I used Fable on Ultracode for a larger task for the first time,
Had it review a large build plan for a new feature.
- 37 sub agents
- 2.1M tokens
- Burned through my 5 hour window (I’m on the 5x max plan)
- $40 in extra usage credits
Lmao.
All that to review a build plan.
I’m on baby duty, who is napping in my lap, so I’m glued to my chair.
So I used Hermes to read my repo, find the files I was looking for, pull my most recent meeting transcript, synthesize it all, and formulate a brief for my next engineering build.
What a time to be alive.
@thsottiaux@OpenAI my banked usage resets aren’t working.
I had 3 this morning, used one and nothing changed. Tried another, still no change.
Wondering if there’s an issue? Any help would be greatly appreciated!
@lennysan If a PM is going to venture to do their own analysis they should be held accountable to 1) back it up when it gets grilled and 2) fix it themselves when it’s wrong.
An accountability culture will ensure PMs are only attempting to do analysis they can handle.
In a previous role, Lenny was my PM mentor that didn’t exist in my own org. Thank you @lennysan.
A newsletter has to be pretty damn good to be better than real humans around you day to day.
@stuholden I can’t speak to whether this exists or not. But if you build it yourself you’ll want to use something like @Replit@boltdotnew@Lovable . With a few prompts you can get a prototype running that works perfectly for you. No code needed at all.
@JoshuaImel@okayzade Thanks for flagging @JoshuaImel ! Now I don’t have to bounce back and forth with a different LLM. And the context available will be helpful!
"PMs need to work 60+ hours a week to be successful."
It's bullshit.
What isn't bullshit?
PMs need to *think* about their product 80+ hours a week.
When you do this, your 40-50 hour work weeks become much more optimized than everyone else who's slogging 60+ hours.
@carlvellotti IMO the EM should be running these frankly. If they aren’t, influence that change.
Worst case, PMs should be there to answer questions and ensure the user POV is maintained.