Learning about an experimental new #rustlang SDK for Beam #BeamSummit
Talend talking about their improved LocalRunner using Reactive Streams! Very cool. #beamsummit
okay, so what GPU are we using for model training these days? I might finally get around to doing this ITX build.
What’s going to be kinda cool is that we’re going to be able to use LLMs to rapidly iterate on API ergonomics. If people are writing code using LLMs then it makes sense to optimize APIs for getting good results from LLMs. Question is how well that moves to humans.
So, I got to thinking… it be nice to have something like dbt for managing large Beam pipelines, especially multi-language ones… Finally got a basic multi-language test working last night. Behold, spd:
github.com/byronellis/beam/b…










