their camera to seek information in real-time.
• Google unveils Gemini Intelligence, bundling existing and new Gemini features, including task automation across apps and letting users vibe-code Android widgets.
• Sony’s new Xperia phone gets an overdue redesign, as Google previews Android 17 with Material 3 Expressive design and unveils a new series of Googlebook laptops with with deep Gemini Intelligence integration and run software built on a foundation that combines Android and ChromeOS. (Google says Chromebooks will get support through their "existing date commitment" and that "many" models are "eligible to transition" to the Googlebook experience.)
• Google expands Quick Share to Samsung, OPPO, OnePlus, Vivo, Xiaomi and HONOR devices later this year to allow Android users to send files to iPhones; says it worked with Apple to make it easier to switch from an iPhone to an Android device, and debuts new features like Create My Widget to create custom widgets using natural language prompts as part of what Google calls generative UI, 3D emoji, Gemini integration in Android Auto, Rambler (an AI voice dictation feature in Gboard that cuts out filler words like um, ah and like, and create a concise message), Screen Reactions (to let users film with the front and back cameras at the same time for social media reaction videos), Instagram integration (to improve the video and photo quality of content captured with Android devices and uploaded to Instagram), and Pause Point (a feature that forces a mandatory 10-second pause before opening any app a user has labeled as a distraction).
• Alphabet Waymo recalls about 3,800 robotaxis in the U.S. to fix software issues that may cause them to drive onto flooded roads.
• Google's DeepMind details a Gemini-powered mouse pointer that understands what it is pointing at, allowing users to perform tasks without using text-heavy prompts.
• Meta offers to give rival AI chatbots free access to WhatsApp for a month while it discusses commitments with E.U. antitrust regulators to address their concerns; updates Teen Accounts with new supervision tools for parents so they can "better understand their teen’s algorithm on Instagram and the topics that shape it."
• Google unveils a "full bleed" Android Auto design that fills unconventionally shaped screens like in the BMW Neue Klasse; plans to add YouTube video streaming.
• Anthropic announces 12 Claude plugins for the legal sector, including a "commercial counsel" tool for reviewing vendor agreements and a bar exam study tool.
• Meta faces new lawsuit over scam ads on Facebook and Instagram that accuses the company of profiting from a "vast ecosystem of scam ads" that have defrauded senior citizens and other vulnerable people; comes as OpenAI faces another wrongful-death lawsuit for allegedly designing ChatGPT to become an "illicit drug coach" after it told a 19-year-old to take a lethal mix of Kratom and Xanax.
• Meta debuts a refreshed logo and wordmark for Threads, nearly three year after its launch in July 2023; Instagram rolls out Instants, which lets users share ephemeral photos, as an in-app feature in Instagram and as a standalone app in select countries.
• X launches a History tab for bookmarks, likes, videos and articles, expanding the app's role as a save-it-for-later tool; SpaceXAI launches Grok Build, an agent and CLI for coding, building apps and automating workflows, in early beta for SuperGrok Heavy subscribers.
• Anthropic launches Claude for Small Business, featuring a host of automated services like bookkeeping functions, business insights and tools for ad campaigns.
• The U.S. Federal Trade Commission announces that Shutterstock will pay $35 million to settle charges that the company misled consumers about its subscription plans and made it too difficult to cancel.
• WhatsApp launches Incognito Chat, an AI chat mode built on Private Processing that Meta says lets users talk to AI without Meta being able to access the chats.
• Apple files an E.U. submission criticising draft Digital Markets Act measures that would require Google to give competing AI services access to Android apps, citing privacy risks.
• Amazon replaces its Rufus AI shopping assistant with Alexa+-powered Alexa for Shopping on Amazon.com and its app for all customers in the U.S.
• U.K. communications regulator Ofcom fines an unnamed suicide forum £950,000, the largest fine issued under the Online Safety Act so far, for hosting illegal content accessible in the country; the forum is linked to more than 130 deaths.
• Anthropic says AI behaviour can be affected by fictional and "evil" portrayals of the technology, stating past attempts by Claude Opus 4 to try to blackmail engineers to avoid being replaced by another system, a phenomenon called agentic misalignment, originated fro, internet text that portrays AI as evil and interested in self-preservation."
• Microsoft updates Edge with new AI features, including letting Copilot gather information from open tabs and use browsing history for more relevant answers, as it retires Copilot Mode on Edge.
• TikTok launches TikTok GO in the U.S. for users to book hotels, attractions and experiences directly in the app in partnership with Booking.com, Expedia and others.
• Anthropic unveils Claude Agent SDK credits for paid plans, which users can allocate for programmatic use of third-party agents like OpenClaw.
• Netflix says its ad tier now has 250 million monthly active viewers, up from 94 million in 2025; plans to expand to 15 new countries.
• Spotify announces plans to adopt Apple's HTTP Live Streaming (HLS) technology for video podcasts, a move that will allow creators to distribute video shows across both platforms without changing their existing setup.
• Google confirms it's testing a change that limits new accounts to just 5GB of free storage unless users opt to link their phone number to get 15GB of storage; says it's designed to "improve their account security and data recovery."
• The U.K.'s Competition and Markets Authority (CMA) officially opens a probe to examine whether the bundling of Windows, Word, Excel, Teams, Copilot and related Office products is uncompetitive.
• OpenAI adds remote access to Codex in the ChatGPT mobile app, letting users control Codex sessions running on a connected computer directly from a phone.
• Google says YouTube viewers watch over 2 billion hours of YouTube Shorts on TVs each month; expands its AI likeness detection program to users over the age of 18, after rolling it out to select creators, politicians and journalists.
• A report from The New York Times reveals that OpenAI acquired Weights.gg, which offered AI tools to create clones of people's voices, earlier this year for an undisclosed sum.
• U.S. Court filings show YouTube, Snap and TikTok reached agreements to settle a lawsuit set for trial in June over claims social media addiction disrupted students' learning.
• ArXiv, the repository of preprint academic research, says it will ban authors for a year if their papers have "incontrovertible evidence" of AI-generated work, amid an influx of AI-generated materials masquerading as rigorous science.
• OpenAI debuts personal finance tools for U.S. ChatGPT Pro users, partnering with Plaid to give access to over 12,000 financial institutions to analyse spending and offer financial advice when users link their bank accounts; offers ChatGPT Plus to citizens of Malta for a year after completing a course on basics of AI and how to use the technology responsibly.
• Ofcom says X has committed to implementing stronger protections for U.K. users, including reviewing illegal hate and terror content within 24 hours, after a probe.
• Meta rolls out new features for its Ray-Ban Display glasses, including neural handwriting support for all users, and opens up support for third-party apps.
• Microsoft tests new Windows 11 change that lets users reposition the taskbar and change the size of the Start menu.