2025 turned out to be quite the year. Professionally, had the chance to continue working on exciting AI for code opportunities at google. Personally, got promoted to Dad. Excited to welcome 2026 and continue these journeys.

Classic working cat

Turning 36 today and once again thankful for all the people I have the good fortune of spending my life with. It's been a great 36 years so far.

Looks like the Danish have made a clear choice.

Looking forward to attending ICSE 2025 next week on Ottawa -- if you'll be there and are interested in AI for code (broad applications including program repair and synthesis), would love to chat.

I wonder what fraction of the US public is following the events in Turkey, and notice we're on the same path here (just 10 years earlier)

I know of multiple talented prospective CS PhD applicants, who unfortunately did not get any admits (I very much suspect funding uncertainty played a role here). Are there any non-US opportunities that are still open? These applicants have done meaningful research in AI for code

If you're in the ATL area, join us next Thursday for a Foundations of AI seminar series instance at Georgia Tech (foundationsofai.github.io), where I'll be presenting work on LLM-based agents for software engineering.

I was catching up on recent papers and did a double take when i saw this. Then after clicking into arxiv saw that the paper title was closer to what i would have expected. @abhikrc.bsky.social seems like google scholar has made this title quite exciting!

Flying is always a bit annoying (verging on miserable depending on the flight), but the fact that i can go to sleep on one continent and wake up on another is nothing short of a miracle

Every time i read about another cryptocurrency scandal, i wonder who is still plowing money into these. Too bad the sec probably wont look at these for the foreseeable future

McCormick's entire argument rests on appeal to authority, that somehow because he's an MD he understands federal staffing needs and thinks it is reasonable to expect AI to replace cut staff. If this guy knows as little about medicine as he knows about technology, I'm glad he's no longer practicing.

Sicco Verwer, Annibale Panichella, and I are hiring! If you are interested in program synthesis and repair, explainable AI, and reducing errors in software, consider applying! lnkd.in/djir-dsh

The CDC has an ultra-competitive 2-year program to train "the best of the best" to lead public health labs. Most of the current class were terminated over the weekend. www.science.org/content/arti... (by @meredithwadman.bsky.social)

Please submit to Onward! papers and essays. One of the best places to publish ambitious non-mainstream ideas on programming. 2025.splashcon.org/track/splash... 2025.splashcon.org/track/splash...

And there it is, a Faustian deal if I've ever seen one www.nytimes.com/2025/02/17/w...

Can someone explain to me how the f. congress/senate members are allowed to trade individual stocks? Truly insider trading for you but not for me.

Lol...someone legit decided this was a reasonable choice...

The incompetence boggles the mind. These people dont know their head from their ass. And dem leaders still cant figure out how to stop them... The dumbest timeline of all.

Donald Trump and Elon Musk’s idea of efficiency is stopping research on curing cancer.

some times you think you're going crazy -- and then some times you realize it's bad hashing in your experiment's keys...

"just one more lane... that's the trick"

Terrarium Saturday (first time but so fun!)

Excited to share that our work on validating synthetic data for spreadsheet formula generation from natural language will be presented at NAACL Findings 2025 arxiv.org/abs/2407.10657 congratulations to the lead author, Usneek Singh, who i had the pleasure of working with at MSFT.

Cats as album covers

There are many magical places out there. It's important we protect them

I am thrilled to finally be able to share this code that I recently open sourced! github.com/JHUAPL/SHIELD github.com/JHUAPL/openf... It is the culmination of work that I was principal investigator of in the ML + FHE space, about a year ago now. I recently got permission to open source it.

I aspire to this level of graphic design in my future posters

1 hr long headways....a deeply unserious public transportation system

Excited to share some recent work: arxiv.org/abs/2501.07531 . We evaluate LLM-based agents for automated program repair in Google's large codebase, generating fixes to real industrial bugs. Joint work with a bunch of great folks in the DevAI team. We'll be presenting at ICSE SEIP 2025.

I recently watched Horizon: An American Saga - Chapter 1. It is the most unintelligible movie I've seen in a while. I'm still shocked someone saw the editing and said "ship it out".

A very serious researcher

A country so wealthy we've wrapped back around to dangerous and primitive practices: www.nytimes.com/2024/12/31/u...

Pretty sure this is how we end up buying a new tv

Nothing like getting laid out by a cold during the holidays...

SoCal Programming Languages and Systems is back and will be @ucsd_cse in February! Submit your abstracts! socalpls.github.io @ranjitjhala.bsky.social @manu.sridharan.net @cristalopes.bsky.social Reposts appreciated!

Gotta respect the hustle (of complete crock).

My wife's gift to me this week. Now gotta figure out how to wear this in a way that is clearly ironic.

Reviews are like pokemon, gotta catch them all. I've previously caught this one too (an FSE submission we had a few years back, got this exact complaint).

17 years after first moving to the US for college!

Ah yes, my often-used business card and my publicly available *internal* corporate directory.

So rarely do i encounter a product that does exactly what it says and over delivers on quality. This tree stand is that.

This whole HF/bluesky dataset debacle thing is interesting. I dont think i was aware of how much the "progressive ex-twitter" hates AI. In any case, I'm using some of the HF blocklists to find new folks to follow.

Sharkie isnt sure about this work thing

I had a great chat with Tim Scarfe of Machine Learning Street Talk (x.com/mlstreettalk) on AI reasoning, program synthesis, and AI for math and science. Here's the video -- please reach out if you would like to discuss these topics more! www.youtube.com/watch?v=XFMk...

I really like the idea of starter packs, but would love a small extension: autoupdate follows for starter packs where i hit follow all (or maybe this *is* already a feature and i havent figured it out)

What if we could only code by using examples? How can we integrate AI into actual programming languages? I made a little proof of concept: austinhenley.com/blog/mirrorl...

Hi BlueSky! I'm an Assistant Professor at Purdue doing research in PL/SE. I'm interested in developing software verification techniques that fit into existing development workflows, so we can produce more secure software! I assess techniques with theory & proofs, user studies, & by building tools.🧵

I made a wee starter pack of some of the computing people I follow. I am sure I have forgot tons of people! go.bsky.app/MWGdzxM

Great resource! Insta follow for folks interested in PL. This should make the Twitter offramp much smoother.