You Believe in AI the Wrong Way
The current obsession of some individuals about using LLMs for everything is partly driven by the fear of missing out; more so it seems to be a fear of becoming not needed and obsolete in the (apparently coming?) age of AI.
Can something be done about such fear? As to concepts that we do not understand well yet (like large language models, AI and gods), in the 17th century the prominent French mathematician Blaise Pascal suggested a risk mitigation strategy that any individual can follow: believe and submit to that god just in case. For if it exists, the submission guarantees life in heaven, but if it does not exist, the small effort in believing in him doesn't seem like a large waste.
It seems that currently many people are unconciously implementing Pascal's wager to the AI situation: some leave "manual coding" behind and try to turn themselves into "agentic engineers", some use the guessing machine as their therapist, some recommend others to read non-existent books ↗ while some allow LLMs to include non-existing scientific papers ↗ into their scientific research.
The issue is that such a strategy has been criticized as non-optimal a long time ago.
First, just a formal belief in a god doesn't guarantee safety, as any omnipotent god can easily catch one's dishonesty. You cannot fake faith by just following religious rituals; you cannot demonstrate your belief in AI just by changing your profile description on LinkedIn to an "AI engineer". Therefore, unless you turn into a true believer of AI, it's not going to work.
As of today, becoming a true believer of AI apparently means never doing anything manually again, delegating to write all your work emails and even personal messages to an AI bot, and giving an unconstrained AI agent access to your personal health records, bank account, photos, and emails so that it can shop and transact on your behalf.
"Happy Birthday, mom! You are now 65 - it's not just a theoretical assumption. It's an operational truth grounded in a documented reality."
Second, similar to the problem in believing in a traditional god, there is no way to guarantee that one has chosen to believe in the correct god. Maybe one has decided to be a believer in the Great Green Arkleseizure ↗, but the only real god is Cthulhu; most probably Cthulhu would be upset about the fact.
The same happens with AI: how do you know that this is the correct god, sorry, the correct future to believe in? What about advanced robotics or biotechnologies? Or interplanetary humankind development? Or new anti-viral medicine? Maybe instead of watching that 100th YouTube video on how to write a perfect prompt for an intergalactic group of AI agents, you should be preparing to live a life where you are allowed to go out of your apartment only for grocery shopping and always wearing a face mask (difficult to imagine, right?)