IMVHO the IT crisis is the cleptocracy crisis. I mean for business reasons there was and still is an immense push toward cloud+mobile, and archaic mainframe model even worse than the original one, and now a push toward ML for stuff ML is useless for or still far from being usable.
For a certain amount of time people bought the hype but the honeymoon period is short. Some recap for instance:
- full-stack virtualization on x86 push, selling the idea that's cheap buy a starship to visualize some OS instances, one per service, instead a much more modest iron or even a simple blade witch cheap iSCSI mounted storage. A push needed to sell early "cloud" VPS model, but definitively NOT needed by anyone else;
- when the bubble burst, time to scale down to paravirtualisation aka "containers" with K8S, a hell of YAML, a monster to orchestrate another big load of crap just to sell pre-made images instead of running OSes and service "infrastructure as code" style simply orchestrating with Ansible/Salt and co, this bubble is still to burst even if the NixOS/Guix growth suggest some start to understand that only giants selling *aaS stuff need it;
- "AI" promise, while all others services start to experience lower quality and higher issues than before, most have already see it's a fallacious promise, with very little to be of actual use.
How this system can last longer? How many in HN community still do not realize that with NixOS/Guix System model, or bult-in IaC in the OS, we can host on bare metal pretty anything slicing much attack surface, costs and crappy complicatedness using a fraction of code, services and in general resources, so costs, for doing the damns same job? How long it would take to realize that people damn need to know IT to be citizens in the modern society and with knowledge we can work back on a desktop model with few servers and at current datacenter resources we can run the entire planet in a much modern world than now?
IT industry it's failed because it's not an industry anymore, it's just manager run pure finance based on psychology more than IT. Current management can only change role allowing technicians to design where we can and should go instead of pretending to have their solution in purely economical terms. And BEWARE IT actually is the sole mass-industry we still master, automotive is Chinese now, automotive afferent like Naval, Aerospace etc as well for simple consequence, natural resources are elsewhere in the world, academia enslaved by big corp can only produce useful idiots, we are essentially broke. Software and semiconductors are the sole sectors we are still leaders with enough margin to keep up.
For a certain amount of time people bought the hype but the honeymoon period is short. Some recap for instance:
- full-stack virtualization on x86 push, selling the idea that's cheap buy a starship to visualize some OS instances, one per service, instead a much more modest iron or even a simple blade witch cheap iSCSI mounted storage. A push needed to sell early "cloud" VPS model, but definitively NOT needed by anyone else;
- when the bubble burst, time to scale down to paravirtualisation aka "containers" with K8S, a hell of YAML, a monster to orchestrate another big load of crap just to sell pre-made images instead of running OSes and service "infrastructure as code" style simply orchestrating with Ansible/Salt and co, this bubble is still to burst even if the NixOS/Guix growth suggest some start to understand that only giants selling *aaS stuff need it;
- "AI" promise, while all others services start to experience lower quality and higher issues than before, most have already see it's a fallacious promise, with very little to be of actual use.
How this system can last longer? How many in HN community still do not realize that with NixOS/Guix System model, or bult-in IaC in the OS, we can host on bare metal pretty anything slicing much attack surface, costs and crappy complicatedness using a fraction of code, services and in general resources, so costs, for doing the damns same job? How long it would take to realize that people damn need to know IT to be citizens in the modern society and with knowledge we can work back on a desktop model with few servers and at current datacenter resources we can run the entire planet in a much modern world than now?
IT industry it's failed because it's not an industry anymore, it's just manager run pure finance based on psychology more than IT. Current management can only change role allowing technicians to design where we can and should go instead of pretending to have their solution in purely economical terms. And BEWARE IT actually is the sole mass-industry we still master, automotive is Chinese now, automotive afferent like Naval, Aerospace etc as well for simple consequence, natural resources are elsewhere in the world, academia enslaved by big corp can only produce useful idiots, we are essentially broke. Software and semiconductors are the sole sectors we are still leaders with enough margin to keep up.