3 ms·

I can't find the article again that uses an analogy for "revolutionary" change vs "incremental" change: that of slowly climbing a mountain in a fog (roughly kno

by yboris 6y ago

I can't find the article again that uses an analogy for "revolutionary" change vs "incremental" change: that of slowly climbing a mountain in a fog (roughly knowing which direction the peak is and going there) vs launching up into the air to randomly land somewhere. The takeaway is that if we've made it far up the mountain of progress, a revolution is more-likely to land us somewhere far below the progress we've already made. Anyone remember anything like this? I'd love to find the blog post again!

tehjoker 6y ago
This metaphor only makes sense if there are no inherit limits to the system you are working in and you're climbing the right mountain. For software engineers, think about single core vs multi-threading or scalar processing vs SIMD. You are limited in the performance you can get in a particular regime without making radical changes.
- luckylion 6y ago
  > You are limited in the performance you can get in a particular regime without making radical changes. But you don't know where you are relative to those limits. You may be in 1999 on a 400Mhz Celeron and you'll still see plenty of increase in single-core performance.
animalgonzales 6y ago
I think it was this Freakonomics episode: https://freakonomics.com/podcast/in-praise-of-incrementalism/ https://freakonomics.com/podcast/in-praise-of-incrementalism...
perl4ever 6y ago
This reminds me of something I was reading earlier (may be a HN thread) about how a possibly optimal procedure for searching an area when you have no information is called a "Lévy flight", which "is a random walk in which the step-lengths have a probability distribution that is heavy-tailed". It's distinct from Brownian motion which keeps you closer to the same area. If you wander too far around your mountain in small steps, you probably get away from the peak, but if you want to find another mountain peak, you have to take some much longer steps.

by yboris 6y ago

tehjoker 6y ago

This metaphor only makes sense if there are no inherit limits to the system you are working in and you're climbing the right mountain. For software engineers, think about single core vs multi-threading or scalar processing vs SIMD. You are limited in the performance you can get in a particular regime without making radical changes.

luckylion 6y ago
> You are limited in the performance you can get in a particular regime without making radical changes. But you don't know where you are relative to those limits. You may be in 1999 on a 400Mhz Celeron and you'll still see plenty of increase in single-core performance.

animalgonzales 6y ago

I think it was this Freakonomics episode: https://freakonomics.com/podcast/in-praise-of-incrementalism/ https://freakonomics.com/podcast/in-praise-of-incrementalism...

perl4ever 6y ago

This reminds me of something I was reading earlier (may be a HN thread) about how a possibly optimal procedure for searching an area when you have no information is called a "Lévy flight", which "is a random walk in which the step-lengths have a probability distribution that is heavy-tailed". It's distinct from Brownian motion which keeps you closer to the same area. If you wander too far around your mountain in small steps, you probably get away from the peak, but if you want to find another mountain peak, you have to take some much longer steps.