The risk is high when investing into a younger company. I think younger companies, like Rivian, tend to collectively attract more talent and more modern ideas. Risky sure, but Rivian seems to gotten amazon's deep pocket investment, which is encourages more trust it would succeed.
Oshkosh has experience with making workhorse trucks (i,e their firetrucks), but their EV initiative is banking on the USPS contract, its like a new startup; As risky as partnering with anyother EV company with just a concept car.
Rivians, for example, already have the talent, manufacturing processes, and financial backing to get USPS trucks much quicker
100%, `terraform plan` is such a useful time and cost saving step when developing terraform.
Ive read about pulumi and casually used it (python) to learn about it. Benefits like taking advantage of proper python language niceties, like running pytest on your IaC.
Additionally, it uses terraform's providers behind the scene, both contributing their success to the other.
from the videos and articles I was able to find about the amazon EDV, amazon's investment stipulation was for Amazon to get first dibs on having their EDV manufactured and delivered first. Fortunately, it was not an exclusive platform for Amazon, anyone else can order it
As much as I support USPS to upgrade their fleet, I dont think its a smart financial idea for USPS to fund Oshkosh for the development of their EVs. Partnering with already EV focused companies would be a better long term investment, because they would be funded by other companies looking for EV delivery vehicles. This would drive down cost per unit and long term support. Companies like Rivian with successful Amazon EDVs, or Canoo with their modular design (MPDV).
Military contracting companies like Oshkosh would charge USPS so much money for long-term support, and creating an expensive co-dependence with USPS. USPS wont be able to escape because they would need them to maintain 66,000 bespoke EVs.
Terraform: mainly because my organization used to make everything manually, using boto3/sdk scripts, and poorly tested/constructed cloudformation templates