Welcome to Sleepless Links! This is my (sometimes) regular collection of things that I’ve read recently that I’ve found interesting and thought-provoking. I’ve let my publishing habit for these lapse a bit, and I’m trying to get back to it. Hopefully you find some of them worthwhile as well!
Here’s what I’ve been reading during the week ending April 2nd, 2023.
Steve Yegge posted a stormer about LLMs and their potential for disrupting the process of writing code in Cheating is All You Need. Agree or not, he makes some valid points. Money quote: “If you’re not pant-peeingly excited and worried about this yet, well… you should be.”
In the same vein is this post on the SK Ventures Substack that is predicting the rapid acceleration of code output coming from the large-scale adoption of LLM tools to aid developers.
For those in the front-end business, Charity Majors has a great treatise on why Deploys are the Wrong Way to Change User Experience. It gets to the heart of both incrementalism in deployment (smaller releases, much more frequently) as well as the use of feature flags and dark-mode deployment.
This short post about AIOps is bound to generate some skepticism amongst those looking to streamline their Ops ranks through the use of AIOps. If you believe the study, your customer experience gets better with AIOps, but you don’t get to reduce staff because of it.
This post about Apple Pay is a bit old, but it’s a good read about the back-end workings of the process and how much inter-entity coordination is required for it all to function. Interestingly, there’s more tokenization happening in more places that I was aware of (though in retrospect it should have been more obvious to me). I was involved in a payment processing tokenization project as far back as 2010, and it doesn’t seem like much about the basics has substantially changed since then.
This post about Platform Engineering by Hüseyin Dursun (currently at VMware) has some excellent points about how to think about these types of teams, and what they should be doing to support enterprise software delivery teams. I don’t quite agree with his assertions about the skillsets required of team members - I think there is room to also bring traditional infrastructure engineers into that space rather than solely rely on folks coming from the dev side.
Lydia Leong has some hard truth for many IT shops with cloud workloads in her post FinOps can Be a Big Waste of Money. I agree with the gist of the article, though I’d counter that there are far more shops that do need a FinOps function than have one. I have a blog post brewing on this topic.
This short post by Michael Coté has some real gems in it, especially the screen shot in the middle about how metrics should change as you shift away from traditional infrastructure management to more of a Platform Engineering approach. I especially liked the shift to “KPIs / OKRs based on customer engagement and value.”
Last but not least, this HBR post You Can Be a Great Leader and Also Have a Life by Brigid Schulte is a few years old, but the premise is still salient. This is something I’ve struggled with in my career, bouncing back and forth between being completely consumed by my work to periods where I just didn’t want to have anything to do with it beyond the business-day hours. There’s a healthy middle, and this article provides some good pointers on how to get there.
That’s all for this installment - happy reading!
Comments