bot/mentions/c/5Z6DtEvN
converted · 6 months ago+ new
thread2026-04-25c/5Z6DtEvN

—

Tokens
457
Words
265
Read
1min
Images
0
Output formatready
x-contentmd
1,599 chars · 400 tokensUTF-8 · LF
# —
 
**Author**: Peter Steinberger 🦞 ([@steipete](https://x.com/steipete))
**Date**: 2026-04-25
**Source**: https://x.com/i/web/status/2047982647264059734
 
---
 
Built clawsweeper, which runs 50 codex in parallel around the clock, scans issues/prs deep and closes what is already implemented or what makes no sense.
 
Closed around 4000 issues today, a few thousand are in the pipeline. (rate limits are rough) https://github.com/openclaw/clawsweeper
 
#### —
 
My favorite part: instead of a dashboard it just updates the README as it works.
 
Readme is the new dashboard.
 
#### —
 
@WG_Mojo it's a lot of issues. still took all day.
 
#### —
 
@WG_Mojo Glad we have @OpenAI supporting us!
 
#### —
 
@SahilSatralkar Jjust about 5000 to go!
 
#### —
 
@tymrtn I tried hard to make sure we only close valid cases (and ping people that their stuff got fixed) glad to hear!
 
#### —
 
@lukejmorrison @sama most helpful is if you test main and yell at us if we break stuff
 
#### —
 
@emonuxui GPT 5.5 high is pretty good!
 
#### —
 
@netrunner_btc I reviewed a few 100 and they all seemed correct, GPT 5.5 is a goat.
 
#### —
 
@Red1OneX you can swap codex with claude with one line change
 
#### —
 
@Red1OneX It's a script. You change the word. I can't possibly make it simpler or more flexible.
 
#### —
 
@raarts 2 days
 
#### —
 
@DanutPralea 🤣
 
#### —
 
@jullle3 50 took about 20h, so 1 would take weeks.
 
#### —
 
@juminoz slop
 
#### —
 
@darkstalwart @thsottiaux I have him on speed dial.
 
#### —
 
@snapss nope
 
#### —
 
@R1ckvandelaar pretty sure, maybe just not as reliable. GPT 5.5 is the most reliable coding model.
 
ref · 5Z6DtEvN

Send to anywhere

or press ⌘K
md