Operations
LOOP: Partial efforts should result in partial benefits with partial blast radius if it goes wrong. KEY SRE philosophy
The brutal math of compounding errors
Why focus so intensely on 99%? Why not settle for 95%, or even 90%?
Because the math gets unforgiving quickly. At 90% per-step accuracy, a 10-step task succeeds only 35% of the time (0.9^10 ≈ 0.35). At 99%, those same 10 steps succeed 90% of the time – the difference between “sometimes works” and “mostly works.”... See more
Why focus so intensely on 99%? Why not settle for 95%, or even 90%?
Because the math gets unforgiving quickly. At 90% per-step accuracy, a 10-step task succeeds only 35% of the time (0.9^10 ≈ 0.35). At 99%, those same 10 steps succeed 90% of the time – the difference between “sometimes works” and “mostly works.”... See more
Just a moment...
How Complex Systems Fail
adaptivecapacitylabs.comMost engineers move too slowly because they treat every decision as irreversible.
I used to waste days overthinking choices I could have rolled back in minutes.
Careful did not mean professional. It just delayed progress.
Stop asking, “What if I’m wrong?”
Start asking, “Can I reverse this fast and with minimal impact?”
If yes, move forward now.
Substack
... See moreSo, what’s the broader lesson? In practice, our risk models (implicit or otherwise) are always miscalibrated: a history of past successes is just one of multiple avenues that can lead us astray. Trying to achieve a perfect risk model is like trying to deploy software that is guaranteed to have zero bugs: it’s never going to happen. Instead, we need... See more
Nothing fails like a history of success
We better get damned good at recovery
1.
What parts of our system (or process) are truly “core and adopted”—and which are still experimental or unproven?
Clarifies where to enforce strict scaling, standards, and review versus where to optimize for learning and speed.
2.
Where have we drawn explicit, non-negotiable principles or “golden lines”—and do all team members know never to cross... See more
What parts of our system (or process) are truly “core and adopted”—and which are still experimental or unproven?
Clarifies where to enforce strict scaling, standards, and review versus where to optimize for learning and speed.
2.
Where have we drawn explicit, non-negotiable principles or “golden lines”—and do all team members know never to cross... See more
Perplexity AI: Ask Anything
Opportunity Cost Engineering - Why now? Where and when and how much?
An Elegant Puzzle , by Will Larson, on the community book club.
It is the best book I ever read about engineering management, and the reason comes down to one thing: systems thinking. This is a wildly overused term, but Will gives a very precise definition for it.
Systems are combinations of stocks and flows :
It is the best book I ever read about engineering management, and the reason comes down to one thing: systems thinking. This is a wildly overused term, but Will gives a very precise definition for it.
Systems are combinations of stocks and flows :
- 🧱 Stocks — are accumulations of
AI bottlenecks, translating impact, and systems thinking 💡
- f the cost of creating apps craters, the cost of distribution will go up .
- Because there will be more competition.
- To unlock the power of infinite software will require a new distribution model.
- A new medium with different physics.
Bits and Bobs 5/19/25
- Enterprise products get more complicated over time and consumer products get more smooth.
- Enterprise products tend to get more fractally complicated over time.
- Consumer products tend to get smoother over time.
- Enterprise customers are individually big enough to demand niche features but no individual consumer is.
- For
- The tyranny of the marginal user.