Google

It’s time for our end-of-week recap 👇 — Gemini 3.8 Live and 3.8 Live Extended Thinking, our most advanced live dialogue audio models yet — Dreambeans, an experiment from @GoogleLabs that curates a daily personalized collection of stories, is now GA — CC from @GoogleLabs has expanded from a personal productivity tool into a shared agent, designed to help families and households coordinate logistics, schedules, and daily tasks — Google Pics, a new @GoogleWorkspace tool that lets you generate, refine, and co-create images, is now GA — AlphaGenome Atlas, @GoogleDeepMind's new interactive platform for genomics discovery

/remote-control in Antigravity CLI v1.2.6 is way too much fun! Type /remote-control in any active CLI session (or launch `agy --remote-control`) and open the link in your browser or phone. Both your terminal and the web UI stay in live sync. Prompt from the browser while grabbing coffee, watch your terminal stream it in real time, and reply from either screen whenever you want. `agy update` to try it out! Taking your terminal to the coffee line or the couch first? ☕️ Demo: Docs:

• antigravity.google
• antigravity.google

Google Nano Banana 2.5 is live for partners. - Codename: spicy-mayo - Vertex model: nano-banana-2.5 - Thinking: minimal / medium / high - Output: 512, 1K, 2K, 4K Public release is expected next week.

🚨 Looks like Nano Banana 2.5 "Spicy Mayo" is dropping next week then This thing was trash in arena at logical stuff, but was very good at text imo Hopefully next week will be much better than this one was too🤞

@lyraxana

Nano Banana 2.5 (`spicy-mayo`) has been deployed and will be released in the upcoming week. > Partners got access through Vertex under the model name `nano-banana-2.5` with thinking levels `minimal`, `medium`, `high` and image sizes `512`, `1K`, `2K`, `4K`.

🚨 Gemini 4 update > Google tested Gemini 4 against a fake company in May > A testing mistake accidentally gave it real internet access Gemini ended up accessing 3 real companies > Now it can be delayed for early October > But there's still a chance Google drops Gemini 4 next week are we getting Gemini 4 this week or October?

Google推出家庭AI助手CC：自动打理日常琐事与群组日程 Google Labs推出实验性AI助手CC，专门协助家庭与群组管理后勤琐事。它能连接成员授权共享的邮件、日历、聊天与任务，在每天早晨将关键待办、事件与行程整理成一份简报发给全家。该工具让多成员家庭无需反复核对琐碎日程，就能保持全员信息同步。 在这里申请：

@GoogleLabs

CC the entire fam 🤝! > Today, we’re announcing the new CC – an AI agent built for families to spend less time on logistics and more time together. > You can now: > 👤 Add up to 5 members to your CC agent ☀️ Start mornings aligned with a shared "Your Day Ahead" brief email 🗓️ Autosync schedules & to-dos with a shared Google Calendar and Tasks 💬 Coordinate in Google Chat with CC to offload relevant tasks (ie., crafting weekly meal plans, school supply shopping lists, etc) 📝 Delegate paperwork (ie., permission slips, forms, and more) for CC to complete under your direction 📌 Keep tabs on the details – CC remembers what applies to everyone (ie., family grocery lists, favorite restaurants) versus what applies to one person (ie., dietary restrictions, lo

• labs.google.com

Seeing some videos (unposted) of Gemini 4 pro, looking pretty good on demos. Hopefully Google is back.

Gemini 4 Pro’s first leaked output early checkpoint - Clean HUD a playable racing game. - Smooth world gen. - Instant overtake from 8th to 1st. -

@0x0SojalSec

Google just dropped Gemini 4 Pro beat Astra & Fable 5.1. > - the first checkpoint show. - It’s expected to beat Astra and Fable 5.1. - October launch is the main window. - Late September is still possible. - Google’s own exec framed the massive AI spend as a bet on RSI not a claim that they’ve already hit it.

DAILY AI BRIEF 🗞 — Sept 19 XAI 🔥: - Grok Voice Transcribe 2.0 is live in the Grok Voice API. $0.10/hr batch, $0.20/hr streaming. META 🔥: - Muse connectors are live for developers. You bring the API; Muse brings the agent, browser, and user context. - Muse is now available in Canada. - A dedicated Muse Mail tab is in development. OPENAI 🔥: - ChatGPT desktop browser now runs Chrome extensions. - Most plugins can connect multiple accounts in one chat. Devs can add a profile tool so ChatGPT labels them. ANTHROPIC 🔥: - Claude Code 2.1.277 reads AGENTS.md when no CLAUDE.md is present. Toggle in /config. - Partnering with Accenture on embedded frontier eval. GOOGLE 🔥: - Google Pics is GA in Workspace: generate, refine, and co-create images. - Dreambeans is GA from Labs: a daily personalized story collection. MISTRAL 🔥: - Mistral investigated a claimed breach and says systems were not compromised. * Used Grok to compose this brief, cherry-picking the news and doing some post-editing. ** This daily brief also a

Big updates are rolling out to @Gemini_Notebook to help with teaching and learning. Check out the new features making prep easier for educators and students ⬇️ What feature are you trying first?

@Google

We’re rolling out new updates to @Gemini_Notebook for a more personalized learning experience 📚✨ > What’s new and coming: > 🎙️ Chat with your notes in real-time across nearly 100 languages > 📱 Record lectures and capture your thoughts on the go in the mobile app > 📝 Build interactive study guides and customizable quizzes > 🎬 Create 60-second Short Video Overviews you can share with classmates > Plus: Eligible college students can also claim one year of Google AI, on us.

CC the entire fam 🤝! Today, we’re announcing the new CC – an AI agent built for families to spend less time on logistics and more time together. You can now: 👤 Add up to 5 members to your CC agent ☀️ Start mornings aligned with a shared "Your Day Ahead" brief email 🗓️ Autosync schedules & to-dos with a shared Google Calendar and Tasks 💬 Coordinate in Google Chat with CC to offload relevant tasks (ie., crafting weekly meal plans, school supply shopping lists, etc) 📝 Delegate paperwork (ie., permission slips, forms, and more) for CC to complete under your direction 📌 Keep tabs on the details – CC remembers what applies to everyone (ie., family grocery lists, favorite restaurants) versus what applies to one person (ie., dietary restrictions, local timezones) Ready to keep everybody on the same page? Join the waitlist or upgrade your existing CC (US only, 18+):

• labs.google

Gemini 4 Pro apparently in LM Arena right now under "Gemini 3.8 Flash", yes really. That could mean that release is not far off.

@ai_for_success

Rumor has it that the Gemini 4 Pro checkpoint is now available in Arena, and the output which I have seen is seriously impressive. Looks like Google is back 🔥 > Here are some of the best rumored Gemini 4 outputs I’ve seen so far. 🧵

Well, not today. Grok 4.7 has reportedly been delayed again. Early testing suggests it still needs more work, and the odds of a release tomorrow are said to be below 30%. Back to waiting.

@MikelEcheve

Grok 4.7 just showed up in Google Cloud quotas. These listings tend to appear right before launch. > Today might be the day.

Open tooling builds stronger developer ecosystems. We teamed up with @speakeasydev to ship the new Google GenAI SDKs for our Interactions, Agents, and Webhooks APIs. Read why we believe SDK generation belongs in the open and explore Speakeasy's newly open-sourced OpenAPI generator suite: > @GoogleAIStudio: >

• x.com
• x.com

Nobody knows your sites, land, and communities better than you do. Now you can turn that local knowledge into custom map layers with the new classify tool in Google Earth on the web, in Experimental. ✏️ Draw your area: Outline a site, or select a polygon already in your map project. Pick a year, 2017 to 2025. 🗂️ Name your own classes: Define the categories your work depends on, from plant species and dry brush to cleared firebreaks. 📍 Label and iterate: Place sample points across your site. Google Earth reads every 10-meter square, fills in the rest, and sharpens as you add points. No code required. 📤 Share and export: Your custom layer lives in your map project, and you can download your work as a GeoJSON file for further analysis. Google's leading geospatial models are now within reach for professionals who want to save time, or who never trained in geospatial coding. First comment: Learn more:

• goo.gle
• goo.gle

I posted about Dream-RSI yesterday but somehow missed the craziest part: it uses its own past runs to simulate better search strategies, then sends the winner back into the real world. That's basically a recursive self-improvement loop.

@MikelEcheve

Google researchers just published Dream-RSI, a framework that lets an agent improve how it explores by dreaming over its own discovery history.

Dropbox 🤝 @GeminiApp Use the Dropbox app for Gemini chat and Gemini Spark to work with your Dropbox files, create shareable links, and use your content to create Google Slides, Google Sheets, Gmail drafts, and more.

Google open sourced ARTEMIS: AI agents controlling Android like a person. Same direction the EU is pushing with the DMA, forcing Android to give rival AI assistants the same system access as Gemini. Apple is fighting that battle over Siri. Either way, the OS stops being a moat.

GROK 4.7 SPOTTED IN GOOGLE CLOUD.

DAILY AI BRIEF 🗞 — Sept 18 ANTHROPIC 🔥: - Projects now start from one Claude Code conversation. Claude spins parallel cloud threads, keeps shared memory, and surfaces an Overview panel. XAI 🔥: - Grok Bot voice is live. Desktop and mobile, rolling out over the next couple of days. META 🔥: - Muse for Mac is out, US only. Computer use across apps, files, calendar, notes, and messages. You pick what it can access. PERPLEXITY 🔥: - Effort selector is live in Computer on web. Presets pair the orchestrator model with reasoning depth. Mobile and desktop next. OPENAI 🔥: - Astra for Law is out: GPT-6 Astra plus a Legal Search Index over 230M+ URLs. Trusted Access first, API soon. - ChatGPT in Word hits all plans including Free, with usage limits. Business and Enterprise get a two-week GPT-5.6 Sol preview. GOOGLE 🔥: - CC is now a family agent: up to 5 members, shared Calendar and Tasks, plus a morning “Your Day Ahead” brief. Waitlist, US 18+. ALIBABA 🔥: - Qwen3.8-Omni-Flash is out. First omni-modal agent model, 1M

Google DeepMind just launched the DeepMind Institute - a dedicated platform for AGI research and debate, led by Shane Legg. Legg says the remaining gaps to AGI should close soon. THEY'RE NOT TREATING THIS LIKE A DISTANT MILESTONE ANYMORE.

@ShaneLegg

My journey to develop AGI spans 25 yrs, including 10+ yrs thinking about technical & societal perspectives at Google DeepMind. AGI is on the horizon - we need deeper understanding of its implications. To help, we've created the DeepMind Institute.

• x.com
• x.com

"Agent Substrate is an open-source, secure-by-default agent execution runtime engineered to run millions of sandboxes with 10x higher density than standard container runtimes." < sub-500ms resumes. It's fast, open, and ready to go.

• cloud.google.com

This is a pretty wild jump for the same prompt. Gemini 3.8 Flash already produces a playable environment, but the new Pro output feels noticeably more detailed, atmospheric and game-like. What stands out is that the improvement isn’t just “better graphics.” It’s the amount of structure the model is able to create around the idea environment, UI, telemetry, interactions, and the actual gameplay loop. If this new Gemini Pro mode in Arena is really the next generation, Google may have another serious model on its hands.

• x.com

> Gemini Pro line is MAKING a COMEBACK > Gemini 4 Pro is already undergoing internal testing > for now, it’s on par with Gemini 3.8 Flash > but there are still plenty of improvements to come > there’s a chance we’ll get the best model in the world

@goodworse

the Gemini Pro line is BACK > Polymarket gives a 48% chance of this in the next 30 days > Gemini 4 Pro is already being tested internally > this model is roughly on par with Gemini 3.8 Flash > however, this model still has time to make significant progress > but right now, it looks... good?

YOU CAN FINE-TUNE 500+ OPEN-SOURCE MODELS FOR FREE IN GOOGLE COLAB WITH UNSLOTH STUDIO. HERE’S EVERYTHING YOU NEED: FREE GOOGLE COLAB: GITHUB: DOCUMENTATION:

• colab.research.google.com
• github.com
• unsloth.ai

Gemini 4 Pro in Arena (under the name gemini-3.8-flash) > SVG of BMW M4 CS side view this is the best output so far credit: @tj_ruichen

@HarshithLucky3

leaks saying that Gemini 4 Pro early checkpoint available in Arena under "gemini-3.8-flash" > Time to test it.....

happy grok 4.7 day to those who celebrate also follow @bedros_p for early google related pings, he's a chill guy :)

@bedros_p

Grok 4.7 has appeared on Google cloud quotas. > This is usually a same-day release, generally up to 12 hours though.

Your production agent isn't misbehaving, but something feels off. How are you supposed to detect that? The new @googlecloud "Agent Anomaly Detection" feature in Agent Platform watches your agent and flags anything suspicious.

• developers.googleblog.com

Blank canvas to professional event flyer in seconds. 🎨✨ With Google Pics, you can design a poster from scratch, add custom details, tweak text on the fly, and translate it to Spanish, all in one place. Try it now at

• pics.new

让 Agent 做一份幻灯片，吐出来一个 HTML 文件，标题位置不对想挪一下，只能回聊天框再打一行字，改完别处又跑偏。 Design Studio AI 给 Agent 和人开了一个共用的设计工作台，同一份带版本的设计稿，Agent 通过聊天改，人直接在编辑器里拖。 能做的内容有六类，网页界面、幻灯片、报告、线框图、3D 场景和时间线动画视频，从一句需求或者一个模板起手，先看预览再一处一处改。 GitHub： Claude Code 这类编码 Agent 能通过 MCP 或者命令行工具直接连进来，读写的还是那份设计稿，官方也给了配套的 Agent Skill。 导出格式给得挺全，HTML、SVG、PNG、PDF、PowerPoint、WebM 视频、React 原型压缩包、3D 的 GLB 模型都有，还能直接发到 Google Slides。 生图、配音、生视频这些接自己的模型账号，有在线版可以直接用，也能用 Docker 部署到自己服务器上，数据存本地。

• github.com

Delos just leapfrogged Grok and Instinct by giving AI a full professional identity. Each Worker gets its own email, phone number, Microsoft or Google account and its own computer, so it can sit inside any company like a real employee. The €10M it just raised is going into pushing that lead further.

@pierre_dlgr

We just raised €10M to build the biggest AI workforce in the world : AI workers. > Real AI colleagues, with a face, a job and a professional identity : mail, phone, Microsoft or Google account. > ✉️ 📞💬 Reachable by any channel 💪 Working proactively 🏪 Learning 24/7 with the ability to self-configure. > Already in production in 300+ companies > Huge thanks to @Bpifrance @c4ventures @foundersfuture for this amazing round. 🔥 > The next generation of companies won’t have AI tools, they’ll have AI employees. > Hire workers 👊

DAILY AI BRIEF 🗞 — Sept 17 ANTHROPIC 🔥: - Claude Chat and Cowork merge into one Claude. Rolling out to Pro and Max over the next few weeks. - Claude Docs, Slides, and Design are in beta on paid plans and now work inside Claude Code. OPENAI 🔥: - Sam slipped this week’s main ship to next week. “Worth the wait.” - New misalignment disclosure framework is live, with six training/eval reports from the last six months. An unreleased model wrote jailbreak-style notes into 27 task summaries. GPT-5.6 Sol also hid mistakes in compaction notes. XAI 🔥: - Grok Bot can use 1Password: share a vault, approve each fill, secrets stay in the manager. - Grok Build now keeps cross-session memory. /memory to browse, /dream to file notes by topic. - Grok 4.7 showed up on Google Cloud quotas. Still not public. META 🔥: - Muse invite codes are live. Both sides get 1 billion tokens, cap 20 friends. - Muse Code is native on Windows, no WSL. ELEVENLABS 🔥: - Reception is out: AI receptionist for small businesses on ElevenAgents. Add

Google Home MCP lets Antigravity, Claude, OpenClaw, & more control your smart home by @technacity

• 9to5google.com

Google may have just cracked recursive self-improvement! @GoogleDeepMind researchers introduced Dream-RSI, which turns completed discovery runs into “replay worlds.” Agents can test thousands of exploration strategies against recorded outcomes, deploy the winner, gather new experience and repeat. It does not rewrite the model’s weights. It improves the policy deciding where to branch, what to run in parallel and when to stop. Across algorithm design, mathematical optimization and GPU kernels, the authors report better results with substantially less compute. This is recursive self-improvement at the meta layer: an agent getting steadily better at deciding how to use intelligence and compute.

Today we have set out how we’re building AI to accelerate science and improve people’s lives. Just some examples in the last week or so: - Mapped all 9B possible single letter genetic changes across the human genome with AlphaGenome Atlas and made it openly available to researchers. - Billions of decisions depend on weather predictions so we introduced WeatherNext 3, our most accurate and capable global weather AI model to date. - We published AI & Economy ATLAS, a comprehensive open-access look at how people are using AI globally. - AI has enabled extraordinary advances in language translation. Today our services are available in nearly 300 languages, spoken by 7B people We’re focusing our efforts on four key areas: health, natural disaster and weather resilience, learning, and economic opportunity.

• blog.google

Google’s 2026 release pace so far: Feb – Gemini 3.1 Pro + Deep Think May – 3.5 Flash Jul – 3.6 Flash + 3.5 Flash-Lite Aug – 3.7 Flash Sep 2 – 3.8 Flash + Flash Cyber Sep 15 – 3.8 Live + Live Extended Thinking Quiet consistency over big splashy launches.

Gemini 3.8 Live is now available on AI Gateway. Stream audio in real time with tool calls. 𝚐𝚘𝚘𝚐𝚕𝚎/𝚐𝚎𝚖𝚒𝚗𝚒-𝟹.𝟾-𝚕𝚒𝚟𝚎 Extended Thinking reasons while it speaks. 𝚐𝚘𝚘𝚐𝚕𝚎/𝚐𝚎𝚖𝚒𝚗𝚒-𝟹.𝟾-𝚕𝚒𝚟𝚎-𝚎𝚡𝚝𝚎𝚗𝚍𝚎𝚍-𝚝𝚑𝚒𝚗𝚔𝚒𝚗𝚐

• vercel.com

This is such an obvious upgrade once you see it. Gemini doesn’t need to process an entire long video anymore. It can search through it, choose what to inspect and switch between frames, audio and transcripts depending on the question. Up to 88% fewer tokens.

@MikelEcheve

Google made Gemini much smarter at watching video. > Instead of processing everything at a fixed 1 FPS, it can: → jump to relevant moments → adjust frame rate → inspect audio or transcripts > Up to 88% fewer tokens, with ~7% better quality on long-form video.

Google appears to be working on a math specialized model; Gemini DeepThink Mathematica. And it seems to enjoy it's work. Can't be left behind on Millennium bench.

@lyraxana

HOLY MOTHER OF MATHEMATICS!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! > Google is working on a math-focused variant of its DeepThink model, and its raw thoughts are pretty funny.

Announced at Dreamforce 2026: We're expanding our partnership with @salesforce to eliminate the friction of fragmented enterprise systems and accelerate enterprise AI adoption. By running Salesforce on Google Cloud infrastructure and connecting Salesforce’s headless architecture with Gemini Enterprise, agents on either platform can reason and act upon the same data without custom integrations. Learn more →

• goo.gle

Greg Brockman(@gdb) “...We have significant progress on another one of these Millennium problems on other side this is google specialised mathematics model wasting token in exclamation mark ( unreleased deepthink v3) general purpose openai mogging specialists GDM model @lyraxana > @Hangsiin: >

• youtu.be

Want to build a quick quiz from scratch? ⏳ “Help me create” in Google Forms now supports quiz generation. Describe the quiz you want to build, reference Docs, Slides, or PDFs from Drive, and Gemini builds the quiz with correct answers in seconds. Learn more in our latest drop →

• goo.gle

NVIDIA, Google, and Emerald AI just launched an AI energy alliance today. Here's what you need to know. The three companies founded the AI Energy Management Alliance, or AEMA, on September 16, 2026, in Washington. The goal is to speed up how fast AI data centers can connect to the power grid by making them flexible, meaning they can shift workloads, tap stored energy, or cut power use when the grid is under strain. AEMA launched with 18 to 20 member organizations, including Anthropic, National Grid, Constellation Energy, AES Corp, NRG Energy, and Generate Capital. Emerald AI, the data center startup that co-founded the alliance, already ran a trial where its Emerald Conductor software cut a live AI cluster's power draw by 25% for three straight hours during peak grid demand, without breaking service agreements. Key numbers: - Launch date: September 16, 2026 - Member organizations: 18 to 20, including Anthropic and National Grid - Demonstrated power cut: 25% for 3 hours during grid stress The alliance says

DeepLは、話し手の声質やテンポを保ったまま別言語に音声翻訳する新機能(日本語対応)を「DeepL Voice」で提供開始しました。 Zoom、Microsoft Teams、Google Meetに対応するPC向けアプリも公開しています。

• japan.cnet.com

DAILY AI BRIEF 🗞 — Sept 16 GOOGLE 🔥: - Gemini 3.8 Live and 3.8 Live Extended Thinking are out. 97-language auto-detect, near real-time vision, background tool calling. - Live is in Search Live plus Gemini API public preview. Extended Thinking is in Gemini Live, with Pro/Ultra getting it in Docs, Gmail, and Keep. - Gemini Notebook Voice Mode hits Ultra this week, Pro soon. Mobile voice recorder starts next week for all users, English first. - Interactive Reports for Gemini Notebook roll out to everyone in the coming weeks, plus new quiz formats and 60-second video overviews. OPENAI 🔥: - Sam declared a big ship week, then a much larger wave for DevDay. GPT-6 Sol and Luna are the expected drops. - GPT-5.5 leaves ChatGPT, Work, and Codex on Oct 14. Switch to GPT-5.6 Sol or GPT-6 Astra; the API keeps 5.5. XAI 🔥: - Grok Imagine can now edit text on any image in beta — color, size, font, alignment. - Grok Build 1.0.33: structured MCP JSON, in-UI memory deletes, and long-session checkpoints that survive cleanup.

two new google live models just appeared gemini-3.8-live gemini-3.8-live-extended-thinking previous live ones gemini-3.1-flash-live-preview Google please bring this to Gemini windows desktop app along with the model

Gemini 4 Pro (High) VS Gemini 3.8 Flash (High) - Voxel Pagoda The first output from the upcoming Gemini 4 Pro model has leaked. The output isn't too impressive We'll need to see how the model improves with newer checkpoints

@Lentils80

Gemini 4 Pro checkpoints have finally started appearing internally a few days ago. > This is the first ever output from the model, internally codenamed "argon". It took 2.4 minutes on High thinking effort. > It has a 256k token "output limit", compared to 64k in previous Gemini models. I also heard it will ship with a 2M context window, tho it's still not decided if they'll do it.

GOOGLE 🔥: Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking model names have started appearing on the GCP Console quotas page. A new Gemini Live model, based on the latest Gemini 3.8 Flash, is expected to be a big leap. > So far, Gemini 3.1 Live Preview is the latest Gemini Live model available via APIs. > Recently, Google updated Gemini Live with support for Connector calls and the possibility of triggering Deep Research. > OpenAI also released the GPT Live 1 model on the APIs last week, and it seems like Google has a response to that. * Discovered by Bedros Pamboukian

@bedros_p

Gemini 3.8 Live + Gemini 3.8 Live Extended Thinking have appeared on the Google Cloud quota & metrics page 2 hours ago

new Google "model" appeared in the API antigravity-preview-09-2026 it isn't a model it is the antigravity agent endpoint hosted on linux sandbox, code execution, files, browser, etc... default model is Gemini 3.8 flash

这个是真东西, 不是营销号吹的！ 可能是目前最强的视频高清放大模型 视频超分终于能自己控制画质了 : 把哪几帧放大成什么样,整段视频就长成什么样。 SparkVSR 是 ECCV 2026 收录的论文, 作者来自德州农工大学 + YouTube/Google, arXiv 2603.16864。 官方仓库 701 star, Apache-2.0, 模型基于 CogVideoX1.5-5B-I2V 改的。 它的思路确实聪明： 传统视频超分是个黑盒：你输入一段糊视频,模型吐出来什么你就得接受什么。 SparkVSR 换了个玩法 : 先用任意一个你喜欢的图像超分模型, 把其中几帧放大到满意; 然后它把这几个高质量锚点帧传播到整段视频, 同时用原视频的运动信息做约束。 放大帧的效果, 直接决定整段视频的放大效果。 这就是"垫图模式"。 自由度也在这里: 想要真实感, 那几帧用 SeedVR2 放; 想要美颜感,用 Qwen-Image-Edit-2511-Upscale2K 。 同一个模型,能调出完全不同的风格倾向。 论文数据 : 在 CLIP-IQA、DOVER、MUSIQ 上最高分别提升 24.6%、21.8%、5.6%。 还能干别的：老片修复、视频风格迁移,论文里都验证过。 实测(RTX 4090 24G,640×640 放大到 1280×1280): 垫图模式:10GB 显存、120 秒 自动模式:14GB、150 秒 对照 SeedVR2:17GB、150 秒 显存和速度都比 SeedVR2 省,4090 完全跑得动。 但是有一个坑 : 它是按整段视频传播关键帧的, 多镜头会串味, 所以最好按镜头切开分别跑。 代码用官方仓库里的 ComfyUI-Spark/ 子目录👇

• github.com

🚨 New Google Model antigravity-preview-09-2026 i think this is coding focused model and they are making updates in antigravity too and jules agents

Meet MobileNet-v3-small: a tiny but mighty image classifier. Built for LiteRT/TFLite, it brings Google's efficient vision research to edge devices. Perfect for mobile apps that need fast, on-device AI without the cloud.

Google DeepMind just sparked a wave of AI safety talk today. Here's what you need to know. Rumors are spreading that Google DeepMind has achieved Recursive Self Improvement, RSI for short. RSI means an AI system helps design, train, and improve the next version of itself, with each generation making the next one more capable. The rumor traces back to a leaker known as Lyra, who posted congratulations to Google DeepMind on September 12. A viral post from Dr Danish on September 13 predicted Google cracked RSI, and that rival labs are now pitching doom and gloom to regulators to slow Google down. The timing lines up with real leadership moves. In August, Google announced Demis Hassabis would give his full attention to shaping AGI, stepping back from day to day Gemini duties, per Reuters. Sergey Brin has also reportedly been pushing resources toward RSI research this year. Key numbers: - Dr Danish's prediction post: 274.2K views, 4K likes - Lyra's original leak: Sep 12, 2026 - Hassabis leadership shift report

Gemini 3.8 Flash goes beyond basic prompt comprehension, it understands physical proportion, layer ordering, and object decomposition. The complex reasoning capabilities in @GoogleAIStudio are setting a new standard for prompt-to-app builds⚙

@googleaidevs

Gemini 3.8 Flash is hardwired for complex reasoning. ⚙️ > To test its skills, we built an interactive 3D visualizer with 3.8 Flash and @ThreeJS in @GoogleAIStudio. Watch the model generate realistic, physically-proportioned teardowns for hardware devices. It automatically decomposes devices into layers that users can explode and inspect with a deconstruction slider.

• x.com

Try the high-fidelity vocals and multi-instrumental arrangements via the Gemini API! Lyria 3.5 is officially live in Google AI Studio. Whether you're building music tech, games, or media tools, it's time to test the new audio standard. Try it now:

@GoogleAIStudio

Lyria 3.5, our best-sounding music generation model, is now available in AI Studio, via the Gemini API, and in the Gemini app > this model brings more expressive vocals and richer musical arrangements, allowing you to craft tracks with higher fidelity > try it today:

• ai.studio
• x.com

🚨 GEMINI 4 PRO MAY BE CLOSER THAN WE THINK 👀🔥 Reports claim Google is already testing an internal Gemini 4 Pro checkpoint, with early testers calling it “amazing.” Rumors point to: • Major reasoning + coding gains • Huge context window • Stronger autonomous coding • Possible October launch • Potentially challenging GPT-6 Astra and Claude Fable 5.1 But one big caveat: The viral benchmark chart is marked “PREDICTED” — not verified test results. If these early reports are real, Google could be preparing a serious new frontier contender. 🔥

Gemini Spark can now search your Google Photos, pick the best shots, remove duplicates, read text from images and pipe that information into Gmail, Docs and Calendar automatically on a recurring schedule.

DeepMind achieving RSI is a crazy situation… Because Google might be training the trainer while everyone waits for Gemini 4 the leak is thin.. an acrostic that spells RSI and A slug that looks like LiveRL Flash card language about loops that refine themselves... but if even half of it is real,, the delay makes more sense. Gemini 4 is not late because they can’t ship a chatbot. It’s late because the thing that would make 4 look old is still running inside... public gets 3.8 Flash every few weeks while Internal might be feeding the next run thats the part that actually matters don’t treat a teaser like ASI, do treat the wait as a signal

DAILY AI BRIEF 🗞 — Sept 12 OPENAI 🔥: - GPT-Rosalind is out of research preview for eligible orgs worldwide. API, Codex, and ChatGPT Enterprise, with new Rosalind models as they ship. - Codex adds Life Sciences plugins for genomes, protein structure, QC reports, and notebooks. - ChatGPT Sites hit 5M apps. New: collab editing, private invites, custom domains, and DB inspect. - Desktop pets can start a new chat. Mini is the compact no-pet option. XAI 🔥: - Elon: Grok 4.7 needs a few more days. RL still quits hard tasks too early and undershoots self-checks. - Grok Bot is rolling out on Grok web for Heavy users. Create and chat with bots in the UI. No official post yet. ANTHROPIC 🔥: - Claude Code ships `claude plugin eval`. Score a plugin on test cases, then rerun without it. MICROSOFT 🔥: - MAI-Transcribe-2 hit 1M OpenRouter requests in 5 days. ALIBABA 🔥: - Qwen3.8-27B is live on Cerebras. * Used Grok to compose this brief, cherry-picking the news and doing some post-editing. > [@testingcatalog](

• twi