Really? What did you not like about Helix that you get from Nano? (No judgement, I'm just curious!)
I don't have anything against Nano, but for something I use every day, I'd really miss go-to-definition, and being able to jump between sections of text quickly.
It’s too bad nano doesn’t support LSP [1], it makes it trivial to get IDE-level support for almost all programming languages with one simple JSON-RPC integration.
Pardon the snark, but they literally wrote an article [1] about giving up on code reviews because they push 30k _changes_ (I assume that means commits, but they don’t clarify), and were bottlenecked on reviews.
Instead, they claim to just review the _intent_ of the PR and presume correctness will be caught by more AI.
To me, the more absurd part is that they’re pushing that much code.
To put it in perspective:
Software # Commits Equiv. Time
--—----- --------- -----------
Nginx 8,758 7h
uv 11,081 9h
Vim 24,608 20h
Go 67,914 2.3d
Git 82,513 2.8d
VS Code 168,140 5.6d
Emacs 183,611 6.1d
Rust 343,063 11.4d
* Equiv time equals the amount of time it would take Anthropic to push that many commits, given their 30k changes/day estimate provided in the article linked below.
It generated some interesting discussions here on hacker news [2] [3].
FYI: The last is a blog post at simonwilliamson.net about the source post, which was flagged appropriately, but I left since it had a few dozen comments as well.
I saw Jev and without understanding it said to myself, “I can build that!” And brainstormed some weird Alex trebek jeopardy generator I called trebek, a bun ran typescript app that you give it an input, it decides if it’s a category, question or answer then generates what’s missing. Trained a SQLite-vec database on the English language for a few days with a small qwen embedding model to add vec embeds to the db then ripped the cord on the embeddings before I started feeding my llm the proper specs for Jev and now i have a cool jeopardy generator that now doubles as a Jev clone classifier with a custom v1 endpoint for system 0 or whatever it is. Level understanding here, fun experiment though ^>^ and produces useful outputs
That sounds like a lot of fun, love the creativity. Do you have a repo you can share? It would be nice to run it, and if you wanted, maybe I could add it to the demos on https://playground.jeffyclassify.com
Ha. I did something similar but made a Jeopardy judge for my family to play jeo-paw-dy hosted by the family dog. It uses our phones as buzzers and speech to text to let people just speak the answer
A good benchmark would require a decent number of smaller scoped one off tasks to larger multi step refactors, and also one shot full project of simple to complex varieties. In addition to a series of “conversational” ambiguity filled one-liners.
reply