分类筛选
找到 1605 个 Demo
ParseBench is now live on . The first document OC...
ParseBench is now live on @Kaggle. The first document OCR benchmark built for AI agents — 2,000 enterprise pages, 167K+ test rules, 5 dimensions that actually break downstream agents. Benchmark your parser against 14 methods including GPT-5 Mini, Gemini 3, Textract, and LlamaParse. Read the full story → https://t.co/KPYA6gm5kH
#gemini3 1800m×7本（R120"） 設定：4’00→3’40/km 結果：3‘58-...
#gemini3 1800m×7本（R120"） 設定：4’00→3’40/km 結果：3‘58-56-51-46-47-41-34 今週は1800mに増やしたけど先週よりも楽に感じたので成長の証🔥そして見てこの綺麗なビルドアップ流石トモコーチ✨NEWコーチ見たことあると思ったら菅野さんだった🤣人も増えたし頑張るぞ😤 #まるお製作所RC https://t.co/NNcnLjV2c4
昨日は久しぶりの #ジェミラン 最近ジョグしかできていなくて、高強度のトレーニングは久しぶり...
昨日は久しぶりの #ジェミラン 最近ジョグしかできていなくて、高強度のトレーニングは久しぶり 最初はキロ4がキツすぎて絶望したけど、最後はしっかり上げて本数もこなせて良かった 着実に積み上げていきたい #まるお製作所RC #Gemini3 https://t.co/V8PgajUVdZ
毎週水曜日の #Gemini3 練習会 メニュー： 1800m × 7 interval（設定：04...
毎週水曜日の #Gemini3 練習会 メニュー： 1800m × 7 interval（設定：04:00 → 03:40 / km build up、rest：60〜80 sec） 先週と同様のメニュー 主観的には 4本目ぐらいまではそこまでキツくない感じで進められ（あれ？こんなもんだっけ？、ぐらいの感覚）、その後は徐々にキツくなった感じ（オールアウトというほどではない） 練習開始前のアドバイスをもとにいきなり突っ込むのではなく、本数を重ねる毎に徐々にペースアップ & rest なしでもいけるか？、ぐらいの感触を意識して走った ある程度はまとまったのかな…？（たぶん…） 回数を重ねる上でこのペースがもっと余裕に感じる & 速いペースで走れるようにしないと #まるお製作所RC
Google just upgraded Gemini inside Chrome. Here...
Google just upgraded Gemini inside Chrome. Here are the numbers that matter. Most people have no idea any of this exists. → Built on Gemini 3 — Google's most intelligent model to date → Up to 10 open browser tabs shared as context simultaneously → 5 Google apps connected natively: Gmail, Calendar, YouTube, Maps, Flights → 0 extra installs required — it's already in Chrome right now → 0 cost for: side panel, multi-tab context, app integration, image editing → 1 paid feature: Auto Browse (Google AI Pro / Ultra subscribers, US only) → Personal Intelligence rollout: coming in the next few months → Available on: Windows, Mac, Chromebook Plus, Android, iOS → Primary rollout: US with Chrome set to English → Image editing: inline, prompt-based, no download required The four core features are free. The persistent side panel is live. What you're currently spending on tab management? The real cost is the mental energy — and Google just handed you a refund. ✅Save this post — you'll quote these numbers when someone asks why you've changed how you browse. Want the SOP? DM me.
Google rolled out a new automated frontend design...
Google rolled out a new automated frontend design feature within Google AI Studio, allowing Gemini to instantly generate and apply complete UI/UX designs while an application is actively being built. Instead of manually writing CSS or piecing together frontend components, users can now seamlessly select from five distinct, pre-packaged design themes. With a single click, Gemini applies the chosen visual aesthetic across the entire application interface right out of the box.
4/22（水）Gemini練習会 🗒️1.8km×7本 🔧4'00/km～3'40/km 👟メ...
4/22（水）Gemini練習会 🗒️1.8km×7本 🔧4'00/km～3'40/km 👟メタスピードエッジ東京 設定通りペーサー完遂 無理ない強度で走ること そのためにもしっかりとペースメイクすることを意識して #ジェミラン #Gemini3 #まるお製作所RC https://t.co/FKFhB0X0Xx
Okay this one’s been living in my head for a while...
Okay this one’s been living in my head for a while… finally built it ❤️🔥 A fully vibe coded interactive planet explorer. No blender models, just pure geometry, pure math, using @threejs with frustum, culling Logic, raycasting, spring physics, and more. Built using Gemini 3.1 Pro in @GoogleAIStudio as always :D Live: https://t.co/OLlwYKRP9m
Unleash multimodal AI like never before 🔥 The n...
Unleash multimodal AI like never before 🔥 The new real-time vision agent built with Gemini 3.1 Flash Live API sees, hears, speaks, and creates music. Get the technical deep dive: https://t.co/I3V9331XxM https://t.co/ytcTfWTB9V
🚨 Google just launched Deep Research Max. It's a...
🚨 Google just launched Deep Research Max. It's a research agent built on Gemini 3.1 Pro with two things that matter: 1. MCP Support You can now connect Deep Research to private data sources through Model Context Protocol. It doesn't just search the web anymore. It can query your company's internal databases, financial terminals (FactSet, S&P, PitchBook), or any MCP server you point it at. 2. Native Chart Generation The agent generates HTML charts and infographics inline with the report. Not screenshots. Not suggestions to "visualize this data." Actual rendered charts inside the markdown output. - Two Modes: Deep Research: Fast. Lower latency. For interactive user-facing apps. - Deep Research Max: Extended test-time compute. Runs iterative reasoning loops before delivering final output. Designed for overnight batch jobs. Connect it to web search + MCP servers + file uploads simultaneously. Review and edit the research plan before it executes. Stream intermediate reasoning steps in real-time. Ground research in PDFs, CSVs, audio, video as input. Already Running in Production: Gemini App, NotebookLM, Google Search AI Mode, Google Finance. This isn't a beta. It's the same infrastructure powering consumer products at Google scale. Available today via Gemini API (Interactions API).
The biggest tech companies in the world are all do...
The biggest tech companies in the world are all doing the same thing quietly. building chips to ESCAPE Nvidia Google built TPUs. trained Gemini 3.0 on zero NVIDIA hardware Microsoft built Maia 200. launched jan 2026. already beating benchmarks Meta built MTIA. 4 generations dropping in 2 years via Broadcom OpenAI stopped being loyal. stacking 10GW of custom Broadcom silicon Elon's xAI building TERAFAB from scratch Then Amazon made it OFFICIAL. $25 BILLION into Anthropic. $100 BILLION back into AWS 5 GIGAWATTS of trainium for the next 10 YEARS The companies building the future are building it WITHOUT NVIDIA at the center.
Google released Design Previews for Build in Googl...
Google released Design Previews for Build in Google AI Studio, where Gemini would offer you the option to choose from one of the custom themes while your app is being built. Build and design in parallel 👀 https://t.co/M5FHgtgRp9