The Twitter engineering team is probably hating this decision. I bet a lot of their infrastructure is built around the assumption that tweet.txt is short :) This has huge implication on the data schema, caching, search, etc.
As an engineer, you can hardcode your assumptions or make them parametric. I get that 140 chars has been a core part of the Twitter identity, and maybe there are some infrastructural aspects that have to be hardcoded to optimize for that. But if they've got `140`s sprinkled all over their code base because that was the assumption of the day, god help them.