The consequences of AI
I avoid AI/LLMs and the things made by it, but I am still heavily impacted by its existence. A general list of reason on why to avoid LLMs can be found on open-slopware repository. this is a great list, but it lack one key topic: time, wasted time.
The websites implement various scRaping protections
This costs me time to get through. The level of the time waste varies from site to site. At times it is a one time check and on some it is a repeated check each few hours or even each few links (depending on the site).
Depending on a region, an environment or a web browser there is also a captcha and various “are you a human” checks. These often require modern JavaScript implementations and do all kinds of intrusive fingerprinting which makes small web browsers (Netsurf, Dillo, Chawan) or RSS readers unusable and the browsers configured to protect my privacy hard to use.
Some sites put the content behind a login wall, which makes it much harder for bots and also people to use. The alternatice is often to disable / remove the abused feature. A good example of this are various only Git frontends which were shutdown or very limited due to abuse from these scRapers. An example of this could be Sigrok gitweb interface which was disabled, or my Gitea instance which has the “expensive” option active and only allows to view the front page of the repository and cloning it.
Sorry, disabled because of misbehaving AI crawlers.
The modern web sucks
The modern web is full of scams and articles made with LLMs which are hard to read. Well, also count video, audio, music, podcasts, programs, comments, well enything is a mess.
The search engines are also a huge mess. I believe this is intentional to push people to use thei LLM generated summaries instead of visiting the websites and making their own opinions, as the websites are a total mess. I understand that people might rather use this summary, as it can potentialy save some time. It is pushed in front of my eyes, so at times I partialy read it and I am not wiser from that.
The server part
Settings up all kinds of protection and monitoring on the server also costs a lot of time and money. These LLM companies do not pay this, they steal all this time by forcing people to do something about the crazy loads. There is nothing like a permanent solution, it is an uphill battle and as I tried to describe, many server guys go the easy route and disable the feature from public view instead of trying to implement some human detector.
Conclusion
I might be a messy writer, but I triedto to write that even when I do not use the AI/LLMs for anything, I am still heavily impacted by their existence. The time wasted on all these things combined for me is hard to count. I bet it is already many hours.
I am aware that this “article” feels more alike a draft as I am trying to put my thoughts on this topic into an order.