Grant Harvey

TwelveLabs launched Marengo 3.0 at AWS re:Invent 2025, enabling natural language video search with composed queries (image + text) that processes hour-long videos 30x faster than competitors, now available on Amazon Bedrock.

Welcome to "The Soul Document": Inside the Mind That Anthropic Built.

An AI-generated country song mimicking Blanco Brown's voice hit #1 on Billboard without his consent, exposing how music labels went from suing AI companies like Suno to partnering with them—even as the platform generates 7 million songs daily and the industry still can't figure out fair compensation for artists.

Remember when everyone said AI would only replace tech jobs? Turns out, they were looking at the tip of the iceberg.

Ilya Sutskever explains why today's AI models ace benchmarks but fail basic tasks, why we're entering a new "age of research" where ideas matter more than scale, and what SSI is actually building—essential viewing for anyone trying to understand where AI capabilities are actually headed.

Two very different AI “personal shoppers” just launched right before Black Friday—one built on a chat interface, the other on search—and they’re about to change how you hunt for deals and gifts.

The authors of the viral "AI 2027" report now say superintelligence is more likely around 2030, sparking backlash over the original branding—while a new 2032 scenario details slower but still dramatic timelines for AI automation, geopolitical shifts, and two possible endgames for humanity.

Catch up on the AI stories we couldn’t fit in today’s newsletter—from VAST’s $1.17B CoreWeave deal and the EU’s “Digital Omnibus” push, to a high-severity runc escape bug, a surprise HDD shortage, and fresh model/tool moves from DeepMind, Cursor, and Inworld.

AI pioneer Dr. Fei-Fei Li argues that the next frontier for AI is "Spatial Intelligence," a move beyond text-based models to create interactive, physics-aware "world models" that will revolutionize everything from creativity and gaming to robotics and scientific discovery.

A chaotic week of communication gaffes from OpenAI triggered a sell-off in AI stocks, but the real story is the growing developer distrust in its platform, creating a massive opening for competitors like Anthropic.

Meta's Chief AI Scientist Yann LeCun argues the path to human-level intelligence isn't scaling up today's LLMs, but building a new "Objective-Driven" architecture that learns predictive world models through observation, much like a baby.

Moonshot AI's new open-source Kimi K2 Thinking model can autonomously execute hundreds of tool calls to solve complex problems, and thanks to community efforts, you can now run a compressed version of the 1-trillion-parameter model on your own hardware.