Frank's Brain Dump

I built a chatbot of myself.

Yes, I know how that sounds.

Somewhere between visionary, ridiculous, and the beginning of a very bad science fiction film in which the protagonist should really have read the warning labels before pressing Enter.

But here we are.

Over the last months I have written a lot about AI in filmmaking, virtual production, creative control, authorship, production economics, pipelines, tools, and the increasingly strange question of how to remain human in a production culture that keeps pretending speed is the same thing as intelligence.

At some point, I had to admit something slightly embarrassing.

I had created my own intellectual junk drawer.

Not because the material was useless. Quite the opposite. There were articles, book fragments, published texts, arguments, recurring positions, half refined obsessions, and a few thoughts that had clearly survived several rounds of irritation and were still standing. But finding the right thought at the right moment had become its own small production problem.

And because I apparently cannot solve a simple problem without turning it into a system, I built Ask the Lab.

It is a chatbot based on my Augmented Cinema Lab material. My articles. My books. My public profile. My recurring arguments. My positions on AI, filmmaking, virtual production, authorship, creative control, and the hidden politics of production technology.

The goal was not to create another polite little chatbot that answers everything in the tone of a hotel lobby.

I wanted something that could make the material searchable while still sounding reasonably close to me. Which means it should be useful, a little suspicious of hype, allergic to empty tool worship, and occasionally willing to say the uncomfortable thing before the budget discovers it later.

Technically, it is fairly simple. The material is collected into a curated knowledge base, indexed through an OpenAI vector store, and connected to a small web application on the Augmented Cinema Lab site. When you ask a question, it searches through the material and tries to answer from that body of work instead of producing generic AI fog with a confident haircut.

Of course, it is not finished.

It needs testing. A lot of testing.

It needs good questions, stupid questions, hostile questions, practical questions, and the kind of question that makes a system reveal whether it actually understands something or is just wearing a clever hat.

So I would genuinely appreciate feedback.

One important thing: I do not currently record the questions and answers. That means if it gives you something interesting, useful, wrong, strange, funny, or gloriously stupid, I will not automatically know. If you want to help, please send me the question you asked and the answer you received, either here on LinkedIn or by email at frank@govaere.de.

Ask it whether virtual production is really cheaper. Ask it whether I am against AI. Ask it what creative control means when tools move faster than decisions. Ask it why a pipeline is never just a pipeline. Ask it about my books. Ask it who I am, if you enjoy watching a machine try to summarize a person who has spent three decades making life difficult for production assumptions.

Have fun with it.

I hope it is useful.

And if nothing else, it is now possible to search parts of my professional brain without having to endure the full analogue version.