Skip to content

#产品更新

14 today

May 20Wednesday

AI HOT (Curated Pool)

Claude Code’s HTML Output: The Unreasonable Effectiveness of HTML

The Claude Code team is shifting its primary output format from Markdown to HTML, and the post names four mechanisms: tables, CSS styling, SVG charts, and JavaScript interactions.

Why it matters: The Claude Code team switched its main output format from Markdown to HTML and listed four concrete uses, tables, CSS, SVG and JavaScript interaction, for structured and interactive content.

AI HOT (Curated Pool)

ChatGPT Image Generation Surpasses 1.5 Billion Uses Per Week

OpenAI says users generate more than 1.5 billion images per week in ChatGPT, and the post discusses new use cases and trends since the release of Images 2.0.

Why it matters: The weekly image-generation volume shows how large image creation has become inside a chat product.

TechCrunch · AI

You Can Now Talk to Your Gmail Inbox, as Seen at Google I/O 2026

Google expanded Gmail’s AI Inbox with conversational voice search, letting users ask Gemini to find details buried in email. The RSS snippet does not disclose rollout scope, supported languages, pricing, latency, or the retrieval mechanism behind Gmail search.

Why it matters: Google put voice conversation search into Gmail, so users can ask Gemini to pull specific details out of their mail.

AI HOT (Curated Pool)

Google launches new AI search box with multimodal interactions

Google launched an AI search box based on Gemini 3.5, combining AI Overviews and AI Mode into one AI search experience that supports multimodal multi-turn queries across text, images, files, and video, with global availability on desktop and mobile.

Why it matters: Google merged AI Overviews and AI Mode into one search box that takes mixed text, image, file and video queries with follow-ups, live worldwide on desktop and mobile.

The Verge · AI

Google’s AI Future Demands Trust — and Your Personal Data

Google presented Gemini Spark, Daily Brief, and expanded Gmail AI inbox access at I/O 2026; the Verge snippet says these tools depend on large amounts of personal information, but the post does not disclose detailed data-handling terms.

Why it matters: Google's I/O 2026 tools, Gemini Spark, Daily Brief and the Gmail AI inbox, all lean on large amounts of personal data, which is where the trust question starts.

Financial Times · Technology

Google to Release Smart Glasses and Add AI Agents to Search Engine

Google will release smart glasses and add AI agents to its search engine; CEO Sundar Pichai says features powered by a new Gemini model will narrow the gap with Anthropic and OpenAI, while the RSS snippet does not disclose specs, launch timing, or pricing.

Why it matters: Google plans smart glasses and AI agents in search, powered by a new Gemini model, with Pichai framing the goal as closing the gap with Anthropic and OpenAI.

AI HOT (Curated Pool)

Production Guide for Claude Operating Real User Interfaces

ClaudeDevs published a production guide for Claude computer use, and the snippet lists four mechanisms: click accuracy, thinking effort level selection, context retention in long sessions, and replayable demonstration logging.

Why it matters: The post breaks down four production mechanisms behind Claude operating real interfaces, from click accuracy to replayable action logs.

AI HOT (Curated Pool)

OpenAI launches Guaranteed Capacity for long-term compute access

OpenAI launched Guaranteed Capacity, a service for customers to secure long-term access to OpenAI compute and plan critical workloads under capacity constraints; the post does not disclose pricing, contract duration, or quota levels.

Why it matters: OpenAI now sells reserved compute, a sign that capacity, not model quality, is what enterprise buyers are negotiating over.

AI HOT (Curated Pool)

Google Tensor ML SDK Beta Released

Google released the Tensor ML SDK beta, letting developers convert, compile, and run PyTorch or TFLite models on Pixel 10 TPUs through LiteRT, with a model library containing more than 100 classic and generative AI models, including Gemma 3.

Why it matters: Developers can now run their own PyTorch or TFLite models on the Pixel 10's TPU, which makes on-device inference testable without Google's cloud.

AI HOT (Curated Pool)

Gemini Omni launches with physical reasoning and multimodal generation

Google launched Gemini Omni video generation for global AI Plus, Pro, and Ultra subscribers, integrating it with Gemini app, Google Flow, and YouTube Shorts, while the RSS snippet says the model combines intuitive physical reasoning with Gemini’s historical, scientific, and cultural knowledge.

Why it matters: Gemini Omni reasons about how objects fall and collide, so generated video holds together instead of just looking plausible frame by frame.

AI HOT (Curated Pool)

Gemini Omni: A new model for creating content from any input

Google AI published a Gemini Omni post, saying the model creates content from any input starting with video; the post does not disclose parameters, pricing, availability timing, or benchmark results.

Why it matters: Google teased a model that generates from any input and leads with video, but has published nothing beyond a tweet.

TechCrunch · AI

Google takes a page from Meta, announces audio-powered smart glasses at I/O 2026

Google announced “audio glasses” at I/O 2026, letting users issue voice commands across its apps and services, including Gemini; the RSS snippet does not disclose price, launch timing, or hardware specifications.

Why it matters: Google put Gemini in screenless audio glasses that work on both Android and iOS, a direct answer to Meta's Ray-Ban line.

AI HOT (Curated Pool)

Google releases Gemini Omni multimodal generation model

Google released Gemini Omni, a multimodal generation model that combines image, video, and text inputs to generate videos grounded in Gemini’s real-world knowledge; the post does not disclose model size, pricing, or availability.

Why it matters: Google says Gemini Omni takes mixed image, video and text input and outputs video, with no model size or release date disclosed.

AI HOT (Curated Pool)

Google I/O Announces Multiple Gemini Updates

Google announced multiple Gemini updates at Google I/O, including a new experience design using neural expression technology, upcoming agent features with Daily Brief and Gemini Spark, plus Gemini Omni and 3.5 Flash; the RSS snippet does not disclose release dates, model parameters, pricing, or benchmark results.

Why it matters: Google used I/O to show where its Gemini line is heading, with design updates, agent features and the Omni and 3.5 Flash models.

AI HOT (Curated Pool)

Gemini surpasses 900 million monthly active users and reviews key annual feature releases

Gemini has surpassed 900 million monthly active users, and the post attributes part of that growth to a faster release cadence; the post does not disclose the specific feature list, measurement method, or time window.

Why it matters: Gemini's app passed 900 million monthly users, which Google credits to a faster release cadence, without saying how it counts them.

Hacker News front page

Gemini CLI will stop working from June 18, 2026

Google Developers Blog says Gemini CLI will stop working on June 18, 2026, and the post title points to a transition to Antigravity CLI; the RSS snippet only includes the article URL, Hacker News comments URL, 36 points, and 10 comments, and it does not disclose migration steps, compatibility details, or replacement behavior.

Why it matters: Google will shut down Gemini CLI on June 18, 2026, cutting off free, Pro and Ultra users alike, so projects built on it need to migrate first.

AI HOT (Curated Pool)

Gemini 3.5 Series Launches With Stronger Agent and Coding Performance

Google AI launched the Gemini 3.5 series with Gemini 3.5 Flash first, stating it targets agent and coding performance; the post does not disclose parameters, pricing, benchmark scores, or context window size.

Why it matters: Google's Gemini 3.5 series starts with Flash, which the company says focuses on agent and coding ability for models that do work inside business processes.

AI HOT (Curated Pool)

Google Search gets its biggest redesign in 25 years with AI-driven interaction changes

Google announced at I/O 2026 the biggest Search redesign in 25 years, powered by Gemini 3.5 Flash, with AI Mode exceeding 1 billion monthly active users and query volume doubling each quarter.

Why it matters: Google is moving search from typed keywords to full spoken requests, with AI Mode already at 1 billion monthly users.

The Verge · AI

Gemini will use Volvo’s external cameras to interpret parking signs

Google and Volvo announced at I/O that Gemini will access external cameras on the upcoming EX60 SUV, with the first stated use case translating hard-to-understand parking signs for vehicle owners.

Why it matters: Gemini reaching Volvo's external cameras shows multimodal models moving from chat into the car itself.

AI HOT (Curated Pool)

Gemini 3.5 Flash launches with stronger performance and speed

Google opened Gemini 3.5 Flash after Google I/O across its products and API; the post says it outperforms Gemini 3.1 Pro on most benchmarks and generates tokens 4x faster than other frontier models.

Why it matters: Gemini 3.5 Flash is now open to product teams and API users, with Google claiming it beats 3.1 Pro on most benchmarks.

The Verge · AI

The 13 Biggest Announcements at Google I/O 2026

Google announced Gemini 3.5 Flash at I/O 2026. It becomes the default model today for the Gemini app and AI Mode in Search, while Gemini 3.5 Pro follows next month; the RSS snippet also mentions Search, Gmail, and Project Aura smart glasses updates but does not disclose the full list of 13 announcements.

Why it matters: Google made Gemini 3.5 Flash the default model for its app and search AI Mode, with the stronger Pro version still to come.

AI HOT (Curated Pool)

Google releases Gemini 3.5 Flash with 55 intelligence score

Google released Gemini 3.5 Flash, raising its intelligence score by 9 points to 55, exceeding 280 output tokens per second, and increasing operating cost by 5.5 times versus the previous generation.

Why it matters: Gemini 3.5 Flash trades a higher intelligence score and faster output for running costs 5.5 times the previous generation.

TechCrunch · AI

Google’s Genie world model can now simulate real streets with Street View

Google DeepMind is integrating Street View with Project Genie for interactive street-level simulations in robotics, gaming, and travel; the post does not disclose model parameters, launch timing, or evaluation results.

Why it matters: Wiring Street View into Project Genie gives robotics and game teams interactive 3D scenes built from real streets.

TechCrunch · AI

With Gemini 3.5 Flash, Google bets its next AI wave on agents, not chatbots

Google launched Gemini 3.5 Flash at its annual developer conference, describing it as its strongest coding and agentic AI model yet; the RSS snippet says it can autonomously execute complex tasks and build software from scratch, but the post does not disclose parameters, pricing, or context window.

Why it matters: Google is pitching Gemini 3.5 Flash as its strongest coding and agent model, shifting its focus from chat to autonomous task execution.

The Verge · AI

Google wants to compete with Anthropic’s Mythos

Google invited select experts at I/O to test the CodeMender API, an AI agent for code security that flags and fixes vulnerabilities; the RSS snippet does not disclose launch timing, pricing, benchmark results, or concrete details about Anthropic’s Claude Mythos Preview.

Why it matters: Google is opening CodeMender's API to expert testers, putting an AI agent that finds and fixes its own bugs against Anthropic's equivalent.

TechCrunch · AI

Google Search as You Know It Is Over

Google is changing Search from a link list into an AI experience with conversational answers, autonomous agents, and interactive interfaces; the RSS snippet does not disclose a rollout timeline, publisher traffic impact numbers, or product parameters.

Why it matters: Google turned the search box into a conversational answer engine at I/O, with agents that research and build small tools, so search now aims to hand over results instead of links.

AI HOT (Curated Pool)

Google releases Gemini 3.5 Flash for complex agent workflows

Google introduced Gemini 3.5 Flash at Google I/O for long-running agent workflows; it outscored 3.1 Pro on Terminal-Bench and MCP Atlas, runs up to 4x faster than other frontier models, and reaches up to 12x speed gains in Google Antigravity.

Why it matters: Google's first Gemini 3.5 model, Flash, targets long multi-step agent workflows and outscores the previous 3.1 Pro on Terminal-Bench and MCP Atlas.

The Verge · AI

Gmail is going to start talking to you

Google is launching Gmail Live for Gmail, letting users tap a search-bar icon and ask voice questions about inbox content; a press demo retrieved school event dates, locations, and an upcoming Detroit trip from the employee’s email.

Why it matters: Gmail Live lets users ask questions out loud in the search bar to pull dates and trip details from their inbox.

The Verge · AI

Google Search is getting its biggest changes ever

Google showed a redesigned Search box at I/O 2026, using Gemini 3.5 Flash to connect AI Overviews with AI Mode; the RSS snippet says natural-language queries will reliably show AI Overviews, but the post does not disclose rollout timing.

Why it matters: Google's new search box, powered by Gemini 3.5 Flash, merges AI summaries and chat-style search into one entry point.

The Verge · AI

Would You Let Robots Spend Your Money? Google Is Betting on It

Google unveiled an AI shopping Universal Cart at I/O that lets users add products while browsing Search or chatting with Gemini, then check out through Google; the RSS snippet says future support includes YouTube and Gmail, while pricing, rollout timing, and retailer coverage are not disclosed.

Why it matters: Google's universal cart lets shoppers add items across retailers in Search or Gemini and check out in one place, with price tracking and restock alerts.

The Verge · AI

Google can now vibe-code you an Android app

Google upgraded AI Studio today to build native Android apps from prompts, with an embedded Android emulator for preview and device connection for installation; the initial release focuses on personal utility apps, and tester invites are planned but not yet available.

Why it matters: Google's AI Studio generates native Android apps from a single sentence and embeds an emulator for previews, aimed first at small personal tools.

TechCrunch · AI

Agentic app coding gets an upgrade with Google’s release of Android CLI

Google released Android CLI for AI coding agents, letting platforms such as Claude Code and OpenAI Codex build Android apps from the command line; the RSS snippet does not disclose version numbers, release timelines, pricing, or performance data.

Why it matters: Android CLI hits 1.0, letting AI coding assistants like Claude Code and OpenAI Codex build Android apps from the command line without opening Android Studio.

TechCrunch · AI

Google’s AI Studio now lets anyone build Android apps in minutes

Google unveiled web-based AI tools in AI Studio that generate native Android apps in minutes. The RSS snippet does not disclose the model, pricing, availability, or supported development constraints.

Why it matters: Google's AI Studio now turns plain language into native Android apps, cutting weeks of setup and coding down to minutes.

TechCrunch · AI

OpenAI is making it easier to check if an image was made by its models

OpenAI announced two measures for detecting AI-generated images: it joined the open C2PA standard and added Google’s SynthID to its products. The RSS snippet does not disclose which OpenAI products include SynthID, whether the checks cover legacy images, or when the measures become available to users.

Why it matters: OpenAI now stamps its generated images with both C2PA metadata and Google's SynthID invisible watermark, running two provenance paths side by side.

TechCrunch · AI

Google's Gemini Omni turns images, audio, and text into video

Google's Gemini Omni generates and edits video through conversation, using text, images, audio, and video as inputs, with Omni Flash named as the starting version; the RSS snippet says the model reasons across modalities, but the post does not disclose launch date, pricing, context limits, benchmarks, or API availability.

Why it matters: Gemini Omni takes text, images, audio and video as input and generates or edits video through conversation, with the first release, Omni Flash, focused on cross-modal reasoning.

TechCrunch · AI

Google launches Antigravity 2.0 with updated desktop app and CLI tool at I/O 2026

Google launched Antigravity 2.0 with an updated desktop app and CLI tool, and introduced a $100 AI Ultra plan that gives users 5x the usage limit of AI Pro; the post does not disclose the desktop app or CLI feature details.

Why it matters: Antigravity 2.0 ships an updated desktop app, a CLI and a workflow SDK, plus a $100-a-month AI Ultra plan with five times the quota of AI Pro.

TechCrunch · AI

Google introduces Gemini Spark, a 24/7 agentic assistant with Gmail integration, at I/O 2026

Google introduced Gemini Spark at I/O 2026 as a 24/7 agentic personal assistant with Gmail integration; the RSS snippet says it uses Gemini base models and an agentic harness from Google Antigravity, but the post does not disclose pricing, rollout timing, or supported Gmail actions.

Why it matters: Gemini Spark is a 24/7 personal assistant built on Gemini models and the Antigravity agent framework, with Gmail access to read mail and handle tasks.

AI HOT (Curated Pool)

I/O 2026: Welcome to the autonomous Gemini era

Google announced at I/O 2026 that Gemini is moving into an autonomous agent phase, with the post saying it can manage email, schedule calendar items, and generate reports automatically, but it does not disclose model parameters, launch timing, or pricing.

Why it matters: Google said at I/O that Gemini moves into autonomous agents that handle email, calendars and reports, but gave no parameters, launch date or pricing, so it reads as a direction, not a product.

AI HOT (Curated Pool)

Google AI subscription updates from I/O 2026

Google announced a $100 AI Ultra subscription at I/O 2026 and added new features and benefits for existing Google AI Plus, Pro, and Ultra subscribers.

Why it matters: Google added a $100-a-month AI Ultra tier at I/O 2026 and refreshed Plus and Pro, giving its subscription lineup three parallel price levels.

AI HOT (Curated Pool)

Google AI Ultra plan gets a price cut and a new tier

Google cut the top AI Ultra plan from $250 to $200 per month and added a $100 monthly tier with 5x the Gemini app usage limit of Pro, 20TB of storage, early access to new features, and YouTube Premium under stated terms.

Why it matters: Google drops AI Ultra from $250 to $200 a month, adds a $100 tier, and raises usage limits to five times the Pro plan.