8th July 2026, 10:27 AM
Yeah, you pretty much nailed it. LLMs don't understand speech, images, or... anything (other than pattern recognition). I remember discussing this with some coworkers before, and one of my gripes is that AI has made some software less reliable. A calculator, for instance, will spit out accurate information pretty close to 100% of the time. Nothing fancy, just an ordinary calculator. But AI? It may get most basic math equations right (again, statistical probability will likely steer it in the right direction), but it's significantly more likely to get the answer wrong, especially if it's a more complex equation. And that's just math, which is the least abstract thing in our entire universe. For any task with the tiniest amount of subjectivity or requiring judgment, AI is ill-suited to the task (incompetent on a level unacceptable for human employees, as you stated). Humans can, in fact, create airtight programming code that works as it should 100% of the time (or very close). The more complex the code, the more likely there are to be bugs, and the more debugging and revising of the code may be necessary. AI cannot replicate that, though. It can do the task more quickly, perhaps, but with far less accuracy and without the discernment to determine where things have gone wrong when things do inevitably go wrong and how to fix those issues in a way that doesn't create more or bigger issues.
And yeah, the obsession with pushing chatbots (or "AI assistants") over the last four years has been... corrosive. People online already lack information literacy skills and tend to believe whatever nonsense they hear on the internet. For some reason, people are even less likely to verify misinformation when it comes from a robot than from a human. In extreme cases, there have been cases of AI psychosis. Richard Dawkins recently posted a series of tweets (or X posts, or whatever they're called now) in which he insists that Claude has developed sentience. Some people are referring to his recent psychosis as "The Claude Delusion."
The pollution sucks, too.
And yeah, the obsession with pushing chatbots (or "AI assistants") over the last four years has been... corrosive. People online already lack information literacy skills and tend to believe whatever nonsense they hear on the internet. For some reason, people are even less likely to verify misinformation when it comes from a robot than from a human. In extreme cases, there have been cases of AI psychosis. Richard Dawkins recently posted a series of tweets (or X posts, or whatever they're called now) in which he insists that Claude has developed sentience. Some people are referring to his recent psychosis as "The Claude Delusion."
The pollution sucks, too.