My understanding is that the author is this superior being trying to accomplish a massive task (damming a beach) while knowing that it could cause problems for these clams. In the real world, Anthropic is trying to accomplish a massive task (building AGI) and they're finally starting to notice the potential impacts this has on people.
Coquinas are clams that bury themselves in the sand very close to the surface [1]. The author worries that while they are playing with the sand, they might accidentally bury coquina clams too deep and kill them because they can no longer reach the surface.
Anthropic apparently is starting to notice the possible danger to others of their work. I'm not sure what they are referring to.
Are you at all concerned that plugging stuff like this into ChatGPT is leaving you with weaker cognitive muscles? Or is it more similar to what people do when they see a new word and reach for their dictionary?
I see AI like the reading glasses I’ll soon need — not because I can’t think clearly, but because it helps cut through things faster when my brain’s juggling too much.
A few years ago, I’d have quietly filed this kind of article under “too hard” or passed a log analysis request from the CIO down the line. Now? I get AI to draft the query, check it, run it, and move on. It’s not about thinking less — it’s about clearing the clutter so I can focus where it counts.
> Are you at all concerned that plugging stuff like this into ChatGPT is leaving you with weaker cognitive muscles?
Couldn't this very same argument have been used against any form of mental augmentation, like written language and computers? Or, in an extended interpretation, against any form of physical augmentation, like tool use?
In fact it has been, dating all the way back to Phaedrus.
> If men learn [writing], it will implant forgetfulness in their souls. They will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external marks.
There is another side to this, which is maybe we don’t need to know a lot of things.
It was true with search engines already, but maybe truer with LLMs. That thing you’re querying probably doesn’t actually matter. It’s neurotic digging and searching for an object you will never use or benefit from. The urge to seek is strong but you won’t find the thing you’re searching for this way.
It's almost certainly going to be bad, and almost certainly going to be unavoidable.
I can't spell for shit anymore. Ever since auto correct became omnipresent in pretty much all writing fields, my brain just kinda ditched remembering how to spell words.
buuuttt
Manual labor has been obsolete for at least 100 years now for certain classes of people, and fitness is still an enormous recreational activity people partake in. So even in an AI heavy society, I still strongly suspect there will be "brain games" that people still enjoy and regularly play.
We aren't talking about something like spelling or digging a hole. We're talking about a fundamental cognitive skill: reading eight short paragraphs of text and extracting meaning from it.
This is increasingly happening to me every day. Hope the alien overlords don't have spelling tests (as their version of IQ tests) to separate the serfs from the field-masters.
I think that they're saying a little bit of playing around with replacing thinking and composing with automated tools is recoverable, but at an industrial or societal scale the damage is significant. Like the difference between shoveling away some sand with your hands to bury the small creatures temporarily and actually destroying their habitat by "lobbying city council members to put in a groin or seawall, and seriously move that beach sand."
I just hope the educational institutions catch on, stick with their principles and don't give them the paperwork. The paper / title should be evidence of students' learning and thinking abilities, not of just their output.