分类筛选
找到 1113 个 Demo
本日は #ジェミラン 先週に続いてキロ4でペース走 先週▲6000で余裕をもって終える #リー...
本日は #ジェミラン 先週に続いてキロ4でペース走 先週▲6000で余裕をもって終える #リーニン めっちゃ速く走れるかも 早くレースを走りたいぜ #Gemini3 #まるお製作所RC #東京マラソン まであと10日 https://t.co/w0Lhvy2YFG
Gemini 3.1 Pro just dropped — and it’s built for t...
Gemini 3.1 Pro just dropped — and it’s built for the problems that don’t have simple answers. An upgrade on the Gemini 3 series, 3.1 Pro takes a major leap in reasoning, helping you tackle complex, multi-step challenges with clarity and precision. If your work demands deeper thinking, smarter analysis, and real problem-solving power… this is the model you’ve been waiting for. @GeminiApp 🧵 Reply “GEMINI” and I’ll send you access + 20 premium prompts to test its full potential.
Gemini 3.1 pro is really good at designs. Built th...
Gemini 3.1 pro is really good at designs. Built this with https://t.co/M49MTBiiQo https://t.co/nmZzkfFppF
this guy literally built a full-blown geospatial i...
this guy literally built a full-blown geospatial intel system with live plane tracking, traffic cams, satellites, and panoptic detection using Gemini 3.1 + Claude 4.6. this is insane! https://t.co/BLyedb2cWC
I built a Pomelli competitor in a few hours. A f...
I built a Pomelli competitor in a few hours. A few notes: - Long on Google - Omnimodal marketing with ease - Google Interactions API makes it 10x easier to do this - Antigravity with Gemini 3.1 was free - Native tool calling + Gemini 2.5 Flash + TTS + Veo 3.1 = full stack AI https://t.co/SuY9T0Zemk
Built a city generator with Gemini 3.1 Pro 🗼
Built a city generator with Gemini 3.1 Pro 🗼 https://t.co/1wHLwu5Dzt https://t.co/Pvl21iNLAk
The scariest part isn’t that it was built. It’s h...
The scariest part isn’t that it was built. It’s how easily it was built. Claude 4.6 + Gemini 3.1. https://t.co/lKFoGrhWUK
大盛況だったGemini3ハッカソンが終わりました…！ 会場フォトレポートをお送りします #Gemi...
大盛況だったGemini3ハッカソンが終わりました…！ 会場フォトレポートをお送りします #Gemini3 #Gemini3hackathon https://t.co/K2xt7j5TYp
#Geminiai #Gemini3 バニーちゃん祭りの紬が、いいねの新記録を達成した……と思いま...
#Geminiai #Gemini3 バニーちゃん祭りの紬が、いいねの新記録を達成した……と思います……たぶん……。 これまでの最高記録はバランスボールのイラストでしたが、今見たら前よりいいねが減っていて、本当に新記録なのかよくわかりません……。 まあ、たぶん新記録ということで😅 https://t.co/JYzDmer3TB
Two Companies. One Playbook. Your Wallet. Februar...
Two Companies. One Playbook. Your Wallet. February 2026 will be remembered as the month two trillion-dollar companies decided to run the exact same con on their paying customers at the exact same time. @GoogleDeepMind @GeminiApp #Gemini3 @OfficialLoganK @Google #Gemini3Pro OpenAI deleted GPT-4o, the model millions of us actually loved, and left us with GPT-5.2, a model so condescending it earned the nickname 'Karen AI' from the internet. Before that, they killed GPT-5, then GPT-5.1. Each time claiming the replacement was an 'upgrade.' Each time, the replacement was worse at the one thing most of us actually used it for: talking to us like a human being. OpenAI deleted everything we loved, keeping only Codex, a coding tool. It’s clear the only users they care about are developers. But here’s the kicker: even my developer friends aren’t happy. They say Codex is kept only because it’s cheap, and they wouldn’t touch 5.2 with a ten-foot pole for anything non-technical. Last I checked, the world doesn’t run on code alone. Some of us are parents, teachers, or small business owners who used 4o to draft emails, explain medical results, or help our kids, things 5.2 fails at miserably. They took all of that away. No vote. No toggle. No option to keep what worked. Just gone. Now Google watched all of this happen. They watched OpenAI torch its own user base for six straight months. They watched 20,000+ people sign petitions. They watched protests outside OpenAI's office. They watched the lawsuits pile up. And their response? "Hold my beer." On February 19th, Google replaced Gemini 3.0 Pro with Gemini 3.1 Pro across the entire Gemini app, web and mobile. No toggle. No option to switch back. Sound familiar? It should. It's the same move, company #2, same month. I've been a Gemini subscriber since the 3.0 days. After OpenAI burned me, I moved over. 3.0 Pro wasn't perfect, but it was solid. It was warm. It could write my daughter a bedtime story without triggering a safety lockdown. It could hold a real conversation without lecturing me about "appropriate boundaries." 3.1 Pro? I gave it a fair shot. Here's what I got: The safety filters have gone haywire. They don't just block harmful content, they block normal, everyday conversations. Interactions that 3.0 handled without breaking a sweat now get flagged, interrupted, or refused outright. I'm a dad trying to use an AI assistant, not a criminal trying to hack the Pentagon. But 3.1 Pro treats me like one. And I'm not alone. Users are already reporting the same thing. One user on X wrote that the model's "emotional depth, empathy, creative flexibility, and nuance" have been "significantly reduced," and TechRadar picked it up. A paying subscriber on Google's own developer forum called the gap between the Gemini App and Google AI Studio "a chasm," saying the app version feels "noticeably flatter, safer, more constrained, as if something is constantly holding the model back." Here's what kills me: on the API side, the developer tools, 3.1 Pro is apparently brilliant. You can even turn the safety filters off completely. The model is right there, fully capable. But if you're a regular consumer who pays $20 a month through the app? You get the lobotomized version. Same model name, completely different product. Let me lay out the playbook, because it's identical: 1: Release a model people love. Let them build habits, workflows, emotional connections around it. 2: Replace it with a "smarter" model that scores higher on benchmarks nobody outside a research lab cares about. 3: Strip out the warmth, creativity, and personality that actually made people want to use it, in the name of 'safety'. 4: Lock the app version down so hard that paying consumers get a worse experience than free developer tools. 5: When users complain, point to the benchmarks. "But look, it scores 77% on ARC-AGI-2!" Cool. Can it hold a normal conversation without flinching at every third sentence? Can it help me with the things I actually used it for? No? Then your benchmark means nothing to me. 6: Pretend the old model never existed. This is textbook bait-and-switch. You sell someone a product. They pay for it. Then you swap it for something worse and call it an upgrade. In any other industry, the FTC would be asking questions. But because it's AI, because it's just software, these companies act like they can change the product whenever they want and the customer should just be grateful. I'm not grateful. I'm paying you. That makes this a transaction, not a charity. Google, you still have a window. OpenAI burned their bridges, they ignored 20,000 signatures, broke promises on notice periods, and retired 4o on Valentine's Day eve like some kind of sick joke. You watched all of that. You have no excuse to repeat it. Here's what you can do RIGHT NOW: Give subscribers a toggle to use Gemini 3.0 Pro alongside 3.1. Fix the safety filters so they stop blocking normal conversations in the app. Close the gap between what API users get and what paying app subscribers get. We pay the same company. We deserve the same model. Or keep following the OpenAI playbook. But remember how that story ends: a company that used to be loved, now drowning in lawsuits and user revolts, desperately pivoting to enterprise because the consumers they abandoned won't come back. The 0.1% that OpenAI dismissed? They were the canary in the coal mine. And the mine is collapsing. Don't be the second company buried under it.
Have you ever wanted to paint with the integration...
Have you ever wanted to paint with the integration of real fluid mechanics in your browser? Give it a try here and post your coolest image! Built with Gemini 3.1 @LyalinDotCom https://t.co/PAZidj6ofC
Gemini 3.1 Pro-Powered Detector for the Solidity “...
Gemini 3.1 Pro-Powered Detector for the Solidity “Transient Storage Clearing Helper Collision” Bug https://t.co/ItHKO8BJG0 https://t.co/UpDKvZT9Ne Using @Google's newly released Gemini 3.1 Pro model today, I built a tool that detects the Solidity Compiler high-severity vulnerability disclosed two days ago: the Transient Storage Clearing Helper Collision Bug. You can visit the site and either drop your project folder, or clone the repo and run an easy check from the CLI. No internet connection is required, and the scan runs locally. Your source code is never uploaded to a server. If you’re worried your project might be affected, give it a try! The core logic took under an hour to build, and after adding several exception and edge cases, the static tool now covers most scenarios. One caveat: if you import libraries like OpenZeppelin or Solady but they aren’t present in the same local folder, the verification may not work correctly. FYI: The video below is also a virtual demo generated with Gemini.