Gregory's Server@quoll @hu_logic @rodhilton the difference is that Twitter was built by teams of competent engineers who followed best practices by implementing proper workflows and code reviews. I don't doubt that it's a complex system, but that's not the same thing as "bad".
The company I was interviewing for was a half-digitized furniture company that was losing millions every month from their simple failure to automate their inventory management. They kept always hiring the cheapest devs, so their code quality was shit and they had absolutely no formalized internal workflows.
I do recognize I made a mistake by pitching a rewrite however. I should have sold them on splitting my time between maintaining the broken system and white boarding a new one over the course of 6 months, rather than spend the interview describeing all the ways their system needed to be replaced.
@joehalfmen @quoll @rodhilton Complexity is a matter of perspective - distributed 10TB database is "complex" to a front-end engineer, but it's relatively boilerplate to a principal data architect.
Same thing with Twitter - it's a highly-tuned high-volume distributed system. It's "complex" to someone who doesn't know what they're looking at, like Musk.
But it's a relatively standard architecture to those who actually work in this space.
Old, but good read:
https://blog.twitter.com/engineering/en_us/topics/infrastructure/2017/the-infrastructure-behind-twitter-scale
@joehalfmen @quoll @rodhilton Complexity is a matter of perspective - distributed 10TB database is "complex" to a front-end engineer, but it's relatively boilerplate to a principal data architect.
Same thing with Twitter - it's a highly-tuned high-volume distributed system. It's "complex" to someone who doesn't know what they're looking at, like Musk.