Can AI determine if your script will make a successful film?

When Quilty went on sale earlier this year, the AI ​​startup promised that its tool could accurately predict a film’s success just by reading the script. When people found out an opportunity to experiment with them Quilty’s products, however, were left in doubt. Even with all the data available in the world, it predicted the records of Christywhich will continue to be box office, it can go beyond the script Sinnerswho became an Oscar winner.

As many AI leaders have already started, the founders of Quilty believe that it can help “democratize” their companies by providing up-and-coming and up-and-coming users with tools to help them – a good score for Quilty, perhaps, can be with the developer, and a slight decline can be a sign that some improvement is needed. But right now, Quilty is just a complication of existing AI systems, and the company has yet to prove that its technology has the taste or analytical ability to detect future hits (let alone be proven).

It was founded by filmmakers Simon Horsman and Daniel WoodQuilty uses AI to analyze documents and generate detailed reports about project opportunities. After being fed raw material, Quilty’s technology assigns a score from 0 to 100 that indicates the type of story the project is expected to be, its commercial viability, whether it resonates with the audience, and how much the production would cost. The platform is selling an idea that can give users a glimpse of the future when they are trying to get their movies/films greenlit. Horsman and Wood believe that Quilty is about to become an integral part of the way traditional studios do business.

When I recently sat down with Horsman and Wood, they insisted that they want to “keep people safe” rather than create for themselves. When they first launched their company, Horsman and Wood solicited feedback from several other experts who often complained about the potential for gen AI to disrupt jobs and leave workers without skills.

“We agree with a lot of negative thinking about AI, but what we’re trying to do is get people to create things,” Horsman told me. “Quilty is really about development and giving users – be it a writer, a designer, a buyer, a financier, or a studio operator – as much information as possible to make a green lighting decision.”

Rather than providing users with a single, bespoke AI solution that provides feedback on content, Quilty integrates a number of widely available AI tools to deliver multi-dimensional analysis. All users have to do is upload their documents to the platform, and after a few minutes, it produces a report that explains things such as estimated budgets, descriptions of important news hits, and personal analysis. The service costs $50 for each individual review, but you can also purchase multiple reviews at a lower price.

The idea for this kind of analytics first came to Wood – who also serves as Quilty’s CTO – a few years back when he was being sued in a business case. Instead of spending money on a lawyer, Wood fired ChatGPT, who quickly told him “I’m not a lawyer; go find someone else to help you.”

“Then I went to Gemini, which worked really well for a while, because I had a big window of experience,” Wood said. “But then I was at X, and he saw Elon Musk as a fool we were talking about Grok getting the best lawyer for AI, and I was like, ‘Let me do the research.’” (Wood did not elaborate on how the legal dispute unfolded.)

The experience left Wood with a better understanding of how similar types of AI can excel at different tasks. And Wood’s use of AI has informed Quilty’s method of calculating script success. Because “Gemini is so good at structure and form,” Quilty uses that tool to help break it down — notes that put an entire film or scene into a complete sequence. In the financial model, the company puts faith in the model of DeepSeek which is stored on servers located in the US. And for description/feature analysis, Quilty uses a combination of Claude and ChatGPT.

Wood told me that the company relies on content creation — a process where you provide additional information — to create quality releases that aren’t filled with facts. Quilty does not self-teach any of the models it uses to create video/information reports. But Wood insisted that it was a strength rather than a weakness because it made it easier for Quilty to incorporate new and cutting-edge styles into his designs when they became available to the public.

“When Claude Mythos comes out and I see that it’s a better LLM, all of a sudden, all of my programs are better,” Wood said, referring to it. a new dynamic model available to a small group of organizations for cyber security purposes. “If some Chinese brand is going to be better than all the US border brands, why not use them instead?”

While the flexibility of Quilty’s technology may make it very quick to adapt to all the changes, it also makes it difficult to understand how the platform takes the script and comes up with metrics that measure intangible things like the audience. strength tan is a video that doesn’t exist. Forecasting has been an important part of the development of movies since the birth of Hollywood, but this work has always been done by professionals who understand the audience well.

No AI company has been able to create a model that can match the human mind or the way we think about art. But the founders of Quilty think that their “engine” is the next best thing to search for documents because of the way it integrates. weapons like VADER (Valence Aware Dictionary and sEntiment Reasoner) — open source software that measures the number of words that come across as positive and negative.

Horsman and Wood are also confident that Quilty can accurately determine how well a project “supports the traditional period” and provide reliable box office estimates. They pointed to Revenge of the Nerds as an example of a classic old film that would receive a low Quilty score mainly because of its setting they try to express sarcasm in a humorous way – something that modern viewers would find out of hand.

When I asked Horsman and Wood why Quilty offered Christy (which eventually raised about $2 million) high score than Sinners (which grossed $370 million), they insisted that the platform decision “came from the fact that Sydney Sweeney is really, really famous.” They said that on paper, Sweeney’s star power combined with the fact that dramas about boxing are cheaper to make than fantasy/events like. Sinners produced Christy safe bet. But this shows how unreliable Quilty’s theory is. Horsman and Wood admitted that sometimes Quilty could not foresee things that would affect a film’s financial performance or audience reception.

For example, Quilty would not have expected that Elijah Bynum Magazine Dreams (which Horsman created) can be messed with by the player The legendary Jonathan Majors is falling from grace in 2023. Similarly, something about Minecraft video‘s script would display on the platform that Chicken Jockey phenomenon It would be part of the monster success of this film. Horsman and Wood told me that, eventually, they want Quilty to be able to see this kind of thing coming, but it’s hard to imagine how that would be possible.

For all his interests, Quilty’s sales revolve around a variety of major languages ​​that are being asked to predict the future as it affects the creative arts. It would be surprising if any of these AI tools could work as well as Quilty claims. But most of them are advanced recognition/imitation machines that are too far to understand what people enjoy.