Pickford Dev Blog — 002

We are continuing to develop Pickford Studio and we’re working towards having the software in a state where we can include exciting feature changes in these blogs.

The past 2 weeks, our work has been focused on making development easier for ourselves and ensuring stability for any current features. Getting these 2 things under control gives us a solid foundation to accelerate bringing new pieces to our tools and letting users truly run wild!

With that out of the way, this entry should focus on the audience, not the creators. We’ve talked a lot about the tools and only hinted at our views on what it will feel like to experience stories in an AI powered world. It’s a bit of a nuanced and ill-defined topic… but let’s give it a try!

Introduction

Innovations are often built on hypothesis. We’ve taken a hard look at the state of storytelling and made a few assumptions to set guardrails for our work. These might not always be correct, but that’s alright since our task is to strategically validate or invalidate these understandings.

Our current theory is that experiencing stories sits on a spectrum of audience agency. Video games and tabletop games on one end and movies, TV shows, and books on the other. Somewhere in between, you have choose-your-own-adventure books and shows like Bandersnatch, but ultimately these are still limited in the kinds of stories they can tell.

This line has been tested over the years in various ways. American Idol, although a show, allowed viewers to call and text to vote — granting the audience agency in a unique way. Video games themselves, drift along the spectrum with fairly linear stories or open worlds and countless, complex tales.

Our ultimate question, and the thing that keeps us up at nights, is this: if generative AI continues to get cheaper and faster, is there a new medium of storytelling possible? And if so, where, if at all, does it fit on this line?

The Conundrums

We run into a few issues as we plan our path towards attempting to answer this line of questioning.

• What does this new experience feel like?
• How do we ensure it still feels like a story?
• How do we guarantee creators can craft their narratives?

Our opinions on current approaches of using AI in storytelling, without prejudice, is that they fail on all 3 of the above issues.

Their experience is still the same as watching a show, or short trailer, with their only uniqueness being to create versions of something that hasn’t be seen before (e.g. Friends but Chandler creates an NFT, Family Guy based in the future).

Additionally, LLMs are bad at telling coherent stories. Even with proper prompting, they don’t do well with automatically and consistently coming up with story beats and meaningful endings.

Current approaches mostly focus on ditching the creator in every aspect except the initial prompt. They aim to automate any controllable part of the pipeline in favour of time-to-half-decent-outcome. This lack of control manifests itself in different ways depending on the medium — AI written books lack unique dialog and generated video will always be “good enough” compared to what the creator saw in their head.

This isn’t to say we don’t see AI being used as a piece of the pipeline that helps in specific ways. But ultimately, anyone making interesting attempts at story telling fall into 2 buckets.

• Using AI sparingly, ensuring they maintain control over their creation.
• Creating variations of previously defined narrative mediums.

What is the reason for these traps being commonplace? It’s easy to stick to what we know. Finding new approaches and breaking out of predefined molds is never easy.

But if it was easy, it wouldn’t be fun.

Pickford’s Upcoming “Show”

So where do we go from here? How does Pickford attempt to find this new, mythical way of telling and experiencing stories?

Well, we’re working on a show of sorts! The tech behind this thing is being built from the ground up as well, a fully dynamic storytelling engine. Designed as our first attempt at breaking out of what mediums currently exist.

Coming back to our 3 issues from earlier — we’ve focused on ensuring that what plays out still feels like a story and that creators have the ability to define the world, characters, and plot lines to properly craft their narratives. What we don’t fully understand yet, is what it will feel like interacting with it… but we are incredibly excited to see what people think and how it changes over time.

Creation is being done in conjunction with some absolutely brilliant creators from the film and gaming industry to ensure diverse and experienced opinions are taken into account.

We are still early days, but with our toolkit being defined and this “show” in the works, we think the next 6 months are going to be a blast. Stay tuned!

Conclusion

Work on our first tele-narrative experience will be finished in the next few months. If you want more detailed updates as development progresses, join our Discord.

Pickford Studio is still in early alpha. If you want to help guide the storytelling tools of the future, join the waitlist!

Onwards and upwards!