my biggest unexpected engineering lesson from my short adventure at merkle:

if you're cracked you have enough confidence to ship code to high-traffic prod without testing it first

stuff that i would estimate in my head for 2 days @gt can do in 2 hours because he just looks at the code, knows what to change and ships it -- while i need to find a way to run in a replicable environment, test, give others and only then im comfortable with shipping

bell curve: test in prod -- automated tests / staging / replicated envs -- test in prod

"short adventure"? 
Supercast revival sooon....