Monet's Painting Called AI Slop

PLUS: Odyssey's AI world model powers playable GoldenEye, Google's Gemini Omni launches video

Author

Gen AI Art
May 21, 2026 • Estimated Reading Time: 3 minutes

In partnership with

Late last week, an anonymous artist labeled a detail from Monet’s Water Lilies as “AI-generated” on X — and watched as 6.7 million people tore it apart. The responses came fast: “emotionless,” “cluttered slop,” and one commenter’s 700-word breakdown of its supposed compositional failures. The painting was a real 1915 Monet, currently hanging in a Munich museum.

While OpenAI has built the feature with strict privacy barriers, will users be comfortable sharing their most sensitive information with an AI? The move tests the boundaries between AI as a creative tool and a trusted personal companion.

Today in AI:

  • Real Monet called “AI slop” — and 6.7 million people agreed

  • Odyssey’s AI world models generate a real-time playable GoldenEye

  • Google’s Gemini Omni brings video creation to YouTube Shorts for free

Blu Dot surpasses 2,000% ROAS with self-serve CTV ads

Home furniture brand Blu Dot blew up on CTV with help from Roku Ads Manager. Here’s how:

After a test campaign reached 211,000 households and achieved 1,010% ROAS, the brand went all in to promote its annual sales event. It removed age and income constraints to expand reach and shifted budget to custom audiences and retargeting, where intent was strongest.

The results speak for themselves. As Blu Dot increased their investment by 10x, ROAS jumped to 2,308% and more page-view conversions surpassed 50,000.

“For CTV campaigns, Roku has been a top performer,” said Claire Folkestad, Paid Media Strategist, Blu Dot. “Comping to our other platforms, we have seen really strong ROAS… and highly efficient CPMs, lower than any other CTV partner we've worked with.”

Using Roku Ads Manager, the campaign moved from a pilot to a permanent performance engine for the brand.

What’s new? Conceptual artist SHL0MS posted a cropped Monet on X with a fake “Made with AI” label — and 6.7 million people took the bait, calling the authentic 1915 Water Lilies painting “emotionless slop” and an “incoherent muddle of inconsistently saturated greens.”

What matters?

  • The painting is housed at Munich’s Neue Pinakothek museum — art historian A.V. Marraccini identified the hoax by recognizing the impasto brushwork typical of Monet’s late period.

  • A 2024 Nature study by Grassini & Koivisto found people consistently downgrade identical artworks after learning AI created them — meaning the label, not the image quality, drove the criticism.

  • One commenter produced a 700+ word breakdown dissecting the supposed fake’s compositional failings — none of it accurate.

Why it matters?

The “AI-generated” label is now so loaded it can make a Monet look bad to millions of people. For creators using AI tools, this perception bias — not the quality of their work — may be the biggest obstacle to having it taken seriously.

GUIDE

What’s new? Odyssey launched Agora-1, a multi-agent world model that places up to 4 players inside a shared AI-generated GoldenEye deathmatch in real-time — with every frame generated on the fly, no traditional game engine involved.

What matters?

  • Agora-1 uses a two-model architecture: one model evolves a shared world state from player actions, a second renders each player’s unique viewpoint — decoupling simulation and rendering in a way that mirrors game engine design but runs entirely on learned systems.

  • Running alongside it, Starchild-1 became the first real-time world model to generate synchronized video and audio together at up to 24 fps.

  • Beyond gaming, Odyssey points to applications in collaborative robotics, multi-agent reinforcement learning, and generating imagined training environments for AI agents.

Why it matters?

AI could already generate video or simulate environments — but not in real-time with multiple participants sharing the same generated world. Agora-1 and Starchild-1 together point toward AI-native game engines and simulation platforms that could reshape how interactive experiences are built.

SPONSORED BY DEEL

Hiring in 8 countries shouldn't require 8 different processes

This guide from Deel breaks down how to build one global hiring system. You’ll learn about assessment frameworks that scale, how to do headcount planning across regions, and even intake processes that work everywhere. As HR pros know, hiring in one country is hard enough. So let this free global hiring guide give you the tools you need to avoid global hiring headaches.

What’s new? Google launched Gemini Omni at Google I/O 2026 — a video generation model that creates and edits clips from any combination of text, images, audio, or video — rolling out today free to YouTube Shorts and YouTube Create users, and to Gemini subscribers.

What matters?

  • Gemini Omni uses conversational editing: each instruction builds on the last, preserving character consistency and scene physics across revisions — unlike one-shot generators where any change can break continuity.

  • All videos are watermarked with SynthID by default, making Gemini Omni one of the first major AI video tools to ship with invisible provenance tracking built in.

  • The current Flash tier is capped at 10-second clips at launch — shorter than Sora’s 60-second maximum — though Google says this is a deployment decision, not a model constraint.

Why it matters?

Free access through YouTube Shorts puts AI video creation in the hands of millions of casual creators overnight. Gemini Omni’s conversational editing approach could shift how creators iterate on video — closer to talking through ideas than adjusting settings in a timeline.

Everything else in AI

Karpathy joined Anthropic’s pre-training team, where the OpenAI co-founder and former Tesla AI lead will build a new unit that uses Claude to accelerate foundational model research.

Musk lost his lawsuit against OpenAI after a jury unanimously ruled the case was filed years past the statute of limitations, clearing a major obstacle for OpenAI’s anticipated IPO.

Cursor shipped Composer 2.5, a coding model matching frontier performance at under $1 per task versus $11 for comparable solutions, built on Moonshot’s Kimi K2.5.

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 →