Notes from the infrastructure
Essays on AI video production, generation vs. production workflows, and how tools will evolve into real production systems.
AI Video Generation vs Production: why generation isn't the real problem
Models are getting better every week, but creating a complete video remains broken. The real bottleneck isn't generation — it's production.
Read article →
From Prompt to Production: the next step in AI video creation
Prompt-based tools hit a ceiling fast. The next generation of AI video platforms will break ideas into scenes, manage references and enable precise control.
Read article →
Why most AI video tools feel the same (and what's missing)
As models improve, outputs converge. Tools start to feel identical. The only thing that doesn't commoditize is production systems.
Read article →
Character consistency in AI video: the hardest problem nobody talks about
Every AI video demo shows a perfect character. Try to use that character across five shots and watch them turn into a stranger. Here's why — and how production pipelines fix it.
Read article →
How to prompt cinematic lighting in AI video (with examples)
"Cinematic" isn't a style — it's a lighting decision. Working prompts for golden hour, low-key noir, Rembrandt portraits and motivated practicals.
Read article →
Studios vs. solo creators: how AI video is splitting the industry
The narrative is "AI is replacing filmmakers." The reality is more interesting. Production houses and solo creators are using AI in fundamentally different ways.
Read article →
The new brand film: how AI is reshaping commercial production
Brand films used to take six weeks and a crew of forty. AI brings them down to days. Here's how Nike, Prada and emerging brands use AI for commercials.
Read article →
Why we killed subscriptions: pay-as-you-go credits in AI video
Subscriptions are SaaS default. They're wrong for AI video, where usage is bursty and unpredictable. Here's why we moved to credit packs.
Read article →