Hacker News · item history
Running local models is good now
558
first seen points
1,535
peak points
1,535
latest points
11
observations
| Posted by | jfb |
|---|
Posted on X
- Ran a 7B model locally for a week on actual agent tasks (200+ calls/day). Median latency: 340ms. Cost: $0. No rate limits, no API outages, no "your request was flagged" errors. The tooling finally caught up to the promise.