Hacker News Reader

A lot of comments are suggesting to use better performing alternatives but they're ignoring the potential cost savings in hardware alone with small, incremental improvements. Performance might be fine but saving off 1-2% in hardware costs could be significant for a large scale app (like Instagram)

View on HN

The thing is, it takes a huge scale for web apps before python becomes your bottleneck. Between caching, async, task queue, reverse proxes, databases and compiled extensions, you have so many performant options to glue with python that any hardware saving is more likely affected by your choice between redis, postgres and clickhouse than by the language you query them with.

Exactly. It's fine that the PSF and others are tackling performance seriously. But in addition to your point, in the 90% (made up on the spot) of the problems python is perfectly suited for, performance is simply not a consideration.

Reply on news.ycombinator.com