Aaron Eline
My good friend John bought 100 old credit card readers off ebay and reverse engineered them: youtu.be/whOf7lb8nqs
Spent today seeing if an LLM could optimize ocaml's pattern compiler. diving-gray.exe.xyz This site tracks its progress, scored by size of decision tree. The LLM runs in a loop attempting to improve and taking notes on what worked. Nothing groundbreaking but neat to watch it hill climb.
Wrote up a little trip report from NJPLS2026! aaronjeline.github.io/posts/njpls2...
Wanted to send a big thanks to @antithesis.com for organizing Bug Bash 2026. Was a wonderful time full of great people. Looking forward to next year!
Great video from my friend John on repairing bits of his Corvette: youtu.be/Max6wBa4EdM?...
Aaron Eline reposted
DC Systems 14 will be Tuesday, March 10th! Peter Tong will talk about Optimal Real-Time Locking John Kastner will talk about Clang’s experimental bounds safety extension. As always: pizza and beer will be provided. Please come hear some great talks: luma.com/90ss30d2
Aaron Eline reposted
DC Systems tomorrow! We just uploaded a bunch of talks from DC Systems last year. Here's just one of Richard Hart's hot takes about systems programming. Tomorrow, come hear @aaroneline.bsky.social talk about flow-directed closure conversion for typed languages.
Let’s chat about one of coolest compilers ever made :)
#DCSystems returns for 2026! We're starting the year with a special homage to one of our favorite meetups, @paperswelove.org On Tuesday Feb 10, @aaroneline.bsky.social will present and lead a discussion of: Flow-Directed Closure Conversion for Typed Languages, by Cejtin, Jangannathan, & Weeks.
Attended USENIX last week. Wrote up some of the work I saw: aaronjeline.github.io/posts/usenix...
Great talk from Joe Cutler: t.co/8Ffra2YkMt on improving the performance of property based testing.
Great article from @mcy.gay on pretty printing. mcyoung.xyz/2025/03/11/f...
Favorite CS work you’ve read this year? I’ll start in the thread below! (It’ll be obvious I’ve been thinking a lot about testing this year)
Favorite CS work you’ve read this year? I’ll start in the thread below! (It’ll be obvious I’ve been thinking a lot about testing this year)
Google’s work on automating the creation of fuzz harnesses seems promising for lowering the barrier to robust software: security.googleblog.com/2023/08/ai-p...
Cool and wacky work from Neil Walkinshaw on trying to get a formal solution to the “stopping problem” in random testing. arxiv.org/pdf/2405.17019. This formulation has a bunch of limitations/assumptions, but I really enjoy the framing and line of thought.
Aaron Eline reposted