English pronunciation <-> writing is so far from a mapping, it causes a lot of resource waste.

For foreigners like me that mostly use it in text, the biggest overhead is in speaking, for example I always mess up the words study and student, it's infuriating.

Even if there are dialects that pronounce the same word differently, you could still find a lot of common ground.

Finding that common ground would mean switching from our current system of somewhat-arbitrary spelling to a different but very similar system of mostly-arbitrary spelling. It imposes the same memorization burden on everyone and the benefit is slightly more predictable pronunciation within each of a set of officially-blessed dialects. That ground gets lost over time regardless; there is a reason predictable pronunciation is a feature of spelling systems that either (1) are new, or (2) have just undergone reform.

If your biggest problem lies in a circumstance you rarely encounter, arguably fixing it is not a priority.

> for example I always mess up the words study and student, it's infuriating

This is a funny example to use, since it fully conforms to the rules I described above -- study uses the STRUT vowel, and student uses the GOOSE vowel. It would be a better example for the complaint that we have more sounds than symbols.

Good points, but I disagree. Decoding and encoding becomes a lot harder, if vowels change depending on consonants coming after them, or something else even further down the line. (I'm not a linguist.)

It's like a config file specification that supports gotos.