[taneq]

I think that's something we lose sight of in software sometimes, where we're used to some token effort improving performance by 10x or more. In the real world, if you can make something 10% more efficient, that's a game changing, industry conquering advantage.

[IronWolve]

Too true. Like the new dodge ram etorque, using its electric motor to replace the alternator, it can launch the truck and let the combustion motor take over at highway speeds. 1-2 mpg is a large improvement for a truck. 1500 dollar option that pays for itself, and a good step towards getting battery tech into trucks.

I've seen it mentioned boats/ships using electric assist will also be another catalyst for battery tech integration, fuel costs on those large ships are always a topic.

Bolt on solutions that are affordable with quick turn around on investment.

[ballenf]

Agree, but with a few caveats:

- The cost needs to be fairly easy to measure

- Some degree of competition in the industry exists (otherwise the capital could be used to increase revenue or acquire small potential rivals before they're too large to afford, among other higher return places to spend the capital)

- The industry is mature and sources of larger savings already exhausted

- Prospect for some future much bigger savings small (or at least not in conflict with the current effort)

- Prospect for the industry as a whole very good (in a shrinking market, 10% savings becomes lower priority than selling off assets quickly enough at high enough cost without subsidizing a competitor or startup)

- Industry consolidated enough (or capital costs of the improvement low enough) to finance the upfront cost

[uryga]

I'd guess (no real evidence) it could be because a lot (most?) software runs on timescales where 10% isn't that significant. Also, maybe something to do with Amdahl's law, where speeding up one part by 10% rarely speeds up the whole thing by the same amount?