Grok 0.9’s video generation is surprisingly impressive — it’s free (with daily generation limits), fast, and delivers natural character movements and camera work.

Open-source Qwen Code just leveled up—now it plans before it edits and sees your screenshots. 🚀

ALS patient Nick uses a brain implant to control a robotic arm—feeding himself in real time.Neuralink’s BCI translates brain signals to precise robot-arm motion.

Blip is a cross‑platform, encryption‑backed direct transfer app that eliminates traditional upload/download delays. With LAN acceleration, unlimited file sizes, project‑bundle integrity, and robust resumability, it streamlines media workflows across devices and geographies.

World’s first all‑weather humanoid robot hits IP66—industrial ops just got outdoorsy. DEEP Robotics’ DR02 is built for real work: full‑body waterproof/dustproof, h quman‑like form for tool compatibility, quick‑detach modules for low O&M, and high compute for continuous upgrades.

GAGA-1 just dropped: an AI “actor” that fuses voice, lipsync, and performance. It’s not fully free—there’s roughly 60 seconds of free generation per month; the API is currently unavailable. It’s built for dialogue-driven scene prototyping.

Figure’s third‑gen humanoid, Figure 03, flips the script on home + factory robots. 🚀

Google’s no‑code AI mini‑app builder, Opal, moves beyond the U.S. into 15 countries and ships workflow debugging + parallel execution. It’s built for creators who want natural‑language app building with real reliability, not just demos.

Gemini 2.5 Computer Use: agents that actually use UIs—click, type, scroll—fast and safe. ⚡ blog.google/technology/g...

NeuTTS Air is a cheeky open-source challenger to cloud TTS—super realistic, on your device. 🔊

Meet Tinker: a techy, playful training API that lets researchers fine‑tune open models with LoRA—minus the infra headaches. 🚀 thinkingmachines.ai/tinker/

OpenAI vs DeepSeek - only this time it’s Sam and I repping the brands! 😂😂😂

Mac VS PC but it's DeepSeek vs OpenAI

Making short ads with Sora 2 is a breeze.

Let’s bring some energy here 🕺😂😂😂

Sora 2 just leveled up text-to-video—now with physics, control, and synced audio. 🤯 OpenAI introduces a more physically accurate, realistic, and controllable video model, plus synchronized dialogue and sound effects. www.youtube.com/watch?v=lEcg...

Claude Sonnet 4.5 just rewrote the coding playbook. 🤖⚡ It’s billed as the best coding model, strongest for complex agents, and the smartest at real computer use.

DeepSeek drops V3.2-Exp: faster long-context, lower bill. Yes, both. Built atop V3.1-Terminus, the experimental V3.2-Exp debuts DeepSeek Sparse Attention for fine-grained sparsity—aiming to keep output quality while cutting compute on long inputs. Oct 15, 2025 (UTC).

Parents and teens, meet smarter guardrails for AI—link accounts, tune safety, stay informed. OpenAI just rolled out parental controls in ChatGPT on web (mobile soon).

Open-source Hunyuan3D‑Omni lands as a playful but serious rival to commercial 3D control systems. 🚀 Built on Hunyuan3D 2.1, it acts like a “ControlNet for 3D,” fusing multi‑condition inputs to fix single‑view occlusions and messy geometry.

Meet HunyuanImage 3.0—a native multimodal, MoE‑Transfusion system that fuses Diffusion with an LLM. It claims closed‑model parity while staying open, built for creators who need crisp typography, complex prompts, and fast workflows.

Google just put AI avatars into your video workflow—directly in Docs’ new sibling, Vids. ⚡ Write a script, pick an AI avatar, auto‑generate and drop it into your timeline. No extra tools, no messy handoffs—production goes from ideation to talking‑head in minutes.

Generative OS, but make it real-time. A Google prototype uses Gemini 2.5 Flash‑Lite to render an entire UI on the fly—each click becomes fresh HTML streamed into the page. No fixed windows, no static menus; the model composes screens from a “UI constitution” + your latest interaction.

AgiBot X2 shows extreme agility: balanced hops, rapid recoveries, and clean flips. The demo focuses on dynamic stability and whole‑body control, pushing humanoid motion closer to athletic human moves while keeping repeatability on hard surfaces.

Suno drops a first-of-its-kind generative audio workstation—Studio—built for creators blending AI with classic music-making. 🎛

My fifth original music MV is officially out! I poured effort into both the music and the AI-generated visuals. Even though I didn’t use the latest AI models for most of the production, the final quality is a clear step up from my earlier work. Click the link to check it out—hope you enjoy it!🩷🩷🩷

Ready to remix your creativity? Meta just launched Vibes - a new AI video playground powered by Midjourney! Discover, create, and share short-form AI videos—all in one feed. Vibes lets you start from scratch, remix others’ work, and cross-post to Instagram & Facebook.

OpenAI just dropped ChatGPT Pulse—a game-changer for proactive AI updates! 🚀 ChatGPT Pulse is now previewing for Pro users on mobile, delivering daily personalized updates from your chats, feedback, and connected apps like your calendar.

Meet Qwen3-VL: the open-source challenger shaking up the vision-language model scene! 🚀 A new player from Tongyi Lab, Qwen3-VL rivals commercial giants by operating GUIs, coding charts from mockups, and recognizing everything from daily life to niche fields.

Meet Qwen3-LiveTranslate-Flash: the real-time translation wizard making global chats lightning-fast ⚡️ Powered by Qwen3-Omni and trained on millions of hours of multimodal data, Qwen3-LiveTranslate-Flash delivers seamless audio and video interpretation in 18 languages.

Design to code, but make it magic! ✨ Claude Code + Figma just changed the game: using MCP, it reads your Figma mockups at the data level—component hierarchies, design tokens, auto-layout rules—and instantly turns them into production-ready code.

Qwen3-Max just dropped, and it’s not here to play—it’s here to dominate! 🚀 Meet the new flagship from Alibaba’s Qwen team, supercharged for coding and agentic tasks. No previews, no waiting—just raw power. Project: chat.qwen.ai

🚀 Ready for next-level 3D creation? Rodin Gen-2 just dropped, bringing massive upgrades to mesh quality, part generation, and texturing. If you’re into 3D assets, this is a game-changer—think smoother, cleaner, and more detailed than ever.

Mixboard just dropped from Google Labs, and it’s not your average mood board. 🚀 Imagine an AI-powered canvas where you can dream up anything—home decor, product ideas, or even a board full of tiny, cute bananas.

🚀 Next-level video creativity just got turbocharged! Kling AI has unveiled its 2.5 Turbo Video Model, promising a leap in creative power for video generation. With enhanced realism, support for both anime and real-life styles, and a more affordable price, it’s set to shake up the AI video scene.

Gamma API just launched on Product Hunt, letting anyone turn raw text, data, or images into polished decks and reports—automatically. Plug it into Zapier, n8n, or your own SaaS to supercharge your workflow. Try it now: gamma.app/signup?r=nqq...

Suno 5 is out—remixing a classic tribute 🫡 Hope you all enjoy it! ❤️❤️❤️

Suno just dropped v5, and it’s rewriting the rules of AI music creation! Meet the world’s best music model—now live for Pro and Premier users. Expect richer audio, lifelike vocals, and creative control that’ll make you rethink what’s possible. Try it now: suno.com/invite/@luok...

The image from yesterday didn’t work well spatially, so today we’re running the Wan 2.2 Animate model again with a different location and a female character design. Now I get why the model caps at 15 seconds by default… longer generations cause the visual details to degrade. 😅

Qwen-Image-Edit-2509 drops with game-changing upgrades—think multi-image editing (yes, you can blend “person + product” or “person + scene” in one go), plus way better consistency for faces, products, and even text.

🚀 Open-source AI is leveling up! Meet DeepSeek-V3.1-Terminus — the open-source challenger to commercial AI giants. This update builds on V3.1’s strengths, tackling user feedback head-on for a smoother, smarter experience.

Pancake optics are about to flip the XR industry on its head! 🥞✨ True human-like 220° FOV is a game changer. Most current XR headsets max out at 120-140°, but HyperVisionXR aims for a panoramic experience that matches natural human vision.

Meet Qwen3-Omni: the open-source challenger shaking up the AI world! The first truly end-to-end omni-modal AI, Qwen3-Omni unifies text, image, audio, and video in a single model—no more trade-offs between modalities.

I just tried out Wan 2.2 Animate, and the results are so convincing it’s hard to believe they’re AI-generated.

Unitree G1 has mastered more quirky skills😱 With its new “Anti-Gravity” mode, G1 can stay stable through wild moves—and if it takes a tumble, it pops right back up. Stability boost: Any action sequence, instant recovery.

This isn’t a leaked clip from KPOP Demon Hunters—it’s actually AI-generated. The more realistic these fake behind-the-scenes clips look, the harder they become to distinguish from the real thing.

Metaballs is redefining 3D modeling on visionOS—think digital clay, but you can grab it in mid-air! Created by the mind behind Vibescape, Metaballs leverages visionOS 26’s next-level gesture controls for a hands-on, real-time modeling experience.

Meet Hailuo AI Agent — the next-gen creative sidekick that turns your ideas into visuals, fast! Just say what’s on your mind and watch it come alive on an infinite canvas. Free for one week: Try images, videos, and audio creation in one place.

Did you know Meshcapade’s motion downloads come with perfectly synced camera data? 🎥 No more guesswork—just plug & play for instant 3D shots. Every motion file includes camera extraction, so your movement and perspective are always in harmony.

Meet Hailuo AI Agent — the next-gen creative sidekick that turns your ideas into visuals, fast! Just say what’s on your mind and watch it come alive on an infinite canvas. Free for one week: Try images, videos, and audio creation in one place.