Ops 'N' Hops

The episode contains a few genuinely useful technical specifics - Terraform 0.11→0.12 HCL type system change, CloudFormation vs ARM vs Terraform trade-offs, using CloudFormation to bootstrap Terraform state - but these are heavily diluted by beer chat, career autobiography, mutual nostalgia about Exchange Server, and generic platitudes about failure and asking for help.

“0.12 introduced actual data types to the language, which meant it could actually understand an object versus a list versus just plain text”

“Terraform was a beautiful union of all the things I liked in infrastructure as code...it maintained state so you could actually use it to manage your infrastructure ongoing”

The episode recycles very common DevOps-community wisdom - pin your versions, learn fundamentals, fail forward - with only a couple of mildly distinctive observations like using CloudFormation purely to provision the Terraform state bucket. Nothing here is contrarian or first-principles.

“You know how you use Internet Explorer to download Chrome? That old joke...the only purpose I use cloudformation to deploy the S3 bucket and other IAM resources that I need to work with Terraform”

“don't be afraid to experiment and don't be afraid to fail, because failure is a learning opportunity”

Ned is a legitimate practitioner with seven years of consulting, deep Terraform/Exchange/VMware hands-on experience, and 30+ Pluralsight courses, making him a credible practitioner-turned-educator; however, he is no longer an active operator at scale and is primarily a content creator at the time of recording.

“seven years in consulting is like 20 years in a regular job. I mean, you know, it's. You just learn so much because you have no choice”

“I started sort of evangelizing Terraform around the consulting office and a lot of us ended up using it for the majority of our projects going forward”

There are concrete version-level specifics (Terraform 0.9/0.10/0.11/0.12, Exchange 2007/2010/2013, AWS provider v4 S3 breaking change, Barracuda email gateway incident) but broader claims - cybersecurity incidents 'doubled in the last two years,' VMware licensing creating a 'vacuum' - are stated without sourcing or quantification.

“the really big change was between 0.12 and 0.11. So in 0.11 was HCL version 1, which didn't have strong data types”

“the Barracuda email security gateway incident...the advice was, we can't even patch this. You just have to pull it out of production immediately”

The host asks a few genuinely useful follow-ups - probing the definition of 'genuinely need help,' drilling into specific Terraform version transitions, asking about CloudFormation vs Terraform trade-offs - but spends significant airtime recounting his own parallel career story, and the closing segment on failure goes entirely unchallenged as a platitude.

“So we've taken a lot of liberties whenever you just explained that one of the big words that I heard was, when you genuinely need it, can you add a little more color on what genuinely is”

“do you ever find yourself going back to cloudformation...do you find that the benefits of Terraform almost always outweigh cloudformation”