Post

We're delighted to have contributed case studies to OpenAI's report on scientific computing in the age of agentic AI. 📃 hubs.la/Q04r75Zm0 Alongside Felix Krueger (Altos), we contributed agent-assisted rewrites of RNA-seq QC tools (RustQC, FastQC-Rust & more), built for @nextflow.io and @nf-co.re.

hubs.laScientific computing in the age of agentic AIA new field report shows how scientists use AI coding agents to modernize scientific computing, accelerating software development and discovery in genomics and beyond.

29/7/26, 12:07 pm1 repost6 likes

Ethan Hetrick

@ethan-hetrick.bsky.social

31 July

Crazy how fast that all the open source tools we've used for years are all being rewritten over the course of weeks by agentic systems. Soon, most tools we use will have minimal human element.. Not necessarily saying it's a bad thing, I use agents too It's just uncanny

Seqera

@seqera.io

29 July

We also contributed to a case study on faithful rewrites, foundational primitives, and the limits of agent self-verification, specifically related to an upcoming, significantly accelerated version of the STAR aligner.

Reid Marlow

@reidmarlow.com

29 July

The useful boundary in those rewrites is where the agent stops being trusted. Porting old scientific code is one thing. Proving STAR-like behavior against real read outputs is what keeps it science instead of a very fast refactor.