What I don't understand with both Muratori/Blow is that they seem to be incredulous that someone wouldn't just make software high quality and fast for the sake of it
Like they can't comprehend the fact that these things only happen if there is an incentive for it
I think its because they work in games where there is a business incentive for performance, users care a lot if framerate suffers in a game
They are so used to working in games that they take it for granted that the business incentive for performance exists
As someone who has worked in both AAA games and big tech it just comes across as a bit naive to me
I mean, it’s a thing in pretty much all lower level software that you want it to be fast. I guarantee you postgres and redis are trying to optimize the software to the cycle, because wasted cycles are wasted for all their users.
I think there could be a version of our industry with people doing better work. And it’s worthwhile to study why we don’t have that industry.
Yes it’s incentives - but why these incentives exist and not others is an interesting question to study.
In the 2010s we pushed everyone to learn how to code and maybe lost some of our higher standards. Maybe we’re doing that now with AI?
Maybe broader social and economic incentives have had a lot of short term thinking of late? It’s hard to organize society or a company to go in one direction very long before incentives completely change everything. Who knows, maybe that’s a feature, not a bug.
It’s worth asking though why we have this software industry and not another one. I’m not convinced it’s for legitimate reasons.
I think the thesis would be there's a colossal structural/market failure in software. This because of many degrees of freedom, consumer ignorance, network effects and such. Compare this to physical things. The difference between the absolutely shoddiest couch sellable and the practically best one is tiny, compared to the difference between the worst software you could impose on the market and what could be achieved by competent engineers if we somehow removed the wrong incentives from capital and bureaucracy. With most software most people use, we are near the bottom of this range. I don't know if I fully buy that, even if we completely ignore whether removing these incentives is at all realistic. But would also say there's something there.
Then we can talk about how we all pay the increased cost in energy and materials, and so forth.
The fallacy is assuming it takes more work to get good performance. When what’s really lacking is clarity about the problem and solution.
Optimizing code for max performance does take work. But just not doing incredibly dumb things and writing simple programs just takes education (and re-education).
I can’t count how many times I’ve replaced a distributed system with for loop.
Like they can't comprehend the fact that these things only happen if there is an incentive for it
I think its because they work in games where there is a business incentive for performance, users care a lot if framerate suffers in a game
They are so used to working in games that they take it for granted that the business incentive for performance exists
As someone who has worked in both AAA games and big tech it just comes across as a bit naive to me