
5 minutes read
September 26, 2022
Batou.xyz: Where creatives, ideas and AI meet
CEO Pol Baladas on AI in the creative process and intelligence augmentation
What were the origins of Batou.xyz?
Pol Baladas started coding when he was young, having been inspired by the computing pioneers of the 60’s and the way that their technological inventions were driven through ‘prototyping and just trying stuff out’.
Along with a team of AI researchers, UX specialists and Software Engineers, Pol founded Batou in as a creative studio founded on these experimental principles combined with a frustration with the limitations of existing tools.
“The quality of the human computer interaction for AI has generally been out of sync with the incredibly enabling capabilities of machine learning - we wanted to change that!”
What is intelligence augmentation?
The company’s mission statement is to;
“create products to think, work & learn, individually or collectively in order to augment human intellect.”
Pol explains that the use of AI is not intended as a replacement for human creativity but as a way to 'push the limitations of what humans are capable of'. The team are delivering cutting edge projects and interactive toolsets based on ‘guiding design principles’ to help users generate new ideas, and to enhance the creative process.
Pol adds that this approach has been the main driver behind their flagship project Fermat.ws although it has also led to some ‘really wacky stuff’ like an attempt to render Wikipedia in VR!

How are Batou working with Pipeline.ai?
As part of the Fermat platform, Batou have a toolmaking function so that technical users can add their own code and build their own features. The Batou team also use these tools to iterate the product and to build more features on top.
Pol describes how the ecosystem is evolving fast and how the creative companies (like Batou) need the ease of mind and assurance that the models are going to work so they don’t have to worry about the infrastructure of putting models into production.
He explains;
“We were using GPT3 and we wanted to move to more models like DALLE and then Stable Diffusion came out and we spent around 3 days trying to deploy the model and really struggling, and then I was like STOP! Because it’s not our focus, right? Our business is to create amazing and creative tools for people so let's just call Mystic and get an endpoint to the API… and so then we deployed in the same day and it was up and running and awesome!”
Where is all this going?
Pol is optimistic and upbeat about the future of AI as an integral part of the creative process, despite not being entirely clear about how it’s all going to play out. Like many companies that are innovating in this space it can be a thrill, if somewhat daunting, to be part of a market that is evolving so fast;
“There’s been a whole lot of investment going into text based LLM’s, and text to image models; the speed of change is incredible and we are only just getting started! We can’t predict the future, but what we do know is that the models are going to get better and get more able - so we all need to be ready to translate that into a creative opportunity for end users.”
And the infrastructure is key. As Pol explains, for a creative person it’s so important that you optimise the creative moment – in the real world the artist doesn't have to wait to make a brush stroke and it’s equally important in the digital world when you are using generative AI; 'speed is of the essence'!
In terms of the market, Pol’s passion for human computer interaction and his vision for Fermat.ws goes beyond the creative users;
“The scope of this kind of AI application is potentially huge - the challenge that we ask ourselves is; how can we take all these generative ai models and make compelling user interfaces for users in every vertical - how can we make it universal?”
Watch this space - Batou.xyz!

#FERMAT
https://fermat.ws Fermat is your personal multimedia space where you can build your own tools. In Fermat, you can compose in any media (hypermedia, text, images, buttons) within an infinite spatial canvas. The clever bit is that the text and images are being generated from text prompts to Large Language (DALLE and Stable Diffusion) and Text to Vision Models (GPT etc) – also there’s a creator community where you can see examples and exchange ideas with other users.
#Augmented desk
https://batou.xyz/content/augmented-desk Augmented Environments are physical spaces where people can work and learn, collectively or individually, by using day-to-day tools such as pen and paper, or other quotidian objects.This affordable and extensible system augments the user’s capabilities without being invasive, recognising their natural language and humane behaviors.
#Arbor
https://batou.xyz/arbor/ At batou, arbor is used for human-computer interaction research and “augmented behaviour” programming. Arbor is an interpreted lisp-like language. It can also be hosted in Java, Kotlin and Javascript systems as a library.
#Explorer
https://batou.xyz/content/explorer Visual knowledge tool that will allow you to experience information in a completely new perspective.Explorer interface improves the understanding and experience of concept searching.
Follow Batou/Fermat on Twitter, on linkedin and Instagram
Try out Large Language (DALLE and Stable Diffusion) and Text to Vision Models (GPT etc) at Playgrounds.ai.