Skip to main content

Live from the build loop · updated Wed 12:15p Central

Shipping right now.

I run an agent pipeline that turns issues into reviewed pull requests across my products. Every row below is a real build, with what it cost and how many tries it took. What each change does stays private; that the work happened, and how, does not.

PRs merged, last 7 days
456
agent builds shown
40
verified on the first try
42%
median compute per build
$1.18
  • Wed 12:05pIdynicAdded a new view of changesfeatureopus20 tool calls3 test runs≈$0.62revising
  • Wed 11:40aIdynicAdded a new view of changesfeatureopus28 tool calls1 test run≈$1.00revising
  • Wed 11:51aOnrunbookShipped a featurefeatureopus54 tool calls3 test runs≈$2.85revising
  • Wed 12:00pIdynicUpdated wording on a main pagecopyopus70 tool calls5 test runs≈$2.81merged
  • Wed 11:30aIdynicAdded a data viewfeatureopus88 tool calls1 test run≈$3.27revising
  • Wed 11:51aIdynicImproved check-in content for usersfeatureopus109 tool calls1 test run≈$4.28verified
  • Wed 9:40aOnrunbookUpgraded a dependencyupgradeopus60 tool calls5 test runs≈$2.65verified
  • Wed 3:34aBuild loopRan the build loop's self-testself-testhaiku14 tool calls1 test run≈$0.11verified · try 2
  • Wed 3:33aBuild loopRan the build loop's self-testself-testhaiku33 tool calls2 test runs≈$0.23failed
  • Wed 3:32aBuild loopRan the build loop's self-testself-testhaiku19 tool calls2 test runs≈$0.18revising
  • Tue 11:50pIdynicFixed a generation bugfixsonnet44 tool calls3 test runs≈$0.82verified · try 2
  • Tue 7:20pIdynicShipped a redesignredesignopus70 tool calls7 test runs≈$2.31verified · try 2
  • Tue 6:55pIdynicShipped a featurefeatureopus57 tool calls6 test runs≈$2.14verified · try 2
  • Tue 5:55pIdynicRedesigned a settings screenredesignopus121 tool calls5 test runs≈$4.55revising
  • Tue 5:30pIdynicRedesigned settings navigationredesignopus235 tool calls3 test runs≈$10.03revising
  • Tue 2:45pIdynicPolished mobile screenspolishopus77 tool calls2 test runs≈$2.28verified · try 4
  • Tue 1:40pIdynicShipped a redesignredesignsonnet4 tool calls0 test runs≈$0.14revising
  • Tue 1:20pIdynicPolished mobile screenspolishsonnet62 tool calls2 test runs≈$1.05revising
  • Tue 1:45pIdynicMade generated content more consistentfixsonnet46 tool calls2 test runs≈$0.72verified
  • Tue 1:05pIdynicPolished screen layoutspolishsonnet54 tool calls4 test runs≈$1.21revising
  • Tue 12:20pBuild loopRan the build loop's self-testself-testhaiku11 tool calls1 test run≈$0.09verified · try 2
  • Tue 12:19pBuild loopRan the build loop's self-testself-testhaiku24 tool calls2 test runs≈$0.17failed
  • Tue 12:19pBuild loopRan the build loop's self-testself-testhaiku19 tool calls1 test run≈$0.19revising
  • Tue 3:33aBuild loopRan the build loop's self-testself-testhaiku25 tool calls1 test run≈$0.15failed
  • Tue 3:32aBuild loopRan the build loop's self-testself-testhaiku17 tool calls1 test run≈$0.15verified
  • Tue 7:05aIdynicMajor toolchain upgradeupgradeopus85 tool calls1 test run≈$2.06verified
  • Mon 9:40pIdynicRewrote key copy wordingcopyopus54 tool calls4 test runs≈$1.15verified
  • Mon 9:00pIdynicRedesigned a main screenredesignsonnet159 tool calls1 test run≈$2.96verified · try 4
  • Mon 8:55pIdynicRedesigned a key screenredesignsonnet48 tool calls0 test runs≈$1.03verified · try 2
  • Mon 9:05pIdynicCorrected a data list issuefixopus49 tool calls2 test runs≈$1.92verified
  • Mon 7:40pIdynicRedesigned a main screenredesignsonnet20 tool calls0 test runs≈$0.37revising
  • Mon 7:40pIdynicRedesigned a main screenredesignsonnet63 tool calls4 test runs≈$1.92revising
  • Mon 8:35pOnrunbookAdded an approval stepfeatureopus43 tool calls1 test run≈$1.25verified · try 5
  • Mon 7:06pOnrunbookAdded an approval stepfeatureopus41 tool calls5 test runs≈$1.56revising
  • Mon 7:15pIdynicRedesigned a main screenredesignsonnet63 tool calls9 test runs≈$1.18revising
  • Mon 7:50pIdynicRedesigned a main screenredesignopus183 tool calls7 test runs≈$18.54verified
  • Mon 6:30pIdynicRedesigned a main screenredesignsonnet79 tool calls3 test runs≈$1.45revising
  • Mon 6:20pIdynicUnified design system stylingdesign-systemsonnet26 tool calls2 test runs≈$0.62verified · try 2
  • Mon 7:00pIdynicPolished screenspolishopus13 tool calls0 test runs≈$0.39verified · try 2
  • Mon 5:20pIdynicUnified the design systemdesign-systemsonnet65 tool calls0 test runs≈$0.95revising

Hover to pause. Failures and send-backs stay in the feed.

How the loop works

  1. 01

    An issue goes in

    I write the issue and mark it ready. That's the only way work enters the loop.

  2. 02

    An agent builds it

    Within minutes an agent picks it up in an isolated branch, writes the change and its tests, and opens a pull request.

  3. 03

    It gets checked

    CI is the gate, and a separate review pass checks the work against the issue. Work that falls short goes back for another try.

  4. 04

    It ships, or comes back to me

    Verified work ships. Anything that needs a judgment call comes back to me instead of being guessed.

What stays private

  • Titles are rewritten into generic summaries from a fixed vocabulary before they are published, then checked again by this site.
  • Stealth products show only that a pull request merged, and when.
  • No code, issue text, customer or repository detail is ever published.

Want a loop like this for your team?

Book an intro call