[…] a lot of AI companies are “selling dreams” that this tech will go from 80 percent correct to 100 percent.
In fact, Marcus thinks that last 20 percent might be the hardest thing of all.
Yeah, it’s well known, e.g. people say “the last 20% takes 80% of the effort”. All the most tedious and difficult stuff gets postponed to the end, which is why so many side projects never get completed.
It’s not just the difficult stuff, but often the mundane, e. g. stability, user friendliness, polish, scalability etc. that takes something from working in a constrained environment to an actual product - it’s a chore to work on and a lot less “sexy”, with never enough resources allocated to it: We have done all the difficult stuff already, how much more work can this be?
Absolutely, that’s what I was thinking of when I wrote “tedious”; all the stuff you mentioned matters a lot to the user (or product owner) but isn’t the interesting stuff for a programmer.
While I agree with the underlying point, the “Pareto Principle” is “well known” like how “a stitch in time saves nine” is well known. I wish this adage would disappear in scientific circles. It instantly decreases credibility. It’s a pet peeve but here’s a great example of why: pseudo-scientific grifters.
Yeah, it’s well known, e.g. people say “the last 20% takes 80% of the effort”. All the most tedious and difficult stuff gets postponed to the end, which is why so many side projects never get completed.
It’s not just the difficult stuff, but often the mundane, e. g. stability, user friendliness, polish, scalability etc. that takes something from working in a constrained environment to an actual product - it’s a chore to work on and a lot less “sexy”, with never enough resources allocated to it: We have done all the difficult stuff already, how much more work can this be?
Turns out, a fucking lot.
Absolutely, that’s what I was thinking of when I wrote “tedious”; all the stuff you mentioned matters a lot to the user (or product owner) but isn’t the interesting stuff for a programmer.
While I agree with the underlying point, the “Pareto Principle” is “well known” like how “a stitch in time saves nine” is well known. I wish this adage would disappear in scientific circles. It instantly decreases credibility. It’s a pet peeve but here’s a great example of why: pseudo-scientific grifters.