分类筛选
找到 437 个 Demo
フライングもちパーティおしるこ編 続編があるかは不明 #Geminiai #Gemini3
フライングもちパーティおしるこ編 続編があるかは不明 #Geminiai #Gemini3 https://t.co/m4LCSNBSIN
🌌 3D Particle Gesture Interaction｜Gemini 3 (Promp...
🌌 3D Particle Gesture Interaction｜Gemini 3 (Prompt Included) I experimented with gesture interaction using the recently trending Gemini 3, and the result turned out pretty well. #AIGC #Gemini3 #Threejs #CreativeCoding #3DParticles #InteractiveArt #HandTracking https://t.co/oJtzseVyj2
Google Labs has launched CC, a new experimental AI...
Google Labs has launched CC, a new experimental AI productivity agent built with Gemini, designed to help users stay organized and get things done. Once connected to your Gmail, Calendar, Drive, and the web, CC delivers a “Your Day Ahead” briefing every morning summarizing your schedule, tasks, and important updates. It also auto-drafts emails, adds calendar links, and responds to direct messages or custom instructions. Users can even teach CC to remember personal preferences, ideas, or reminders.
Can you guess which ai built each landing page out...
Can you guess which ai built each landing page out of these? Which one is best in your opinion? The choices are: Gemini 3 Pro Grok 4 Expert chat gpt 5.2 Claude Opus 4.5 https://t.co/d0rwHDKhwE
Vibe Coding Robotics Part 4 Implemented a more re...
Vibe Coding Robotics Part 4 Implemented a more reliable gripper that no longer glitches through objects when picking them up. Designed with Nano Banana Pro, built with Gemini 3. https://t.co/5ENHhC9zsK
ESP32 oscilloscope project built with the ESP-IDF...
ESP32 oscilloscope project built with the ESP-IDF framework and Gemini 3 LLM. https://t.co/BmnYLnviEl Developed in a few hours, the ESP-Scope project transforms ESP32 targets into web-based oscilloscopes using the ADC input from the chip. It was tested on a XIAO ESP32C6 board, and a 3D printed case enclosure was also designed for it.
🚨 Google Just Dropped Gemini 3 Flash. The Fastest...
🚨 Google Just Dropped Gemini 3 Flash. The Fastest, Most Practical Model Yet Gemini 3 Flash is a speed-optimized model in the Gemini 3 family built for real-world workflows where latency, reliability, and cost efficiency actually matter. Why it stands out: • Ultra-low latency near-instant responses • Best-in-class cost-to-performance ratio • Strong function calling & tool integration • Ideal for coding, automation, agents, and daily workflows • High-quality reasoning without the premium price tier This isn’t a benchmark flex. It’s a production-ready daily workhorse. Big credit to @GoogleDeepMind and @GeminiApp for pushing practical AI forward this is the kind of upgrade builders and creators actually need. 📌 Save this Gemini 3 Flash will quietly power a lot of workflows going into 2026. What’s one task you wish AI could handle instantly? Reply below 👇 #Gemini3 #GoogleAI #AI #Productivity #AIDevelopers
Just built a Red Flag Scanner in literally minutes...
Just built a Red Flag Scanner in literally minutes using OnSpace App goal? Scan text or images uploaded by users, giving red flags if it's Al-generated. >Whole task is running on Gemini 3.0 for the psychological analysis (and the replies are actually hilarious). >It also handles the App Store push and payments, too. >Feels like a cheat code for shipping apps. >Click my app link and remix if you want : https://t.co/HO0nOFosDk
So came with SERA Crypto, a new reasoning archite...
So @sentientagi came with SERA Crypto, a new reasoning architecture built only for real crypto research. Quick hits: • number one open source agent on DMind • beats GPT 5, Grok 4, Gemini 2.5 Pro and Perplexity Finance • all under 45 seconds end to end What makes SERA different: it routes by embeddings, reasons with an llm, then runs parallel data pulls, so every query feels like a full research cycle instead of a single model guess Core flow in simple points: 1. intent extraction 2. ticker resolution 3. embedding based routing 4. parallel API and deep search execution 5. optional loop for open ended tasks 6. final grounded synthesis Why people care: SERA sits near GPT 5 Medium on DMind while staying open source and fast enough for live DeFi, tokenomics, and security workloads sentient basically proved you can get state of the art crypto reasoning with open models that anyone can inspect, run, and build on top of. and they said they are open sourcing the full SERA stack next, benchmarks included.
Two models, same prompt, insane difference in the...
Two models, same prompt, insane difference in the result One of the models extremely nailed it 🤯 Its now my favourite model > Gemini 3 Pro vs Opus 4.5 > Same prompt > No reference > Built inside rork ai Guess which was gemini & which was opus https://t.co/f5L1RzUJtg
Gemini 3 Flash is now available via Firebase AI Lo...
Gemini 3 Flash is now available via Firebase AI Logic 🔥 Frontier intelligence built for speed at a fraction of the cost, Gemini 3 Flash: ⚡ Offers Gemini 3’s Pro-grade coding performance with low latency 🛠️ Is made for iterative development 💨 Enables more intelligent applications that demand both quick answers and deep reasoning Learn more → https://t.co/INSIFp4ycm
Gemini 3 可以直接引用Notebook，然后进行对话，这个功能太爽了。等于是Notebook...
Gemini 3 可以直接引用Notebook，然后进行对话，这个功能太爽了。等于是NotebookLM可以直接使用最强的Gemini 3 Pro模型了。那么至少对我来说，所有的RAG类型的需求，毫无疑问第一时间创建一个notebook再说。 #Gemini3 #NotebookLM https://t.co/vU5c0OTkON