>Then again, with the release of each new model, paper, etc. it seems more and more likely that all knowledge-based professions are at risk of being automated sooner rather than later.

I do agree that with the release of each new model we do inch closer to the day when the world of knowledge worker has been greatly disrupted and changed beyond recognition, but I don't think that that day is particularly close.

Knowledge work (in general) is a lot more than just crunching numbers, shuffling papers, etc. Anybody who works in a knowledge-based field (or is familiar with a knowledge-based field) knows this.

AI that's capable of fully replacing what a significant amount of knowledge workers do is still pretty far out, IMO, given how much human interaction, task variety/diversity, abstract thinking, precision, etc. is involved in much of knowledge work (not to mention legal hurdles, adoption, etc).

Will these upcoming models change knowledge work and make some white-collar jobs obsolete over the next 5-10 years? 100%. There's no point in even denying that, nor is there any point in denying that much of the rest of knowledge work will undoubtedly change over the next time span and even more so after that, but I'm pretty confident we're a ways away from it being totally disrupted by AI.

My 2 cents 😊.