March 16, 2023

A Collection of My Recent LinkedIn Posts

Some time ago I decided to become active in LinkedIn. Don't ask why. I knew it was going to be a mistake and it was a mistake. The posts get some attention and that's fine, but they feel so out of place and posting stuff there never feels good.

(Complaining about the (relative) lack of attention you get is not a good idea either, because you get ridiculed for being a pathetic attention-seeker and so on. I know, I am!)

Anyhow, I do think that I can occassionally make semi-interesting points and I figured (a) I could collect links to those posts here (b) someone else might find them interesting. You don't need to have a LinkedIn account to see the posts. The posts themselves are quite short, a couple of paragraphs only so don't expect a marathon read here.

So here we go:

There's disconnect between something being 'obvious' and acting on it

There's a huge gender and minority gap in programming industry and we should fix it

Psychological safety still is probably the biggest factor in organization and team success

Physical working environments are still not designed with their users in mind

You should correlate interview candidates' score to some actual performance metric (not for a performance review, but to assess your interview process)

Virtanen's Postulate: the choice of architecture drawing tool affects your architecture decisions

The Feeling of Knowing is a brain mechanism and doesn't mean that you're right

My personal feeling is that cloud vendors' icons shouldn't be used in architecture diagrams

Doing cloud certifications is not a good way to learn, but they are sometimes useful anyhow

"Increasing the amount of staff in a software project is an architectural decision. And often a bad one."

(No link, because that was all. This refers to all of those situations where an outsider of a project has decided that the project needs more developers without considering how it might affect it.)

ChatGPT and Github Co-pilot are impressive but they are solving the trivial parts of software development and not The Hard Thing(s)

Having an automation framework (for say cloud infra) already in place is beneficial because it makes it trivial to make automation improvements

You should stop saying that your code documents itself. You should document the 'why' of your code.

The length of one's experience is a terrible measure for skills and aptitude

You should have a Service-Level Agreement for shitty user experience and fix that as swiftly as you fix server downtime

Now that I think of it, most of these could have been posts on this newsletter even though they aren't that technical. Then again, I'm not sure if the content of this newsletter should be that technical. Oh, well. I hope you get something out of my desperate and failed attempt to become a LinkedIn thought leader!