4 ms·

What software do you think could help speed up discoveries/inventions? Ideally, something which doesn't require immense computing (like materials simulator).

6510 7y ago
Ill pester you with a riddle I came up with the other day. I'm calling it the apples and oranges problem. If we could answer it it would increase progress a thousand fold: The question is: How do you compare unrelated features? To put silly: What is more important, a led on your coffee machine or gears on your bicycle and by what margin? How much is CPU speed worth vs memory vs screen resolution vs battery life? We could make a chart for each for what it costs, we could make a chart for the technical advantage or disadvantage, we could make a chart for consumer demand. I'm probably forgetting 10 things here.... Now imagine each of those to be dynamic(!) If the CPU does zero cycles per second it will render the entire computer useless. If it is 0.8 Ghz and the goal is [say] web browsing it is almost as worthless as the 0 Ghz. In a few years 1.5 wont be enough either. The infinitely large memory costs an infinitely large amount of money. If we could resolve this puzzle we could make giant leaps forwards. Until we do grandma will have to pick her phone on gut feeling and the market will just have to anticipate that.
- mariushn 7y ago
  :)
- nradov 7y ago
  There's no puzzle or riddle. You compare unrelated features by looking at customers' revealed preferences in terms of how much they are willing to pay. Economists do this type of analysis all the time.
nradov 7y ago
There is a whole niche industry of case-based reasoning (CBR) tools which do exactly that. https://en.wikipedia.org/wiki/Case-based_reasoning https://en.wikipedia.org/wiki/Case-based_reasoning (CBR does have applications beyond just invention.)

4 ms·

What software do you think could help speed up discoveries/inventions? Ideally, something which doesn't require immense computing (like materials simulator).

by mariushn 7y ago

6510 7y ago
Ill pester you with a riddle I came up with the other day. I'm calling it the apples and oranges problem. If we could answer it it would increase progress a thousand fold: The question is: How do you compare unrelated features? To put silly: What is more important, a led on your coffee machine or gears on your bicycle and by what margin? How much is CPU speed worth vs memory vs screen resolution vs battery life? We could make a chart for each for what it costs, we could make a chart for the technical advantage or disadvantage, we could make a chart for consumer demand. I'm probably forgetting 10 things here.... Now imagine each of those to be dynamic(!) If the CPU does zero cycles per second it will render the entire computer useless. If it is 0.8 Ghz and the goal is [say] web browsing it is almost as worthless as the 0 Ghz. In a few years 1.5 wont be enough either. The infinitely large memory costs an infinitely large amount of money. If we could resolve this puzzle we could make giant leaps forwards. Until we do grandma will have to pick her phone on gut feeling and the market will just have to anticipate that.
- mariushn 7y ago
  :)
- nradov 7y ago
  There's no puzzle or riddle. You compare unrelated features by looking at customers' revealed preferences in terms of how much they are willing to pay. Economists do this type of analysis all the time.
nradov 7y ago
There is a whole niche industry of case-based reasoning (CBR) tools which do exactly that. https://en.wikipedia.org/wiki/Case-based_reasoning https://en.wikipedia.org/wiki/Case-based_reasoning (CBR does have applications beyond just invention.)