This is a category error antithetical to the whole industry.
While I agree with the author in sentiment; if you trace back this argument - it boils down to “AI shouldn’t be a thing”. Knowledge work is application of intelligence to make decisions - decisions about what to put in code, decisions of what to purchase, legal decisions, forecasting decisions etc.
If you say - an AI should never make decisions as it can’t be made accountable - then using AI for knowledge work shouldn’t happen - and this is antithetical to the industry. It is naive to think that anyone can delegate to an AI while fully foreclosing decisions and achieve any kind of productivity gain - as productivity is decision-making.
Which is not to say the author is wrong; they have a coherent position - just argued at the wrong generality level AFAIK.
While I agree with the author in sentiment; if you trace back this argument - it boils down to “AI shouldn’t be a thing”. Knowledge work is application of intelligence to make decisions - decisions about what to put in code, decisions of what to purchase, legal decisions, forecasting decisions etc.
If you say - an AI should never make decisions as it can’t be made accountable - then using AI for knowledge work shouldn’t happen - and this is antithetical to the industry. It is naive to think that anyone can delegate to an AI while fully foreclosing decisions and achieve any kind of productivity gain - as productivity is decision-making.
Which is not to say the author is wrong; they have a coherent position - just argued at the wrong generality level AFAIK.