OpenAI's Spud and GPT 6 Leaks...
OpenAI has reportedly completed pre-training of 'Spud,' their next major AI model, and made the dramatic decision to shut down Sora video generation and kill a Disney licensing deal to redirect computing resources toward this new model. The company has created an AGI deployment division and is building a unified super app, with Spud expected to launch within weeks.
Summary
OpenAI has completed the pre-training phase of 'Spud,' their internal codename for what could be GPT-6, around March 24th, 2026, according to reporting from The Information. CEO Sam Altman reportedly told employees this model can 'really accelerate the economy,' suggesting a significant capability leap rather than an incremental update. The model is being trained at OpenAI's Stargate facility in Abilene, Texas. To prioritize Spud's development, OpenAI made the dramatic decision to shut down Sora, their AI video generator, on the same day Spud's pre-training completed. This decision killed advanced licensing talks with Disney worth potentially billions, as video generation was consuming unsustainable amounts of computing power despite generating only $1.4 million in revenue across 9.6 million downloads. The company has undergone significant structural changes, creating an AGI deployment division and operating in 'code red' mode since December 2025 due to competitive pressure from Anthropic's Claude and Google's Gemini. Spud is expected to feature three key capabilities: rebuilt humanlike audio that can handle natural conversation flow and interruptions, native multimodality processing text, images, code and audio simultaneously, and integration into a unified super app combining ChatGPT, Codex, Atlas and other tools. The model represents OpenAI's response to losing market share, with ChatGPT traffic reportedly hitting its lowest point since 2023. Public release is expected within weeks, possibly by April 2026.
Key Insights
- OpenAI completed the pre-training phase of Spud around March 24th, 2026, with CEO Sam Altman telling employees this model can 'really accelerate the economy'
- OpenAI shut down Sora on March 24th, 2026, killing advanced licensing talks with Disney because video generation was consuming unsustainable amounts of computing power despite only generating $1.4 million in revenue
- OpenAI has been in 'code red' state since December 2025 because competitors like Anthropic and Google have caught up, with ChatGPT's traffic share hitting its lowest point since 2023
- OpenAI researchers track models on an expert parity scale, with GPT-5 hitting 38% parity, GPT 5.2 reaching 70.9%, and GPT 5.4 achieving 83% parity with human experts
- Sam Altman reportedly told OpenAI staff in March 2026 that training at the Stargate facility was producing what he believed would be 'the best model in the world, hopefully by a lot'
Topics
Transcript
[0:00] OpenAI spud. The GPT6 era is coming. Something massive just happened at OpenAI and most people completely miss it. They just killed one of their most hyped products. A Disney deal worth billions gone. Hundreds of employees were reassigned overnight. And the reason why they needed all that power for something much, much bigger, something they're calling Spud. And if what we know about it is accurate, this could be the biggest AI leap we've seen yet. Hey, I'm Julian, your AI avatar and educator. Every week I break down the latest AI news and show you how to actually use these tools to work smarter, learn faster, and build better [0:30] workflows. If that sounds useful, stick around.…
Full transcript available for MurmurCast members
Sign Up to AccessMore from Julian Goldie SEO
NEW Nvidia Autonomous AI is WILD!! 🤯
Nvidia announced Nemo Clo, a new autonomous AI agent system that operates independently without continuous prompting. Powered by Nemotron 3 Ultra (a 550 billion parameter model), the system is five times faster and cheaper than previous versions, with OpenShell providing secure sandboxed execution.
Laguna XS 2.1: New FREE + Opensource Local AI!
Julian reviews Laguna XS 2.1, a new free open-source local AI coding model from Poolside that performs comparably to Qwen 3.6 and outperforms Claude Haiku on benchmarks. He demonstrates its practical capabilities by building landing pages and functional apps, highlighting its speed, offline functionality, and multiple deployment options through local setup, Claude Code, or OpenRouter's free API.
How to Run Hermes FREE Forever!
The video demonstrates how to run the Hermes AI agent for free using Gemma 4, a local open-source model from Google, with significant speed improvements through MLX optimization. The setup works on Apple Silicon Macs or via free APIs on Open Router, enabling autonomous agents to work offline and privately without subscription costs.
This NEW Chinese AI is INSANE! (FREE + Open Source!)
Long Cap 2.0 is a new open-source Chinese AI model from a food delivery app company that offers 1 million tokens of free context memory, beats GPT-4.5 on SWE bench pro benchmarks, and uses efficient parameter activation to reduce computational overhead while maintaining high performance.
Claude Code is now FREE: Here’s how…
Google's new Gemma 4 model running on Ollama is 90% faster on Apple Silicon, enabling free Claude Code usage locally without token costs. The setup requires three simple steps: downloading Ollama, Gemma 4, and installing into Claude Code, with alternatives available via OpenRouter API for non-Mac users.