A new AI-powered debugging tool developed by #ManningCICS PhD students Kyla Levin and Nicolas van Kempen and Professor Emery Berger integrates LLMs into a conversational interface to help programmers diagnose and resolve software bugs more efficiently. Read more: bit.ly/41wQdaa

At #OSDI2025!

Just opened @maxhuibai.bsky.social's game, hoping to get better reviews in the game than IRL…

East coast powder day.

Norway, here we come! ChatDBG and CoverUp to appear at FSE 2025. github.com/plasma-umass... and github.com/plasma-umass... - AI-assisted debugging and automatic high coverage unit tests!

Proud to celebrate Professor Emeritus Andrew Barto and alumnus Richard Sutton, who received the 2024 ACM A.M. Turing Award for establishing the conceptual and algorithmic foundations of reinforcement learning, a cornerstone of modern AI research. Read more: brnw.ch/21wR7jr #UMassAmhert #UAlberta

One more TurboTax ad and I'm losing it.

We have Central Park at home:

Man Says ‘Fuck It,’ Eats Lunch At 10:58 A.M. theonion.com/man-says-fuc...

Piano has been untouched since 1791.

It's Not Easy Being Green: On the Energy Efficiency of Programming Languages arxiv.org/abs/2410.05460 @nicovank.bsky.social

Reviewers when they read my paper:

Reducing CO₂ emissions with faster software: pythonspeed.com/articles/co2... TLDR: 1. compute less (a.k.a. improve performance), 2. run on multiple cores, and 3. if the code won't run often, these concerns aren't important.