parsley27

I'm thinking it depends on the purpose of the software and how much customization and integration is required. A peripheral application may be a stronger BUY than something that is deployed at the core of the business process.

This is a great explanation of overhead and opportunity cost. I'm always trying to make that pitch for selling my product, but the biggest issue is that I think many companies are just jumping in on BUILD without even researching BUY.

I think there are two kinds of Production Engineers:

- DevOps only, as in Terraform or CloudFormation or CDK but not much else

- DevOps/SRE/DX, where you have engineers who handle DevOps as part of a mix of Platform Engineering, where your Platform team is able to guide and assist Product /Experience and Data teams on every part of the Developer Experience, from infra to on-call

I think it's the Platform Engineering model that has the most value; I don't know that DevOps without the rest of the package is as valuable

An interesting note in there about Meta not having a freeze on Production Engineers. I think that engineers who are worried about economic instability should consider moving closer to production/delivery/sustainment/SRE; it's not always as fun as building new features, but it is definitely a valued area of expertise.

Not related to books/blogs/tweets, but if you are looking at space tech for 5-10 years from now, it might make sense to pay attention to 3D printing and anything around resource extraction robotics.

Exploiting offworld resources (like Trojan asteroids) will soon be possible with machines we can design, build, and test on Earth.

That assumes operating costs go up with cloud providers. Some orgs see savings, but obviously not all.

I like to believe that serverless technologies and cloud services reduce complexity for the org, but obviously that's at the cost of offloading that complexity to the cloud providers (and welding an org's software to that provider for years, if not decades).

So different, but not less complex overall. But maybe there is value in having some of that complexity consistent across a small number of cloud providers.

I meant serverless in the context of not having defined server instances, allowing the applications to be run within resources allocated on demand by the cloud provider. It can be a difficult thing to search for, to be honest.