Due to my company’s line of work, I am of the opinion that some busywork is essential for the proper understanding of the larger task at hand.
We deal with massive amounts of data. Part of dealing with the data is the busywork of analyzing, cleaning and preparing that data to he ingested, and then making sure to validate the state of the data once in the system. That’s a whole lot of mostly mindless busywork. But here is the thing, because of that busywork that nobody likes doing we have a deeper understanding. We know what’s in there, where it is located, how it is formatted and how it will be accessed. We know how a query is going to impact production before it is even run and we can recommend alterations.
Yeah it’s interesting. With no “busy” work there’s no career path to maintain the people with the knowledge of how to design, build, operate and maintain the AI infrastructure. Who will know what Garbage Out is if they haven’t spent a career doing the thing?
That’s why AI exacerbates inequality between more and less experienced workers. More experienced workers will know what garbage to look out for and its manifestations in poorly cleaned data sets. Newer workers will just have to trust the AI did it.
That’s not busy work. Busy work, as explained in the article, is work that doesn’t really accomplish anything, like re-folding towels that have already been folded. Or as I’ve had to do before, sweep a perfectly spotless sidewalk. Data validation is valid work.
Due to my company’s line of work, I am of the opinion that some busywork is essential for the proper understanding of the larger task at hand.
We deal with massive amounts of data. Part of dealing with the data is the busywork of analyzing, cleaning and preparing that data to he ingested, and then making sure to validate the state of the data once in the system. That’s a whole lot of mostly mindless busywork. But here is the thing, because of that busywork that nobody likes doing we have a deeper understanding. We know what’s in there, where it is located, how it is formatted and how it will be accessed. We know how a query is going to impact production before it is even run and we can recommend alterations.
Yeah it’s interesting. With no “busy” work there’s no career path to maintain the people with the knowledge of how to design, build, operate and maintain the AI infrastructure. Who will know what Garbage Out is if they haven’t spent a career doing the thing?
That’s why AI exacerbates inequality between more and less experienced workers. More experienced workers will know what garbage to look out for and its manifestations in poorly cleaned data sets. Newer workers will just have to trust the AI did it.
Right and then forty years later everyonr with experience is retired and no one invested in newbies
That’s not busy work. Busy work, as explained in the article, is work that doesn’t really accomplish anything, like re-folding towels that have already been folded. Or as I’ve had to do before, sweep a perfectly spotless sidewalk. Data validation is valid work.
That makes it sound like nearly the entire thing could be ran by AI. Not just the busy work portion.
Eventually, but not today.