Decart's Lucy Edits Live Video

PLUS: Google and The Met turn art history into a mirror, and a Toyota-backed startup lands $300M for factory robots

Author

Gen AI Art
July 19, 2026 • Estimated Reading Time: 3 minutes

Sponsored by

Decart just gave live video an edit button. Its new Lucy 2.5 model can swap outfits, change backdrops, or erase objects from a video stream in real time — no waiting for a render, no separate editing pass.

Every AI video tool so far has worked on clips you generate and then fix afterward. Lucy 2.5 collapses that gap to near-zero latency, and if live-stream editing holds up outside a demo, it changes what’s possible for anyone shooting video as it happens — creators, streamers, and product teams alike.

Today in AI:

  • Decart’s Lucy 2.5 edits live video in real time

  • Google and The Met’s AI mirror for your art taste

  • Toyota-backed Walden Robotics raises $300M for factory robots

How Jennifer Aniston’s LolaVie brand grew sales 40% with CTV ads

The DTC beauty category is crowded. To break through, Jennifer Aniston’s brand LolaVie, worked with Roku Ads Manager to easily set up, test, and optimize CTV ad creatives. The campaign helped drive a big lift in sales and customer growth, helping LolaVie break through in the crowded beauty category.

What’s new? Decart unveiled Lucy 2.5, a live AI video model that transforms footage while it streams — swapping characters, backgrounds, and effects at 30 FPS and 1080p with near-zero latency.

What matters?

  • Self-Anchoring technology keeps edits consistent frame after frame across extended live streams, instead of flickering or drifting the way earlier real-time models did.

  • The model applies physically-aware effects like water, fire, sand, and slime, plus full object removal with natural background reconstruction, all driven by text prompts and reference images.

  • New MXFP8 and NVFP4 quantization deliver up to 4x faster compute, which is what makes genuine real-time editing possible — try it live yourself.

Why it matters?

Live video has been the one format AI editing couldn’t reach — you could generate a clip and fix it, but not reach into a stream while it’s happening. Lucy 2.5 turns editing into a live performance, opening up use cases that clip-based tools like Gen-4 or Kling were never built for.

GUIDE

What’s new? Google Arts & Culture and the Metropolitan Museum of Art launched Art Aura, a Gemini-powered tool that lets you drag artworks, styles, or phrases onto a canvas to reveal your own taste across 200,000 digitized objects from The Met’s collection.

What matters?

  • The tool grew out of a technologist-in-residence program, where a creative technologist spent six months prototyping generative AI directly inside the museum’s galleries with curators.

  • Anyone can try Art Aura for free and walk away with a personalized visual portrait built from the art they gravitate toward — no account or purchase required.

  • The launch marks 15 years of Google-Met collaboration, which started with virtual gallery tours in 2011 and expanded into full open-access digitization in 2018.

Why it matters?

Museums have mostly used AI for cataloging and search, not as a creative mirror for visitors. Art Aura treats centuries of art history as raw material for understanding your own taste, and it’s live right now for anyone to test.

SPONSORED BY ATTIO

Introducing The First Agentic CRM

Get revenue agents, workflows, and automations across every stage of your motion. Access customer data in real time through Attio's web app, MCP, API, and SDK.

Then Ask Attio anything about your business and get instant answers.

It's the CRM that runs the work behind every win.

What’s new? Walden Robotics emerged from stealth with a $300 million seed round and a $1.1 billion valuation, and its general-purpose robots are already working shifts on a Toyota factory floor in North America.

What matters?

  • The round was co-led by Toyota and Deviation Capital, with Nvidia, Boeing, and Samsung NEXT among the backers — a rare mix of automotive, aerospace, chip, and electronics money in one seed round.

  • Walden spun out of Toyota Research Institute in January and moved from its first pilot to real production work in under two months.

  • The robots run on Large Behavior Models and Diffusion Policy learning, so they pick up tasks like part transport and machine tending by learning physical skills instead of following hard-coded scripts.

Why it matters?

Most humanoid robot startups are still stuck in demo mode. Walden skipped that step and put robots directly onto a real assembly line, which is the clearest signal yet that general-purpose robots are closer to factory-ready than the hype suggested.

Everything else in AI

Google published an ultimate prompting guide for Nano Banana, showing how scene-level natural-language prompts now beat stacking disconnected keywords.

God of Prompt packages prompt libraries, templates, and tutorials into one bundle for turning ChatGPT and Midjourney skills into a side income.

OpenAI launched Codex Micro, a $230 physical keypad built with Work Louder that lets you monitor and steer AI coding agents with a dial, joystick, and status lights.

China unveiled WAICO, a 29-nation AI cooperation alliance led by Xi Jinping, positioning itself as a rival framework to the U.S.-led approach to global AI governance.

Essential AI Guides - Reading List:

Let us know!

Work with us

Reach 100k+ engaged Tech Professionals, Engineers, Managers and decision makers. Join brands like MorningBrew, HubSpot, Prezi, Nike, Ahref, Roku, 1440, Superhuman, and others in showcasing your product to our audience. Get in touch now →