Lenny's AI Builders - LAB 0041 the conductor paragraph, the four steps to run it, and the honest numbers from my 40-hour run. i ran my viral Fable 5.2 conductor for 40 hours on my real app. one conductor, 73 worker threads, 25 pull requests merged. and my weekly Fable limit was gone in 13 hours. so here's what it actually built, and the one paragraph that runs yours. why it mattersthis is the new meta. you use Fable 5.2 as the conductor and Opus 5.5 as the worker threads. Fable 5.2 is twice as costly as Opus 5.5. so Fable does the thinking and Opus does all of the work. it's not perfect and it's not god level. but it makes a huge difference to how much you have to hold in your head while you run lots of agents at once. Fable directs the show now. Opus does the work. what you get
the real example: my runi ran one conductor on Eddy Studio, my Mac app, for 40 hours. this was in T3 Code, with my full conductor skill. the paragraph below is the short version of it. Fable 5.2 plans, then gives each task to a worker thread. when the worker is done, Fable checks the work. 25 pull requests got merged. three of them: an update button, click a scene and the video jumps to it, and a feedback card that opens whole. the stats: Fable sent 259 messages to 73 threads. 150 of those went to Opus 5.5. that's 58%. and here's the stuff that wasn't so clean:
i hit my weekly Fable limit 12 hours and 55 minutes after my first Fable 5.2 prompt. about 158 of the 268 turns came after that notice. how i counted: a pull request counts when it links to a thread the conductor ran. i didn't check who wrote every line. 8 of the 25 were opened before the conductor started. the 13 hours covers all my Fable use, not only the conductor. try it
start with three workers, not 73. 73 threads is where mine ended up after 40 hours. it won't spin up 73 all at once. copy this promptadd this paragraph to your job:
that paragraph is the short version. the cleaned-up skill that ran my 40-hour build is in my skills repo, with the rest of my skills. check it against the paragraphmy 40-hour run used my full conductor skill, not this paragraph. so check yours against the paragraph, line by line:
limits and dates
the rest of this LAB
sources
next stepwatch LAB 0041 on YouTube - the practical starts at 12:59. want the next one? get the next field notes. optional - these notes stay free to read. |
i’m Lennox. i sift through AI Twitter and share one thing i tried with ChatGPT. for people making products, content and useful systems. five emails a week, with a weekly option.
Lenny's AI Builders - LAB 0045 one Claude Code setting, 36 to 38% fewer tokens in every pair of my test, the catch on cost, and a prompt that gives you 10 versions. if you build with Claude Code, one new setting can make it use fewer tokens. i tested this myself. every pair of runs used 36 to 38% fewer tokens. but there is a catch. my first run cost 83% more at api prices. the other two pairs came out 15% and 17% cheaper. why it matters the setting turns on Code Mode. Claude writes code that...
Lenny's AI Builders - LAB 0042 the one Grok Bot upgrade that matters, and the paragraph i add to claude.md for long Fable 5.2 jobs. Grok Bot 1.0 isn't new, but the launch post pulled 46 million views like it was. so don't worry, you're not missing out. the one upgrade that's actually decent got something like 40 times fewer views. and it let my bot walk into my Mac. why it matters Grok Bot has been out since August. xAI's own page calls 1.0 a major update to its AI teammates. the headline...
howdy so i let Claude reread six months of my agent logs. (yes, i called it dreaming) that became the centre of this LAB Roundup. my agent's memory went from 7 to 9 Claude proposed 125 edits to the notes. then i ran the 10-question quiz i wrote before the test. the score moved from 7 to 9 out of 10. but i still had to check the work: 54 edits held up, 58 only partly, 2 did not, and 11 had no log day to check. watch LAB 0039 · listen the video skill found 7 of 14 cuts i gave it my best edit...