And at the end of the day, does it matter?
But as a non-native english speaker, I do use AI to help me formulate my thoughts more clearly. Maybe this is off putting? :)
The non-native speaker point is understandable, of course, but you're much better off writing in your own voice, even if a few mistakes sneak in (who cares, that's fine!). Non-native speakers are more than welcome on HN.
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
Comment 2: https://news.ycombinator.com/item?id=46873809 2026-02-03T17:13:40 1770138820
Comment 3: https://news.ycombinator.com/item?id=46873820 2026-02-03T17:14:25 1770138865
All detailed comments in different threads posted exactly 45 seconds apart, unless the HN timestamps aren't accurate.
That's very impressive if the account is not "generated comments", even using speech-to-text via AI. I'll leave it at that.
In practice, I've found the economics work like this:
1. Code generation (boilerplate, tests, migrations) - smaller models are fine, and latency matters more than peak capability 2. Architecture decisions, debugging subtle issues - worth the cost of frontier models 3. Refactoring existing code - the model needs to "understand" before changing, so context and reasoning matter more
The 3B active parameters claim is the key unlock here. If this actually runs well on consumer hardware with reasonable context windows, it becomes the obvious choice for category 1 tasks. The question is whether the SWE-Bench numbers hold up for real-world "agent turn" scenarios where you're doing hundreds of small operations.