Thread by @Elijah_Meeks

There's a lot of innovation happening in the #notebook space these days, with notebook-focused startups (like my own @noteable_io) and notebook products from existing companies both presenting more and more visions of what the computational notebook can be.
1/8

I put the different notebook approaches in three categories:
1. Notebooks as a way of selling other services
2. Notebook++
3. Notebook as the modern data collaboration tool
2/8

Notebooks are a great way to sell storage or compute. They're not particularly innovative in UI/UX because it's all about integrating with the compute or storage or other product you sell. This includes Notebook bolted onto an app just so you don't leave the product.
3/8

Notebook++ is a different approach. It sees notebooks in their traditional use case (data scientists doing data science) and tries to make all those things easier. Connecting to models, easing datasci workflows or enabling non-data scientists to do more datasci-like work.
4/8

The last one, which we're focusing on here at @noteable_io, is looking at the future of data collaboration and saying notebooks are the best basis for how to enable it. Notebooks can take on features of collaborative docs (like commenting) or BI tools (like no-code viz)...
5/8

...then you can enable data teams to collaborate within the same tool to produce meaning from data. This is the most exciting future for notebooks, in my mind, because you have all the power of code and no-code optimized not for lone data science heroes but for teams.
6/8

Some kind of future document for data teams is coming. The question is whether it will be built out of a code-first environment with UI/UX affordances or whether it will be a GUI-first environment with code escape hatches (like the scripting available in BI tools).
7/8

I think the former is going to do more to empower folks who don't have a strong coding background because coding isn't an obscure escape hatch to do weird things but rather a first-class citizen in the app. And as we all know, data work eventually requires at least some code.
8/8