Two years ago this week, as the implications of widespread LLM use began to crystalize, my (Serbian) podcast co-host and I discussed how ChatGPT would change medicine. It was apparent even back then that we were looking at an onslaught of bullshit — since then rebranded as “slop” — that would inundate platforms of the written word from medical journals to progress notes. Sadly, that ended up being the case sooner than I remember thinking.
What I did not expect was for the already odious video calls to become even worse. Friends, co-workers, business partners potential and current: we can tell when you have a chat window open during the call. The reflection in your eyeglasses changes at the rhythm of our conversation. Your mouth stands half-open, ready to parrot the non sequitur comment that your LLM spat out. You cannot provide context or clarify your question beyond reading it more slowly and spelling out the acronyms.
Job interviewer, beware. If you hire people based on calls only you may be shocked when you meet your new employee in person. If there is absolutely no chance for a real-life meeting before hand, at least make the call video. And be ready to bring a shovel.
Of course, this is a people issue not a technology issue. I wonder, however, if at the margin the availability of LLMs is making people worse. That those who would have otherwise made some cognitive effort now rely too much on this crutch. You can make the same argument about the written word On the note of reading and writing, today’s Experimental History newsletter dealt with the topic. I completely agree with Mastroianni’s conclusion and look forward to adding his book to the pile. and how writing caused the decay in our collective memorization capabilities, just as Socrates foretold. Clearly, that particular tradeoff was on net a good one to make. Will we be able to say the same thing about the one humanity is making now?