Does anybody else ever think about how "wow, that person is REALLY smart, they are my go-to for tough questions" used to mean knowing lots of stuff in your actual head, whereas now it mostly means you're better at search engines than most people you know?
I think about that rather a lot.
@jimsalter You're right and I question my own competence a lot, when searching the web for some C++ template specific thing for the n-th time.
Maybe it's due to a lack of time to really learn and understand the stuff at a fundamental level.
But maybe it's also the result of the explosion of complexity.
For example: I admire devs and maintainers of large software projects and their ability to navigate millions of lines of code, legacy crap and technical debt that accumulated over sometimes decades of development. I admire that capability more then knowing some obscure API call or protocol specification off the top of their head.
@jimsalter
I don't ever call someone "smart" based on their knowledge. Smart people are people who understand many complex topics.
Knowledge is not the same as understanding.
@jimsalter To put in in a slightly rosier perspective, it’s not just about knowledge but about how to apply that knowledge. No amount of web searching or asking an LLM will by itself compensate for actual experience and ability to draw correct conclusions from the available data - and more importantly: to understand when the available data is missing key components. People are bad at describing symptoms. A lot of what I do beyond searching the web, is to ask additional questions.
@mikael yep. Every computer problem is ultimately a human problem.
@mikael one of my favorite prompts is "tell me what you need to do--forget the computer. What is it that you wanted to DO, computer or no computer."
I find that if you don't make people strip the computer out of it completely, you'll find and fix problems--frequently genuine problems, if minor--that the user **never experienced and doesn't care about**, while leaving the issue that they **want** help with entirely unaddressed.
@lounges started happening in the mid-late oughts.
Before full powered smartphones became ubiquitous, we all expected to be functioning in standalone mode too often to mistake google-fu with real knowledge. Some of us were already JOKING about being search engine integrated cyborgs well prior to that, but it was a lot more ha ha and a lot less oh no.
@jimsalter This was about the time I stopped buying tech books, so yeah I agree. Feels like we're in for a much worse version of this as generative machine learning takes over.
@lounges yeah. I mean, I'm no medical imaging tech, but... Those before and after EEG tests MIT did on chatgpt users were pretty damn alarming. Particularly when even Time ran the study.
