if its something automated that runs in like ci/cd then i guess its there but i wouldn't actively use it myself and i wouldn't want to rely on any of the outputs (and would probably start looking for a new job on the side)
personally i try not to use it as much as possible, ive had a ton of experiences where llms have given me incorrect and unhelpful information. for debugging purposes i usually just talk to myself
also dont worry!! research is hard as fuck and ur doing incredible work:)
i found this paper which unifies itrees and stutter bisimulation and might be relevant? dl.acm.org/doi/abs/10.1...
thinking about this a bit more, this sorta kinda reminds me a bit of the problem that itrees are meant to solve, which is that you want to encode some sort of non-determinism, so u have these silent tau nodes which can take infinite steps but are in some sense invisible to the language.
fucked up
completely understandable! yeah defining some sort of quotient would have been my immediate next thought as well but it seems highly nontrivial. but i feel like someone has to have run into this problem before
i see… yeah that’s really unfortunate. but afaik, don’t you need to establish *some* sort of correspondence between cfg nodes and code points for the proof to go through at all?
i think the writeup is really really great! the weaker soundness theorem reminds me a lot of incorrectness logic (but im not sure if thats exactly what you want)
also i didn't realize u worked on snakt :0 ive been working on contributing as an open source developer
omg somehow i missed this??? reading it rn
is it the ai code thats wrong ? nah must be the engineers themselves
have you ever thought javascript spent too much effort on type safety?
have you ever wished that runtime error could have been a segfault?
do you just hate useful documentation that takes the fun away from guessing how a function should be used?
then you should try llvm's c api!
sooooo true
woag...
buy less domain ?
noooo i missed it :(
congrats !!!! ur work is so cool and very happy to hear that u get to do it full time:)
holy goated fit
this is so stupid (the waiver thing)
sooooo true
new thing to add to the reading list
woah !!! this is super cool:)
yuh huh
yes :3
woah !!! ur so cool…
everytime i see ast i think of abstract syntax tree
oh dang sorry :( what are u looking for ?
holy based
