you (probably) didn't need that anyway

I love the concept of microservices. I like the Unix philosophy of doing one thing really good and leave the rest for other specialists (e.g., other microservices).

I love the focus on efficient communication. Well-designed architecture: databases, storage and code. Micro-optimizations in each of those little buggers. Optimized to the point where it’s actually just diminishing returns, but you still keep going for them. Because that’s what you want, that’s what makes services respond in sub-SLA time.

Summarized: I love microservices. And you probably do, too, if you clicked on this title.

And I love(d) to work with and around them.

There’s just one problem: I just never needed them. Not once. And i’m pretty sure that most of the time microservices were implemented/used, they were overkill. And the person who decided on microservices probably read an article on how Netflix scaled to infinity using them. And then they thought “if it’s good for netflix, it can’t be bad for us!” . Don’t get me wrong, i was at this stage, too. Already small services with a definition-wise limited scope where I so much wanted to make use of super-fancy tech like Message Queueing, Kubernetes and other shiny tech stuff. In that phase of the project we didn’t even have automated workflows, CI/CD or any redundancy apart from what Heroku offered. But I wanted to go for microservices. I didn’t know Jack (and i still only know Ja without the ck, probably) - but man was I convinced that this was the way to go.

And then somehow life around microservices got pretty quiet. Then came Blockchain (don’t even get me started on that) and the whole buzz about Web3, of which i think there are still people trying to convince that everything needs to be on a Blockchain. This hype died off, too. And in the end, NFT became worthless and despite all the LinkedIn Marketing I for one can’t remember a single project name that was pushed at that time.

And I deeply believe that the same will apply for current AI hype technology. Maybe not in the same way, but similar enough (for us ’tech people’) to recognize the pattern.

People are way too eager to use them everywhere instead of figuring out key areas where they actually do a great job. These people read about ChatGPT and then build companies around them. My problem with it is that those people are usually not tech-driven. It’s people that hope to skip the Computer Science Degree and sell their slop without anyone noticing. They don’t need to be smart, they only need to be smarter than the people they’re selling to. And then they’re in this lions den with all their ilk, and they fight for the loudest mouth, they fight for the ‘sickest’ workshop, join the group, become one of us, but give ME all your money, i’ve no software dev credentials but man did Claude design a good homepage for my project. You, too, can spawn a swarm of 1000 worker bees that work for you for pretty much no money at all. Theoretically there should be a thousand new unicorn-SaaS projects. I’m sure all that’s missing is a stable code base.

And that’s what this is really about — not the technology, not even the hype. It’s about grifting, plain and simple. But this time it’s different, because for the first time we’re the target. The dev-people. They’ve decided software development should be like driving a car: you don’t need to understand the engine, you just need to hold the wheel. They deem the mechanic obsolete. And maybe ten of them actually get rich off it. But like so many times before, most of them will only cry wolf and then disappear. And that, more than anything else, is what this is all about.

Because in the end they’re not excited about AI. They’re sad they missed Bitcoin.