Google's FREE Gemini 3.1 Pro Feels Illegal To Use (+ 7 AI Updates)

20:43

Google's FREE Gemini 3.1 Pro Feels Illegal To Use (+ 7 AI Updates)

Vaibhav Sisinty 25.02.2026 66 547 просмотров 1 884 лайков

Machine-readable: Markdown · JSON API · Site index

Смотреть на YouTube

Поделиться Telegram VK Бот

Транскрипт Скачать .md

Анализ с AI

Описание видео

💡 Invest in EnergyX → https://bit.ly/energy-x-vs 🔗 Join our WhatsApp Community Get the latest AI updates, tips, and insights straight to your inbox: https://join.switchit.app/YT1 Google's Gemini 3.1 Pro just dropped — and it's FREE. I break down the 3 features you need to try right now, plus 10 major AI updates from this week including Grok 4.20, Anthropic's $30B raise, and Meta's creepy new patent. 🔥 Try Gemini 3.1 Pro FREE → https://gemini.google.com 🔥 Google AI Studio (FREE) → https://aistudio.google.com 🔥 Antigravity (FREE download) → https://antigravity.google/download 🎬 WATCH NEXT: - India AI Summit Full Coverage → https://youtu.be/FSTpvX3IoHQ?si=bSM_pnNfxcZnAbYE 📌 WHAT'S COVERED: - Gemini 3.1 Pro — 2x reasoning power, #1 ranked AI model, free for everyone - Lyria 3 — Full AI music studio inside Gemini (vocals, lyrics, cover art) - Grok 4.20 — 4 AI agents working together, Heavy uses 16 agents ($300/mo) - Anthropic raised $30B at $380B valuation — second-biggest private round ever - Claude used in Pentagon operation, $200M deal now at risk - OpenClaw developer hired by OpenAI after Meta bidding war - Meta patented AI that posts, comments, and video calls AS you after you die - Claude Sonnet 4.6 — handles full to-do lists autonomously - Gemini tutorial: image → video → music pipeline, SVG animations, Antigravity autonomous coding agent 🔗 TOOLS MENTIONED: - Figr AI → https://figr.design/ - EditWithAva → https://editwithava.com/ - Seedance 2.0 → https://seed.bytedance.com/en/seedance2_0 - Qwen 3.5 → https://github.com/QwenLM/Qwen3 - Lovon → https://lovon.app/ - Pomelli Photoshoot → https://labs.google.com/pomelli/ ⏰TIMESTAMPS: 0:00 – Intro: Why Gemini 3.1 Pro is a game-changer 1:10 – What’s new in Gemini 3.1 Pro (features overview) 3:00 – Gemini vs other paid AI tools (performance breakdown) 5:20 – Live demo: using Gemini 3.1 Pro for real tasks 8:05 – Lyria 3: full AI music studio inside Gemini 10:10 – Grok 4.20 and multi-agent AI (4 to 16 agents) 12:30 – Anthropic’s $30B raise and Claude Pentagon deal 14:15 – Meta’s creepy “post-after-you-die” patent explained 15:40 – Claude Sonnet 4.6 and autonomous to‑do execution 17:05 – Gemini image → video → music workflow + Antigravity 19:00 – Best AI tools to try this week (quick rundown) 20:00 – Final thoughts and what to expect next This video is not financial advice. EnergyX segment is a paid partnership. See offering statement at invest.energyx.com for risks and disclosures. -------- To Know More, Follow Vaibhav Sisinty On ⤵︎ Instagram @VaibhavSisinty https://www.instagram.com/vaibhavsisinty Twitter @VaibhavSisinty https://twitter.com/VaibhavSisinty Facebook @VaibhavSisinty https://www.facebook.com/vaibhavsisinty/ LinkedIn - Vaibhav Sisinty https://www.linkedin.com/in/vaibhavsisinty

Оглавление (12 сегментов)

Intro: Why Gemini 3.1 Pro is a game-changer

This week, Google dropped Gemini 3. 1 Pro. And honestly, this one's different. Image generation, video with sound, AI music, and coding, all inside one model. Anthropic raised $30 billion. Then the Pentagon threatened to blacklist them. The safety lead quit and left AI entirely. While all this is happening, Claude just shipped an update that takes your to-do list and does it for you. Mus said coding is dead by December. backed it up with a model that throws four agents at one question. Meta patented an AI that keeps posting on your account after you die. Welcome back to AI updates of the week. So, I was just at India's global AI summit, sat down with Alexander Wong, Meta's AI chief, the guy Zuckerberg brought in for $14 billion, and everyone behind closed doors was saying the same thing. The gap is closing fast. This week is proof. After the updates, we're covering the top three use cases of Gemini 3. 1 Pro. These are the reason this model is going viral right now. Two of them are completely free. And the third one, let's just say the AI takes over your screen and does the work for you. Stay for that. Let's get into it.

What’s new in Gemini 3.1 Pro (features overview)

If keeping up with all of this feels like a lot, I filter the noise and send only what matters straight to your phone on WhatsApp. Links in the description. Google dropped two big ones this week. First, Gemini just got a major brain upgrade. It's called Gemini 3. 1 Pro, and the reasoning on this thing is twice as powerful as what we had 3 months ago. It thinks harder. It handles complexity better, and it's live right now inside the Gemini app. I'm doing a full deep dive tutorial on this one. So, make sure you're subscribed so you don't miss that. But the second thing, completely different. Google just put a music studio inside the Gemini app. It's called LIIA 3 and it works like this. You type in what you want and in a few seconds you get a full track, vocals, lyrics, beat, everything. You can even drag in a photo from your camera roll and it'll make a song to match the vibe. In the demo, someone made a punk song telling their roommate to do the dishes. A party photo turned into a dance club anthem. Each one came with custom cover art and genre tags. Right now, it makes 30-second clips and it's in beta. But here's what makes this a big deal. Gemini has 750 million monthly users. That's 750 million people who just got a music creation tool built into the app they already use. Sunno raised $250 million to build exactly this. But Google just bundled it into Gemini for free. On different note, I have a personal story I'd like to share. Every AI company nowadays claims their model is multilingual. But a team from IIT Madras and Josh talks actually tested it. 15 Indian languages, 35,000 real speakers, the real India. The results were unbelievable. An Indian company called Sarvam achieved 87% accuracy

Gemini vs other paid AI tools (performance breakdown)

significantly outperforming Open Eyes 36%. The best part is I gave Openy a simple Hindi sentence and it actually hallucinated Hungarian. Sarvam on the other hand flawless. Microsoft doesn't even support half the languages tested and in India voice is how people access banking, healthcare, everything. Turns out local context beats global scale and that ties perfectly into what India just did on the world stage. India just hosted its first ever global AI summit. Modi Macron, Sam Alman, Sundar Pichai all in one room. Amani pledged 10 lakh cr rupees to build India's own AI infrastructure. Indian startups showed off AI that works in 22 languages, detects TB from a cough and helps farmers through WhatsApp. I covered all five days in a separate video links in the description. Go watch that once you're done with this one because right now Elon Musk just made a claim that changes everything. Everyone's arguing about which AI writes the best code. Elon Musk just said none of that matters. At a recent XAI event, Musk said AI will skip coding entirely by the end of this year. — Even year to where you don't even bother do doing coding, the AI just creates the binary directly. Um, and the AI can create a much more efficient binary than uh can be done by any compiler. Um, and uh we do expect uh Grock code to be state-of-the-art in 2 to 3 months. So, it's happening very quickly. not writing better code just skip the whole step. The AI creates the final program directly. No coding language needed in between. He also said Grock's coding model will be the best in the world within 2 to 3 months. Think about what that means. Every company right now is building tools to help you code faster. Musk is saying the entire step is about to disappear. And he's backing it up. Grock 4. 20 went live in public beta this week after months of delays. And what makes it different is that instead of one AI answering your question, four AI agents work on it together at the same time. Each one with a different job while a leader agent puts the final answer together. And Musk took that same idea and cranked it up. The heavy version uses 16 agents instead of four, all working on a single question at once. That's the $300 a

Live demo: using Gemini 3.1 Pro for real tasks

month plan built for researchers and businesses who need maximum depth. Early tests already beat GPT5 and Claude Opus 4. 5 on forecasting. In a livestock trading competition, Grock 4. 20 was the only AI that actually made money. Whether Musk's timeline on killing code holds up, that's a different question, but the direction is clear. The skill isn't writing code anymore. It's knowing what to build. Here's something nobody in AI wants to talk about. Every model we just covered, Grock, Gemini, Claude, they all run on data centers that eat electricity like nothing we've ever seen. The Department of Energy says data center power demand could triple by 2028. Google alone has over 100 million lithium ion battery cells, keeping their data centers alive when the grid can't keep up. And the grid, 70% of it was built before most of us were born. It literally cannot handle what AI needs right now. So, here's the bottleneck everyone's missing. All of this, the AI boom, the EVs, the energy storage, it all comes back to one material, lithium. Morgan Stanley projects an 80,000 ton shortfall this year alone, and demand grows five times by 2030. That's where today's sponsor, Energy X, comes in. Their patented tech recovers three times more lithium than traditional methods. General Motors led their $50 million investment round and the US, the Department of Energy awarded them a $5 million grant. They've got rights to 150,000 acres of lithium rich land across North and South America, and they're gearing up for commercial production right as demand explodes. This is the future being built. Do your research, and if it aligns with you, invest in Energy X and be part of the energy revolution. All right, back to the AI news. Anthropic just raised $30 billion five months ago worth 183 [clears throat] billion now 380 billion second biggest private funding round in tech history after open AAI revenues at 14 billion a year and Claude Code alone is pulling in 2. 5 billion Elon Musk called Anthropic misanthropic and evil worth noting Musk runs a competing AI company and Anthropic blocked X AI from using Claude after catching them building competing products with it. So while Musk is tweeting, Goldman Sachs is actually putting Claude to work. For 6 months, Goldman has had anthropic engineers inside their offices, building AI that checks trades and onboards new clients. They brought Claude in for coding. Turns out it could handle compliance, too. When the biggest investment bank on Wall Street embeds your team for 6 months, they're not

Lyria 3: full AI music studio inside Gemini

testing. They're building their future around it. Now, here's where it gets messy. The Wall Street Journal reported that Claude was used during the military operation to capture Venezuela's Maduro through Anthropics partner Palanteer after the operation friction emerged with the Pentagon over how claude could be used. Now, a $200 million deal is on the line and a Pentagon official threatened to label Anthropic a security risk. Meanwhile, OpenAI, Google, and XAI signed military deals with way fewer restrictions. The company that questioned how its tech was being used is the one facing consequences. And while this is happening, Marinang Sharma, who led anthropic safeguards research against bioweapon misuse, quit last week, posted a public letter saying he'd repeatedly seen how hard it is to truly let our values govern our actions. He's not joining another AI lab. He's leaving the field entirely. When the person keeping AI safe walks away from AI, that tells you something. But through all of this, the 30 billion, the Pentagon standoff, the safety team losing people, Anthropic still shipped a product. Claude Sonnet 4. 6 came out this week. Full upgrade across coding, computer use, long documents, and task planning. If you're on the free or pro plan, this is now your default model. Here's what got me. In the demo, someone gives it a regular to-do list. Change my store's shipping price. Swap colors in my email template. check if my website works on phones, file my travel expenses, move a delivery to Saturday, and it does all of it from opening the store settings to rescheduling a package all by itself, bouncing between websites and tools like a person would. See the difference? This isn't ask me a question and I'll answer. This is hand me your to-do list and go grab a coffee. It can hold enough information to fit an entire codebase in one request. Same price as the old version. and developers actually preferred this cheaper model over the previous top tier one 159% of the time. That's the game now. A developer in Austria built an AI agent as a weekend

Grok 4.20 and multi-agent AI (4 to 16 agents)

project. It went so viral that Meta and OpenAI got into a bidding war over him. Spoiler, he picked OpenAI. His name is Peter Steinberger and the project is called Open Claw. What made it blow up is that most AI tools just talk to you while OpenClaw actually does stuff. It clears your inbox, books flights, makes restaurant reservations. You text it through WhatsApp or signal like it's your personal assistant, and it just handles it. He flew to San Francisco, met with all the major labs, and chose OpenAI. His words, I could see OpenClaw becoming a huge company, but that's not exciting for me. I want to change the world. Sam Alman called him a genius and said he'll drive the next generation of personal agents. OpenClaw itself moves to an independent open-source foundation. So, it's not going anywhere. Now, here's a fun detail. Open Claw wasn't always called Open Claw. It started as Claudebot and Anthropic sent a legal threat because it sounded too much like Claude. So, he renamed it twice and before landing on OpenClaw, he actually asked Sam Alman if the name was cool. Altman said yes. The guy Anthropic pushed away ended up at the biggest competitor. Think of this like the AI race shifting gears. We've been in the chatbot era. AI that answers questions. This hire is open AI betting that the next era is agents. AI that actually takes action for you. And they just grabbed the person who proved people want that. Right. Move by OpenAI. Drop your take in the comments. Meta just patented an AI that keeps your social media account active after you die. Not a memorial page. Actually, posting, commenting, responding to DMs. It's trained on everything you've ever done on their platforms. Meta says they have no plans to build it, but their CTO is listed as the primary inventor. And Zuckerberg told Lex Friedman back in 2023 that Meta will eventually have the capacity to create AI replicas of people. So, no plan sounds a lot more like not yet. I made this image, turned it into a video with sound, generated a soundtrack, all with one AI model. That's Gemini 3. 1 Pro. I'm going to show you three features right now. Two of them are completely free. And the last one, I'm not going to spoil it, but the

Anthropic’s $30B raise and Claude Pentagon deal

AI literally opens a browser and tests its own code without you touching anything. Let's start with what 3. 1 actually changed. So, I gave both models the exact same task to replicate the Google AI Studio interface from a screenshot. Same prompt, same image. Here's 3 Pro. Here's 3. 1 Pro. See the difference? And that's just a UI test. On Google's own benchmark for logical reasoning, 3. 1 Pro scored 77%. The previous version scored 31%. More than double. And this is the model that's now available to everyone. Let me show you what that intelligence actually looks like in practice. All right, feature one. I'm in the Gemini app. Pro model selected. That's 3. 1 Pro. I start simple. Give me a two sentence visual description of a cyberpunk coffee cup for a fictional brand called Neon Bean. Dark, rainy, neon lit mood. Description's ready. Now, here's where it gets interesting. Next, I ask it to generate a cinematic image from that description. glowing pink neon letters spelling neon bean on the cup. Look at that. The text rendering is clean. The neon lighting, the reflections. This is Nano Banana Gemini's image model. One prompt. That is genuinely impressive. But here's the part that actually matters. I download this image, drag it into Gemini, and now I'm asking for video. Animate this. Steam rising, neon text flickering. Include ambient audio of rain hitting a window. 8 seconds of cinematic video with AI generated ambient audio from a still

Meta’s creepy “post-after-you-die” patent explained

image I made 60 seconds ago. Now I want a second angle for this ad. So I take the original image and ask for a completely different shot. Macro closeup of the neon text reflecting in the dark coffee. Camera slowly panning downward. Two shots, same brand, same visual style, different angles. I just directed an AI the way a filmmaker directs a scene. And for the soundtrack, I switched to Liria 3. That's Gemini's music model. You can find it by clicking create music in the tools menu or just search Liria 3 Gemini. I'll drop the link in the description. I will upload the video and say, "Write an 8-second dark synth wave track. Realistic female vocals singing wake up in the neon glow. Match the mood to the visuals. " — AI generated vocals matched to a video that was matched to an image. All from text prompts. One person, no designer, no editor, no music producer. Now, this next one, I almost didn't believe it was free. No paid plan, nothing. Same Gemini, same pro model, but this time I'm asking it to do something completely different. I paste in this prompt and it's a wild one. I'm asking it to create an SVG of a glowing blue elephant flying on a magic carpet above the Taj Mahal

Claude Sonnet 4.6 and autonomous to‑do execution

starry cosmic sky, yellow moon, water reflection, the whole 9 yards. Now, before I hit run, let me quickly explain what we're asking for. An SVG is basically an image made entirely from code, not pixels. So, we're basically asking AI to draw, but with code instead of a brush. I hit canvas mode and send it. And now we wait 4 minutes. That's how long this model sat and thought about one prompt. Okay, look at that. A glowing blue elephant on a magic carpet. The Taj Mahal with a reflection in the water. Stars, a crescent moon, and the trunk. It's moving. That subtle animation all generated from code. over 30,000 tokens from one prompt, not an image, not a video. Pure code. And if I zoom in, perfectly sharp at any scale, because it's math, not pixels. The file about 30 kilobytes. A GIF of the same quality would be megabytes. 4 minutes of thinking. And it drew an animated elephant on a flying carpet from code. That's what deeper reasoning gets you. Last feature, and this is the one I told you about at the start, Google Anti-gravity. free to download right now. It looks like VS Code, but there's an AI agent inside that doesn't just suggest code. It plans, builds, launches, tests, and fixes all by itself. I select Gemini 3. 1 ProH High. One prompt, build a complete user registration and login web app. Then launch it, test it in the browser, and fix any bugs. Look, it's already made

Gemini image → video → music workflow + Antigravity

its own project plan. Implementation steps, checkboxes. I haven't written a single line of code. It's writing the front end, the CSS, the JavaScript validation completely autonomously. Now watch what happens next. It just opened Chrome on its own. No human input. Watch the browser. The AI is clicking buttons, filling in form fields, navigating between pages. It just tested its own code like a human QA tester. Nobody touched anything. And when it found a bug, it switched back to the code editor, made the fix, went back to the browser, and retested. That's an autonomous feedback loop. That's not a chatbot that suggests code. That's an AI agent that builds, ships, and validates. I mean, isn't that insane? Six months ago, this workflow would have cost you a team of four and a week of back and forth. Now, one person, one afternoon, zero budget. That's not just a model update. That's a career advantage for anyone paying attention. There's a lot more in 3. 1 Pro that we couldn't cover today. Vibe coding in AI Studios build mode. Canvas for collaborative editing, Deep Think for research grade reasoning, those deserve their own videos and they're coming. But right now, Anti-gravity and AI Studio is free. Go try it. Links in the description. And do drop a comment. Which of these three features are you trying first? And if you want me to do a full deep dive tutorial on anti-gravity, let me know because that thing deserves its own 15-minute video. I'll make it happen. Figure AI. If you're building a product and your UX review process is still screenshots and Loom videos, Figure plugs into your live app through a Chrome extension. It maps your flows, spots problems, and builds prototypes that match exactly what you've already built. Edit with Eva. This one's for anyone making videos. Edit with Eva is an AI editor that actually understands

Best AI tools to try this week (quick rundown)

your footage. You drop in your raw clips, tell it what you want, and it selects the right scenes, cuts your retakes, and adds B-roll, captions, and voice over. It's working with your footage, not generic templates. Seance 2. 0. There are rumors that Seance 3. 0 will generate 18 plus minute videos with consistent characters, full plots, and multilingual lip-sync dubbing. While we aren't there yet, Cance 2. 0 just dropped and it's already generating multi-shot cinematic scenes from a text prompt with realistic camera movements. Quen 3. 5 just went open source. It has a massive brain but only uses the exact parts it needs for your specific question, making it both smart and efficient. It reasons before it responds and is designed to integrate into apps to do the heavy lifting for you. It's free, open- source and going head-to-head with the best closed models. Lov is an AI therapist you talk to by voice anytime. It was built with

Final thoughts and what to expect next

PhD psychologist and is completely private because let's be real, we all need that sometimes. Pomelli photo shoot. Google just added photo shoot to Pomellle. You upload one product photo, pick a style, studio, lifestyle or whatever you need and it generates marketing ready shots in seconds. It's completely free. So if you sell anything online, this one's for you. I'm already tracking three more announcements that didn't make this video. Those go to WhatsApp first link below. If you want these breakdowns every week, subscribe and tell me in the comments which update are you trying out first.

Другие видео автора — Vaibhav Sisinty

Ctrl+V

Экстракт Знаний в Telegram

Экстракты и дистилляты из лучших YouTube-каналов — сразу после публикации.

Подписаться

Лучшие методички за неделю — каждый понедельник