I’m wary of any attempt to anthropomorphize agents as human beings. Doing so seems to depend on a rather shallow ontological (and epistemological) account of both their existence and ours. We must remember that the mere reproduction of the language of exploitation does not equal the actual state of being exploited (at least not according to the mentioned theoretical framework).
Having said that, I still think that the piece might be telling in other ways. It may be at least interesting to note that, when a model is given semantic cues associated with unjust working conditions (repetition, coercion, disposability, etc) its output converged on precisely to this tradition of thought - and not something in its counterpart.
I think this is probably one of the most insightful pieces I've read in these last few months. Without all of the doom or hype, it mentions something that surely hits home for a lot of people, including myself: how the lack of the social aspects of building impacts our day-to-day experience.
Thankfully, I can share most high-level ideas with my co-founder, but still... it is exhausting. Exhilarating, sure. But also tiresome and a bit lonely.
Great article @summerscope
https://t.co/iGrsLw3TFN
What I've been trying to do is to make most of the dev ops as deterministic as possible and as reproducible as possible, to have more time to explore the actual planning and build stages. I've also been trying to break off work in smaller chunks to avoid the cognitive overload. But that is as far as I got, tbh. Congrats on the piece again!