New group at Google with focus on ML debugging is looking for qualified candidates. Reach out to [email protected] if you are interested. The group will be responsible for ML debugging infrastructure covering performance debugging,

Visual explanations of core machine learning concepts mlu-explain.github.io

1/ Since reproducibility in PL seems to be hot topic today, here's some (personal) historical perspective. I'm writing this mainly for the benefit of, say, junior grad students, who may not be aware of some of this and may benefit from the background. ↵

Solve a challenge case and submit your search-based software engineering solution to SSBSE'22. We 👇 are looking forward to your submissions.

"Perhaps we might move away from basing grants and financial security on how many papers we can produce. After all, with the help of our AI first author, we’d be able to produce one per day." Could #AI algorithms upend publish-or-perish? augmentedrobot 2.0 pprpl.co/3OW7AZG

This, via Python Weekly, seems pretty impressive. I'm still disinclined to let GPT-3 *write* my code, but I gotta admit, having it explain my old #Python code to me sounds tempting! Although I hate to think what kind of bad habits this wil reinforce ... simonwillison.net/2022/Jul/9/gpt…

I'm always impressed and happy after reading the testing strategy for SQLite! sqlite.org/testing.html

Hey TOSEM World! You can now submit registered reports! RRs feat. a 2-phase review process and host the exploration of risky ideas, eliminating publication, confirmation and impact bias, offering early feedback and revisions of methodology! #goTOSEM ow.ly/wCx550JLGpg

I was recently speculating if AI “Prompt Designer” could one day become a profession. Thanks to José Campos I now realize you can already make money out of that promptbase.com

⭐WE ARE HIRING⭐ We have several exciting opportunities to join our brilliant, growing team. We’re recruiting for a Lecturer in Verification, Teaching and Research Assistants (x3), a University Teaching Associate and a University Teacher. Apply here ⬇️ ow.ly/ALYr50L15W8

wow 😲😵

Think ChatGPT can write all of your homework for you? Think again, for machine learning is a double-edged sword. Folks from Harvard teamed up with OpenAI to create an "output detector" to scan for AI-plagiarized text... ...and it works wonderfully. Check this out 🧵

Last week, thanks to help from Daye Nam, we used a class in Foundations of Software Engineering (cmu-313.github.io) to collaboratively learn about #ChatGPT. It was very interesting, and I would like to share some observations here. a 🧵

Two weeks ago, at the IDE workshop at ICSE, I announced the IDE Research Integration Grant program. The idea is to support researchers who are eager to put their work into practice by creating a plugin for a JetBrains IDE. Please share this news! jb.gg/ide-research-g…

Quote of the day from Nathan Sobo: "The code you can test is the code you can change. The code you can change is the code you can make good."

LLMs can write code, but they can't maintain mental models. Engineers test as they go. When tests fail, they check their mental model to decide whether to fix the code or the tests. Conrad Irwin on why LLMs can't really build software: zed.dev/blog/why-llms-…

Another amazing patchset from "mkver", with a 3.5x speed increase vs C! He also explains some of the mistakes the compiler makes, a key reason we use hand written assembly.

Thanks to Ubitux, FFmpegs coverage site is fixed coverage.ffmpeg.org If you ve got time, help improve coverage - add tests and send pull requests on Forgejo: code.ffmpeg.org More testing coverage helps us all.

If you have a pizza with radius "z" and thickness "a", its volume is Pi(z•z)a. 😋 Happy Pi Day!

Type I and type II errors are widely used terms for false positive and false negative. It's really hard to remember their meanings, so I have drawn an image with an easy to remember mnemonic. Inspired by Chris Albon and his #machinelearningflashcards