Thanks for the slides, it doesn't look like the challenges are insurmountable, most of the things outlined in the slides are not so much lirbaries as organising your code to avoid circular dependencies (a good thing anyway), and which data store to use, which is a problem any app faces, I don't think it's esp. hard on Go. It would be nice if we had a better abstraction in database/sql so that this was the only package you used, but it's not bad.
This is a little short. It'd be nice if it had some more detail on the different options for logging and the tradeoffs involved. For example for some apps you want to log to a file so you have something to inspect, for distributed apps you probably want to centralise your logging somewhere - log to stdout and have the logs shipped somewhere else, or send them from the app to another server. This just goes over the basic usage of the log pkg, which is pretty straightforward anyway.
Not sure about this one - seems to be a lot of logging and testing packages there, but that's just because everyone does those things, and many people reach for a library to help them test after coming from other languages which required it. Personally I find the built-in testing absolutely fine with table driven tests, sub-tests, I don't feel the need for things like assertions (that's what if is for !).
The logging is more interesting - perhaps a case there for structured logging in the standard library.