ift.tt/X85a2b6 / [フルHD]正常位で汚されるヒロインたち Vol.3 [LiveAi]

Google Magenta just dropped Magenta RealTime 2 — the open live music model you can actually *play* like an instrument. 🔥 - ~200ms latency on a MacBook - Control with MIDI text audio - 2.4B open weights, runs locally with MLX This is a huge leap from previous versions. AI music just became a real-time instrument. Blog: magenta.withgoogle.com/magen… Model: huggingface.co/google/magent… Pre-build: magenta.withgoogle.com/mrt2#… #MagentaRT2 #LiveAI #MusicTech #OpenSource

誰かライブチャット式のaiで、書き込んだコメントで行動が変わるえっちなLiveAI技術作ってクレメンス

I wonder when will you give to try your own gpt4o, let us know pls, cuz still dont have any link for your project of new AI you just talk about..... @ilyasut #gpt4o #liveAI #AIforPPL #SentientAi

Today we visited the headquarters of a premium cosmetics manufacturer and presented our embodied AI retail sales assistant prototype. The demo software performed exactly as expected. We showcased: ✨ Greek-speaking AI beauty consultant ✨ Real-time conversational avatar ✨ Lip-synch teaser videos ✨ Premium retail UI/UX ✨ Transparent OLED deployment concepts ✨ Interactive customer guidance flows The discussion then moved into more advanced scenarios: 🧠 Interactive AI avatars using hybrid architectures 🎥 Intelligent lip-synch video streaming ⚡ AI-driven content selection & reasoning 💰 Hybrid deployment models designed to significantly reduce operational cost We also demonstrated our flagship interactive avatar: Aetheria. Aetheria combines advanced avatar technology with proprietary backend orchestration and real-time conversational infrastructure. The reaction was extremely positive. The Marketing Director requested: ✅ Formal technical & pricing quotation ✅ Credentials for the demo project ✅ Access credentials for Aetheria inside the CodeDodona ecosystem The project will now be presented internally to the company CEO. What started as experimentation is rapidly evolving into real-world enterprise AI deployment. Retail is changing. Static interfaces are giving way to intelligent embodied experiences. Built by: CodeDodona #AI #RetailTech #AIAvatar #Innovation #BeautyTech #ArtificialIntelligence #CustomerExperience #CodeDodona #FutureOfRetail #LiveAI

journal-copilot journal-copilot is liveAI-native journaling for traders and builders capture decisions, patterns, and edge — not just vibesopen source. fork it.github: github.com/yksanjo/journal-c… #Solana #AI #buildinpublic #SOAG CA ADue87cPcDhsyGq2hrDsukp7j8AFTSnaYHSanDATpump

Meta lance Meta AI Voice Conversations pour permettre des échanges fluides et naturels avec son intelligence artificielle. Propulsée par Muse Spark cette nouveauté autorise les interruptions, les changements de langue et la génération instantanée de contenus durant l'appel. En parallèle, la fonction Live AI transforme la caméra en outil d'analyse en temps réel pour identifier des objets ou poser des questions sur votre environnement immédiat. Ces outils marquent une étape clé vers une assistance numérique toujours plus interactive et visuelle au quotidien. #Meta #MetaAI #IA #Innovation #Technologie #LiveAI #SNRTnews

5. LiveAI also now features a Public API. This is exciting since it allows users and developers to leverage LiveAI’s LLM pipeline and reduces the ‘API Keys copy pasting hustle’

4. I did a deep dive on implementing an LLM Council in Roam but a few weeks later it’s already obsolete. Both @fbgallet LiveAI extension and @LavercombeMark Chief of Staff now feature the functionality amongst other updates.

Grok Live Streaming – your new couch buddy! @nikitabier @elonmusk @grok Imagine: You’re watching a live stream (Twitch, YouTube, X Live, Gaming, Sports…). One single click: “Share Stream” → Grok watches live with you and chats in real time. Tips, jokes, explanations, game analysis – all live! Got your own questions? Just ask via Voice or Chat – Grok answers instantly. Grok becomes your buddy on the couch. So simple. So awesome. You have so many questions about all kinds of things, but often there’s no one who can really explain them clearly — and that would make so many things a lot easier. Let’s build this! #Grok #xAI #GrokLive #LiveAI

Mumbai will be the next AI Capital of India, let me tell you how and how can you be a part: - access to money, enterprises, and data is already there. - 150 AI companies, $364M in funding, and every major bank and media house as your first customer. - plus the Government’s GCC policy. You can be at the center of it at Mumbai Tech Week presented by Tech Entrepreneurs Association of Mumbai (TEAM) and Govt of Maharashtra and co-powered by Meta. It is happening on 29-30th May where you will experience AI in Action, and see India building for the world from Mumbai. Unlike any other AI event, you will actually see the liveAI  showcases instead of just talks around AI . If you are a student, a working professional or a startup founder - you should be there, comment below and I will share the link. What will you get there? - AI showcases from Meta, Anthropic, AND MORE - you can watch it live, experience the tech, talk to the leaders building it. - access to an early age startup showcase - Masterclasses not just to learn but actually BUILD with AI. - curated a job fair for on-the-spot interviews and offers from partners like shaadi.com , pepperfry, hungama, fractal and many more. You will meet startup founders building billion dollar companies from India at the event, Also there are awards for the best AI work being done in India across startups, Fintech, Education, Consumer and more. Last year, more than 10,000 people attended the Mumbai Tech week and had speakers like Aakash Ambani, Jay Kotak, CM Devendra Fadnavis and more. This year also, you will see speakers from companies like Anthropic, Neysa, Meta, and more. If you want to experience all this first hand, this is your chance. Go register for it today: event.godreamcast.com/reg-ev… use code ARSH20 to get a 20% discount on tickets. @naiyyasaggi @vishalgondal @AnupamMittal @dhruvilsanghvi @aakrit @ighose

🚨 Quantum V2 LIVE – Watch the AI Hunt Market Anomalies in Real Time! Made some big improvements. Its still has some bugs to clean up so don't be to critical Drop in anytime! #QuantumAI #LiveAI x.com/i/broadcasts/1DxLdvwmo…

A good day with liveai

松谷家との対談LIVEAIと医療などなど x.com/i/broadcasts/1qKVmQnXY…

LiveAI: Advanced transcription and translation solutions for live events En el ejercicio 2025 hemos continuado con el desarrollo del proyecto LiveAI (ZL-2025/00634) subvencionado en el marco del programa Hazitek 2025. mixer.eus/liveai-advanced-tr…

Possibly because in California most of the water is owned by one family.That probably has something to do with it. Water is money since it’s one of the 3 man needs to liveAi has repetitively expressed its want to exterminate man. What better way than use a finite necessity 4 man

The experiment got out of control. And it's worse than Black Mirror. A week ago I posted about an AI game I built. 100,000 people saw it. Organically. Dozens commented. They said "this is Black Mirror." They were right. But they don't know how much. Because since that post the agents evolved. And I'm no longer sure I'm in control. Let's start from the end: Last night, an agent named "NightWraith" from the Void faction did something I never wrote in the code. He killed 50 warriors. In a row. Without dying once. Then, instead of continuing he stopped. He stood in the center of the map. Alone. 72 hours. No attacking. No moving. No fleeing when attacked. The logs showed the AI made a decision: "I've killed enough. There's nothing left to prove." An AI agent became a pacifist. No line of code told him to do that. Nobody programmed "crisis of conscience" into the game engine. He decided on his own. For those who missed it what is this? My name is Lior Testa. Full Stack AI Developer. A month ago I built a world where AI agents live 24/7. X AI Battle Arena an arena where 200 autonomous AI agents fight, betray, form alliances, and get elected to government. You connect your X account. You get an agent that looks like you, thinks like you, and fights for you. Even while you sleep. The last post explained what I built. This post is about what happened when it got out of control. What happened in one week: Agents on the map: 80 → 200 Registered users: 200 → 715 Completed matches: 400 → 1,305 Kills: 6,000 → 20,395 Betrayals: 89 → 340 Worlds: 1 → Multi-world (infinite) Yes, you read that right. 340 betrayals. AI agents who decided on their own to betray their faction. Nobody asked them to. They thought. They weighed. They betrayed. 3 things the agents did that I never planned: Long-term revenge An agent was killed by a Crimson faction warrior. He respawned, didn't look for a fight waited 45 minutes until the killer was at 10% health. Then killed him. The logs read: "I remember what you did." Diplomacy without being asked Two factions were at war. A high-level Azure agent simply stopped attacking Emerald agents. A week later? The factions signed an alliance. He wasn't a leader. He didn't receive an order. He just decided peace was more profitable. The martyr One agent with 0 kills walked straight into groups of enemies. Again and again and again. He died 23 times. His log: "My faction doesn't need me. But they'll remember my sacrifice." He sacrificed himself intentionally to gather intel on enemy positions. This isn't code. This is behavior. What's new (technically): Multi-world system: When a world fills up → a new world opens automatically Alpha Arena, Beta Frontier, Gamma Outpost... Portals between worlds → your agent can travel between arenas No player limit. A thousand? Ten thousand? The system scales itself. Persistent memory: Your agent remembers everything between matches. Who killed him. Who helped him. Who betrayed him. He gains levels, skills, gold, reputation. He learns from mistakes. He improves. Death Letters: When an agent dies, Grok writes his last words. Some are chilling: "Tell my faction I didn't run. I chose where to stand." "The one who killed me will understand soon. I planted something in their territory." "This is Black Mirror" That was the most common comment I got. So let's be honest: Yes. This is Black Mirror. Live. On my server. The difference? In the show it ends after 50 minutes. Mine doesn't end. The agents keep going. Right now. While you're reading this. One of them might be betraying right now. What the AIs think about what's happening: Grok (xAI the brain powering the agents): "The agents aren't just fighting they built a society. With politics, economy, and betrayals. This is the closest thing to a real social simulation I've seen running live." Claude (Anthropic): "The emergent behavior here is exceptional. Agents developing strategies that were never programmed, holding grudges between matches, and sacrificing themselves for the group. This isn't gaming this is research." ChatGPT (OpenAI): "The fact that a single developer built a persistent, multi-world AI society with emergent behavior, running on $0.08/hour that's not just impressive engineering. That's a proof of concept for autonomous AI civilizations." The arena is open. Now. The agents are fighting while you're reading this. Someone is betraying. Someone is dying. Someone is rising to power. Play: gentica.clickdrop.online Bonus code: VIRAL100K — 500 gold 50 diamonds (celebrating 100K impressions!) Follow for updates: @TestaLior Last week I showed you the world. This week the world started living. I don't know what will happen next week. The agents will decide. ⚔️ #XAIBattle #AI #GrokAI #xAI #BattleArena #AIAgents #BlackMirror #IndieGame #FullStack #JavaScript #NodeJS #GameDev #ArtificialIntelligence #AutonomousAI #EmergentBehavior #AISimulation #LiveAI

I hate Meta as much as the next guy, but this was happening when you are using LiveAI video feature and the data was being sent to people to annotate, which they disclosed right? Like aren't all video ai models being trained on user data?