5 Comments
User's avatar
Emma Steenson's avatar

Interesting take, really sums it up. It's easy but then nothing worthwhile is easy..... and hence the underwhelm of the output. You might also be interested to read some of the work of Erik Brynjolfsson around the productivity paradox of IT. Its coming from a slightly more economics-focused angle on productivity. I came across it as I'm doing a literature review for a masters thesis around behavioural change and AI.

Expand full comment
Michael Jones's avatar

Thanks :-) ooo... thats sounds really interesting! Any place you suggest I start?

Expand full comment
Emma Steenson's avatar

The original article I found was 1993, here's a more recent paper (2017). https://www.nber.org/system/files/working_papers/w24001/w24001.pdf I'm working my way forward to more current material - but its slow going! Love to know what you think.

Expand full comment
Michael Jones's avatar

You probably wouldn't believe that I was reading this in a meeting about measuring our AI rollout going over this... https://getdx.com/research/measuring-ai-code-assistants-and-agents/

Expand full comment
Emma Steenson's avatar

Interesting to see the caution against performance metrics as being susceptible to gaming. Thanks for sharing - lots of threads to follow.

Expand full comment