Mikio L. Braun

ML and AI expert

Back to archive

Thread

3 tweets

1
@lalleal @joelgrus What I am thinking about is how databricks/sagemaker and others are selling notebooks as the one-env-for-everything, making them schedulable, callable, etc. Which is cool, but they are also just a bunch of shell scripts essentially. Living in a file system.
1 1
2
@_fogfish_ @lalleal @joelgrus So what about unit tests, what about code reuse? What about abstractions? What about collaboration. And many DS don‘t have a software engineering background, which makes matters worse in terms of production code.
1
3
@_fogfish_ @lalleal @joelgrus Of course, there are ways, you can put stuff in libraries and then import them, and so on, but I think there is a steep change in complexity if you start doing that. I may be wrong, but that is how it looks to me currently.
1