People still need to think critically and interact with the "output" of the LLM / AI and have some knowledge of the subject matter.
The other thing to take into account. We went from not being able to fly to men on the moon in a very short period of time. What's the next "leap" that can happen from now to then?
I think the bottleneck will be the engineering to make the science work.
That's kind of where I was going with this. This isn't the first time we have made strides in science and technology, but it feels a bit more hands off this time. Like I said, is the AI doing all the work or is it just a tool? I'm not sure I know the answer and it probably varies depending on who you talk to.
I know a lot less than what I know.
I feel like it depends on who is using it. A hammer is a tool in the hands of a skilled wood worker. In the hands of a crazed lunatic, its a weapon.
In my hands? its something I drop on my toes, probably hurting only myself.
I think people that treat "AI" as a "human", or a "partner" (or co-pilot, if you will) have probably failed the Turing test by some measure, but they're ... aware they're interacting with the machine.
But yet, they see the output as some sort of collaboration between them and the machine, and tend to feel a sense of ownership of that output. That feels strange, because I can't be seen (or would feel like I should) own the "output" of a conversation between you and I.
(Unless we were academics at a conference or something, or publishing a co-written paper)
Those are some good points and while I agree with you that a tool in the right hands versus the wrong hands can have much different outcomes, I think we also need to be remind of the hubris that can quickly lead from one outcome to another.
Hubris is certainly something that is not in short supply of late. It sometimes takes an uncomfortable duration become apparent in the marketplace of ideas, but it is always there, lurking like a spectre :)
To come back to your top level topic - "Are we cheating?" - Maybe a follow up question could be, what aren't we including in the models - and where are we definitely still doing the hard work?
I think for your follow up question a lot of it is still manual these days. I just have a feeling we are heading in a direction where things start compounding and there will be a point where we can't necessarily stop it. Who knows though. I may not even be alive to see it.
Perhaps it will just be narrower and narrower fields of expertise that we will not be able to fathom. Time has no corporeal voice, yet, it will somehow tell.