123dev #49: Skills, stories, and software every dev should know

Isn’t it great when a problem fixes itself? I’ve seen this happen more often with people systems than with technology. Ignore that email a couple of days and it’s likely not to be a problem anymore.

With technology we usually have to take some action. Reboot a server, add a sleep, or add more resources. It’s not a “fix” but the problem isn’t visible for a little while so we can go back to doing other things.

Sometimes you can push the problem off long enough to another person, team, or even a provider. Other times you push the problem off to the future you hoping you’ll have more time or more knowledge to fix the problem right this time.

I watched from afar as 20,000 people attended re:Invent this past week in person. It looked like a lot of fun and I can’t help but think a lot of people learned a lot of things.

More important than learning things, conferences are about building trust and networking. Humans build trust by doing things together and having shared experiences. Usually, trust is built faster when people are physically together because the most memorable experiences are created by our senses.

If you want to learn something, there are more efficient ways to do it than attending a conference. If you want to build trust and meet like-minded people there are cheaper alternatives.

But I know not being efficient and not finding the cheapest solution also have value. When it comes to doing both things and being a good way to get funds—conferences are the best we currently have.

I worked on a new autoscaler for Kubernetes called Karpenter which was launched this past week as now being “production ready.” The main question I heard was “how is this different than the existing cluster autoscaler?” I made this quick video to describe some of the differences. In true “Justin” fashion—I used props.

I don’t enjoy doing coding puzzles for fun. It’s weird to me. I’m sure it would make me a better developer and probably help me interview better, but I just can’t get into it.

If you do like this sort of thing, I always hear about people who like advent of code.

This was the most level headed review of blockchain and web3 technologies that I’ve read. Many of the technical limitations can be overcome, but I wonder how much of it is already too late. Even if the design is changed, is there any real benefit compared to alternatives? None that I’ve seen.