分类筛选
找到 1112 个 Demo
Just saw the Alethia numbers drop. Google DeepMind...
Just saw the Alethia numbers drop. Google DeepMind's internal beast (built on Gemini 3 Deep Think) is posting 91.9% on Advanced ProofBench, 92.1% novel problems, 100% on hardened IMO 2024+, and 83.3% USAMO 2025. 😳 That's not incremental. That's a canyon. This isn't just better at Olympiads but It's doing PhD-level verification loops: Generator → Verifier (natural lang proof checking!) → Reviser, iterating until it's bulletproof. We went from "AI wins IMO gold" in mid-2025 to "AI is now a research collaborator publishing breakthroughs" in under a year. The verifier-reviser cycle is the key because it mimics actual rigorous human scrutiny instead of probabilistic guessing. Implications? - Human mathematicians just got the ultimate force multiplier to offload proof grinding/verification, and keep the creative leaps for humans. Alethia isn't a model. It's a system showing what inference-time scaling + agentic rigor actually unlocks and more fast tokens = more discoveries. Trying to wrap my head around this.
#Geminiai #Gemini3 ハッピーバレンタイン！その３ 美月の身長が高いように見えます...
#Geminiai #Gemini3 ハッピーバレンタイン！その３ 美月の身長が高いように見えますが、これが本来の身長です。 単独で生成した際の背が低いのは 私 の 好 み で す 。 https://t.co/RH11RJJz2M
某ギャルゲー 理知的な眼鏡の優等生を実写風に再現 ハンカチを差し出し、心配が静かに滲む 的なイメー...
某ギャルゲー 理知的な眼鏡の優等生を実写風に再現 ハンカチを差し出し、心配が静かに滲む 的なイメージ #Gemini3 #NanoBananaPro https://t.co/SwVGLBYX4X
I built my Anti-Gravity: Private AI PDF app using...
I built my Anti-Gravity: Private AI PDF app using this tech stack DAY 32nd of building Anti-Gravity: Private AI PDF Here's what actually powers Anti-Gravity: Building - Claude + Gemini Code - reactjs + vitejs Mobile - capacitorjs AI - Google AI (Gemini 3.0 Flash) Hosting - vercel Auth - supabase + google devs OAuth Database - supabase (PostgreSQL) Payments - Google Play Billing PDF Engine - pdf-lib + pdfjs-dist Security - Play Integrity + HMAC Styling - tailwind css If you're building apps in 2026: What's in your stack?
#Geminiai #Gemini3 ハッピーバレンタイン！その２ ゲームスタート！
#Geminiai #Gemini3 ハッピーバレンタイン！その２ ゲームスタート！ https://t.co/DvzkG183ww
#Geminiai #Gemini3 カピバラはクールに去るぜ
#Geminiai #Gemini3 カピバラはクールに去るぜ https://t.co/KQ9Y5HfT9M
#Geminiai #Gemini3 ハッピーバレンタイン！
#Geminiai #Gemini3 ハッピーバレンタイン！ https://t.co/6Fig4ygNWz
MiniMax M2.5 is out. Already running Minimax in m...
MiniMax M2.5 is out. Already running Minimax in my OpenClaw agent setup. This update hits different: • Beats GPT-5.2 and Gemini 3 Pro on coding • $1/hour to run continuously • Open source • Built natively for agent scenarios The AI race is moving fast. https://t.co/62uXYqdncp
We built a Gemini API skill for devs. If you're b...
We built a Gemini API skill for devs. If you're building with Gemini, make sure your coding agent uses the latest SDKs/models/best practices and doesn't rely on it's knowledge cutoff. Deets on the hub: https://t.co/UtDRDhhIm0 ⬇️Or install it with @vercel or @Context7AI ⬇️ https://t.co/HR8U5TRijD
How my Cowork agent built by GPT-5.3-Codex is goin...
How my Cowork agent built by GPT-5.3-Codex is going so far UI and how the agent works/responds needs work B U T it's going well! Gemini 3 Flash also needs DCF help https://t.co/q4TJEVZVRD
Google Gemini is the smartest AI right now. But 9...
Google Gemini is the smartest AI right now. But 90% of people prompt it like ChatGPT. That's why I made the Gemini Mastery Guide: → How Gemini thinks differently → Prompts built for Gemini → 2000+ AI Prompts Comment "Gemini" and Follow me so I can DM You Free. @codi_fyy https://t.co/Xor75ybUoh
Google Gemini is the smartest AI right now. But 9...
Google Gemini is the smartest AI right now. But 90% of people prompt it like ChatGPT. That's why I made the Gemini Mastery Guide: → How Gemini thinks differently → Prompts built for Gemini → 2000+ AI Prompts Please do Like Comment "Gemini" And Follow me I'll DM it free. https://t.co/hvCiGPhHg1