thanks, mago is a cool project. probably not as a dep tho, the parser isnt where the time goes (the 55x gap is all in evaluation, thats what the bytecode vm is for) and our parser is deliberatly tuned to match php's exact parse error messages, which is itself worth tests in the corpus. but using it as a second oracle to cross check my parser against theirs is actually a neat idea, same trick as the phpt suite but at the syntax layer.
well, thanks for the tips on how to run my own blog :) but the post already tells you this, the last paragraph literally says an LLM drafted it and i edited it. the whole project is an experiment in what a non rust/php guy plus AI can ship, so hiding the AI in the writing while disclosing it in every commit would be a weird place to draw the line.
its still iterating, 17% is just where the counter is today. three weeks ago it was at 10%, two weeks ago 13.8%. i didnt post this as a final result, i posted becuase wp-admin rendering surprised me.
but no, it cant reach 100%. around 55-60% of the suite tests C extensions, gd, curl, soap, intl, mysqli, ffi, sockets etc. passing those would mean writing all those extensions from scratch too (libcurl, ICU, an image library...) which is a completely different project. the realistic ceiling for a from scratch engine is around 40-45% and thats the number im climbing towards.
I have used a couple of laptops with touchscreens, and the experience was awful, even with the latest technology. If Apple gave us an iPhone or iPad-quality touchscreen on MacBooks, I am 100% sure the experience would be perfect.
the real question isn't "should AI write readable code" but "where in the stack does human comprehension become necessary?" we already have layers where machine-optimized formats dominate (bytecode, machine code, optimized IR). the source layer stays readable because it's the interface where human judgment enters.
maybe AI should write better readable code than humans. more consistent naming, clearer structure, better comments. precisely because humans only "skim". optimize for skimmability and debuggability, not keystroke efficiency.