Welcome to eurAIka!
We’re standing on the shoulders of giants in building eurAIka. It started as a proof-of-concept project to determine how easily the components could integrate into the open-source note-taking app Obsidian. With hundreds of built-in and community written plugins, we were able to find the ones to make eurAIka possible.
I began building eurAIka because while writing articles for our blog Wild Peaches I was generating quite a lot of code which resulted in at least two problems. The first is that scientists aren’t software engineers. We primarily write code to get answers and making it beautiful is a distant second requirement.
I had a professor in grad school who explained to me that the way to write code was to assign the letter “a” to the first variable, “aa” to the second, then “aaa”, and so on. Documentation was unnecessary because the code was self-explanatory, and everything should be written on a single line because putting instructions on separate lines and including any white space just slowed it down.
I try to write clearer code than that, but I’m sure there’s room for improvement. The second problem I often face is that there isn’t one software tool suitable for all problems. We may have a preferred language that we use to write most of our code, but sometimes a special purpose tool is needed for a particular problem. Instead of solving the problem you’re most interested in, you must learn to use new software tools.
This is where the new large language models can really help. You can interact with them as you write code to help with syntax, or even have them write complete sections. The goals of eurAIka are even loftier, and I’ll have more to say about that in a later post, but for now, it’s possible to get a lot done with AI assistants like Jupyter AI.
Scientists aren’t trained to be great communicators, either. Carl Sagan and Niel deGrasse Tyson are wonderful counter-examples, but in general we’re first interested in solving the problem and then getting the word out. New AI packages are now available to help organize thoughts and generate clear and concise papers.
These AI platforms are available in stand-alone form, but I thought it would be easier to focus on the research problem if all of the tools were available in a single user interface. The Coder is mostly intended for working in a JupyterLab notebook environment, but you might sometimes want to use Geogebra, Tracker Physics, Google Colab, or some other special code. The Coder gives you one-click access to these and more, most of which will open in a new tab.
Originally named “The Magic Whiteboard”, the construction of each component is documented on Wild Peaches where you can see how we build The Whiteboard, The Librarian, and The Coder. While eurAIka isn’t yet ready to download as a single package, refer to the posts on Wild Peaches to get started.
By combining AI assistants into an integrated platform, eurAIka streamlines literature review, writing, programming, and collaboration. Spend less time on logistics and more time making discoveries.
In the future, posts about eurAIka will appear here, and STEM projects will be hosted on Wild Peaches, but they will continue to be symbiotic. As needs arise in Wild Peaches we’ll add new capabilities to eurAIka, and if we learn interesting science while building eurAIka it might appear in Wild Peaches.