One thing I've learned this year is you can win a lot just by hanging with your bestie and being yourself #teambestfriends #gotree
March 28, 2025 at 12:48 AM
One thing I've learned this year is you can win a lot just by hanging with your bestie and being yourself #teambestfriends #gotree
Recent AI models excel at MirrorCode tasks. Claude Opus 4.6 successfully reimplemented gotree — a bioinformatics toolkit with ~16,000 lines of Go and 40+ commands. We estimate this would take an unassisted human SWE 2–17 weeks.
April 10, 2026 at 3:22 PM
Recent AI models excel at MirrorCode tasks. Claude Opus 4.6 successfully reimplemented gotree — a bioinformatics toolkit with ~16,000 lines of Go and 40+ commands. We estimate this would take an unassisted human SWE 2–17 weeks.
AI can solve many long-horizon MirrorCode tasks. E.g., Claude Opus 4.7 passed 99.95% of tests when reimplementing gotree — a bioinformatics toolkit with 16k lines of Go and 40+ commands.
We estimate this would take an engineer 2–17 weeks. Opus 4.7 solved it in 14 hours for $251.
We estimate this would take an engineer 2–17 weeks. Opus 4.7 solved it in 14 hours for $251.
June 26, 2026 at 3:25 PM
AI can solve many long-horizon MirrorCode tasks. E.g., Claude Opus 4.7 passed 99.95% of tests when reimplementing gotree — a bioinformatics toolkit with 16k lines of Go and 40+ commands.
We estimate this would take an engineer 2–17 weeks. Opus 4.7 solved it in 14 hours for $251.
We estimate this would take an engineer 2–17 weeks. Opus 4.7 solved it in 14 hours for $251.
How do we know that it's actually doing a cleanroom re-implementation?
It doesn't have internet access, so it can't *search* for it, but the original gotree code has been on the internet for a decade and almost certainly ended up in Claude's training data.
It doesn't have internet access, so it can't *search* for it, but the original gotree code has been on the internet for a decade and almost certainly ended up in Claude's training data.
GitHub - evolbioinfo/gotree: Gotree is a set of command line tools and an API to manipulate phylogenetic trees. It is implemented in Go language.
Gotree is a set of command line tools and an API to manipulate phylogenetic trees. It is implemented in Go language. - evolbioinfo/gotree
github.com
June 26, 2026 at 8:44 PM
How do we know that it's actually doing a cleanroom re-implementation?
It doesn't have internet access, so it can't *search* for it, but the original gotree code has been on the internet for a decade and almost certainly ended up in Claude's training data.
It doesn't have internet access, so it can't *search* for it, but the original gotree code has been on the internet for a decade and almost certainly ended up in Claude's training data.
Claude Opus 4.6 reimplemented gotree, a 16,000-line bioinformatics toolkit, from its specification alone. No source code. New benchmark from Epoch AI and METR. The catch: it works because gotree has a checkable spec. Most scientific software does not.
hallway.aris.pub/no/004/
hallway.aris.pub/no/004/
No. 004 - The Hallway Track
The Hallway Track No. 004 - How AI is affecting the practice of science
hallway.aris.pub
April 21, 2026 at 7:33 AM
Claude Opus 4.6 reimplemented gotree, a 16,000-line bioinformatics toolkit, from its specification alone. No source code. New benchmark from Epoch AI and METR. The catch: it works because gotree has a checkable spec. Most scientific software does not.
hallway.aris.pub/no/004/
hallway.aris.pub/no/004/
I love how it looks a little like a pair of legs and hands on hips with all the white space right now. GOTree is not amused 😆
June 25, 2025 at 6:12 PM
I love how it looks a little like a pair of legs and hands on hips with all the white space right now. GOTree is not amused 😆
New MirrorCode benchmark shows AI can reimplement complex programs like pkl or gotree in hours for $100-$400, tasks that would take humans weeks. 8 of 25 programs remain unsolved at 100%.
Source: Import AI
Source: Import AI
Import AI 466: The bitter lesson for robotics, AIs complete week-long programming tasks; and OpenAI's accidental AI hacker
The warning shots will continue until civilization wakes up
importai.substack.com
July 27, 2026 at 4:30 PM
New MirrorCode benchmark shows AI can reimplement complex programs like pkl or gotree in hours for $100-$400, tasks that would take humans weeks. 8 of 25 programs remain unsolved at 100%.
Source: Import AI
Source: Import AI