Denoised

Ben Affleck’s AI company InterPositive was recently acquired by Netflix, raising new questions about how filmmakers are post-training open-weight video models. Joey and Addy break down the technical details behind Affleck’s workflow, the ethics of model training, and why local, studio-controlled models are gaining traction. Plus, a look at new precision editing tools from Ideogram and Flux.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

Addy and Joey discuss Jeffrey Katzenberg’s argument for AI in entertainment, questioning whether his focus on studio pipelines misses how fast individual creators are adopting these tools. They also share real-world tests using Anthropic’s Opus 5.5 in Blender and evaluate Alibaba's open-weight Qwen model.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

Joey returns from TIFF with observations on Kling's presence and the industry's ongoing misunderstanding of hybrid AI production. Plus, a close look at ChatGPT Image 2.5’s character consistency updates, how TypeSafe's Jev AI model tackles non-LLM tasks, and new US-hosted compute options for Seedance 2.0 on fal.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

Addy and Joey explore GPT-6 Astra’s ability to generate complex, accurate 3D cityscapes in Blender from simple prompt goals. They also review DaVinci Resolve 21.1's new AI agent integration and discuss why an Indonesian TikTok meme character might force the courts to clarify AI copyright law. --

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

In this episode of Denoised, Joey and Addy break down the latest breakthroughs in AI world models, from World Labs’ photorealistic Atlas to Runway’s GWM World 2 and Solaris interface. They also dive into tech news, including OpenAI’s Mac Mini hoarding and YouTube creator Markiplier’s recent investments in GoPro and Strada.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

NVIDIA agreed to buy Hugging Face for $13B. Addy and Joey break down what's behind the deal and what it signals for open source AI. They also debate gray box vs. natural environments for AI hybrid shoots, cover Apple's new M5/M6 chip updates, and compare Wan 3.0 to Seedance 2.5.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

The MPA and ByteDance reached a deal on Seedance IP and likeness protections — Addy and Joey break down what it actually means for studios and why Chinese-based inference remains a blocker. Topaz releases Hyperion 2.5, converting 8-bit video to HDR. Plus, The Odyssey's IMAX 70mm depth of field and what drove its box office run.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

An AI-generated "behind-the-scenes" disaster video hit 750 million views in four days — and most viewers had no idea it was fake. Addy and Joey dig into why it went viral, what it reveals about audience appetite for movie magic, and why the filmmaking details don't hold up. Then, a live demo of Decart's Lucy 2.5 real-time video generation model, a look at MiniMax H3's multi-input open-weight release, and a breakdown of AI model pricing shifts and Chinese chip optimization trends shaping the industry.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

Addy and Joey recap SIGGRAPH — covering EVRN's image-to-3D tool, ComfyCode.ai's metadata-driven JSON prompting, and Beeble's SwitchX 2.0 push toward pro-grade outputs. Neill Blomkamp releases a fully AI-generated short film, Flux 3 enters the video model race, and the duo discuss what the AI film pipeline might look like in two to three years.

--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.

Joey tests OpenAI Codex + Blender MCP to reconstruct a real building in 3D from phone photos — and the results have real implications for production workflows. Plus: what to expect at SIGGRAPH and a free AI Workflows Summit in LA.

Also: Nuke's SmartRoto update, the AI-generated Odyssey trailer debate, Seedance 2.5, and Thinking Machines' open-source Inkling model alongside Kimi 3's massive parameter count.
--

The views and opinions expressed in this podcast are the personal views of the hosts and do not necessarily reflect the views or positions of their respective employers or organizations. This show is independently produced by VP Land without the use of any outside company resources, confidential information, or affiliations.