Not showing any code, but after working with SDAs ...
# community-showcase
v
Not showing any code, but after working with SDAs as the core abstraction for quite a while now, I've observed a subtle but important shift in how I think about my job. I wrote my thoughts on this substack article: https://askvinnie.substack.com/p/now-youre-thinking-with-assets Would love to discuss!
daggy love 9
❤️ 3
s
What if developing a use case could be about creating new physical objects?
Yes, this is the effect. It's like a mental rotation from process-land to artifact-land. But the physicality of assets is really what makes it so much more different and so much more mentally scalable. nice piece!
Brainstorm: Assets are physical things. You could eat an asset. They are nourishing, unlike the fluffy abstractions most frameworks encourage you to write. Abstraction --> concretion
Let's not talk about "machine learning frameworks". Enough with the methods. Give me the pickle. Tell me where it lives. I want to devour it.
pickle 4
y
amazing writeup!! so glad to see dagster was able to lead this mindset shift daggy love as an engineer who works with data all the time myself, one thing i realized was when i’m designing the “data pipeline” in a traditional ETL way (or even just saying things like “from external sources to s3 and then to snowflake”), the nodes that show up in my mind actually have always been those data itself (external sources, s3, snowflake, etc — aka assets), not the task. we were complicating our thinking and, as you perfectly pointed out, our communication too, by writing the pipeline where the nodes are the movements/tasks not assets. loved that you and more folks start to feel the same way — i.e. just let me focus on those data assets that are in my head in the first place, and let a tool figure out the rest for me.
v
I wonder what an asset would taste like. Can we get a sensory panel assembled?