There is nothing permanent except change. – Heraclitus
Rahul Pandita
@rahulpandita.bsky.social
In no particular order: developer, researcher, dad, engineer, student-pilot. Tweets are mine & are not endorsed by my employer. #GitHubNext #Copilot #AI
Richer issues are the key to OSS contributions as code slinging becomes trivial. https://ferryman.io/r/gXXVozjd
The New OSS — Samuel Kahessay
The bottleneck in open source is moving from coding to diagnosis. This is what happens when it does.
ferryman.io
Paper accepted at MSR 2026 : "Investigating Autonomous Agent Contributions in the Wild." It was an absolute pleasure working with the extremely talented coauthors Razvan Mihai Popescu, David Gros, Andrei Boţocan, Maliheh Izadi , and Prem Thomas Devanbu.
My experience jerry-rigging a web app to build software on my phone. It's more feasible than you think. rahulpandita.me/blog/2026-01...
Coding on a Phone: What I Learned Building Software on Mobile in the Last Three Weeks
Over the past few weeks—largely out of a sense of FOMO—I’ve been forcing myself to do something I’ve fantasized about for a long time: building software primari...
rahulpandita.me
Weekend XC with my favorite passenger(s). The mini-mes were not impressed
TIL that adding a screenshot as an input to iterate with Copilot is not that widely known. I've been using it for a while to streamline/debug my UX code.
Lately, I've been experimenting with agentic PR reviews. Here's the prompt I use. I'm curious what you folks use: https://gist.github.com/rahulpandita/ebc337ca7a6c67823cdcca065d04fc4a
Model builders, if you're taking recommendations, I would love to have a coding model that is really good at git rebase, merges, and resolving conflicts.
Karpathy's suggestion of using LLMs as a simulation machine "Who are the best people to answer XYZ and what would they say about it?" is a wonderful self-reflection heuristic. Leverage it to the fullest before the next round of RLHF makes it ineffective for upcoming models. 💡
How the programming evolved from pressing ctrl+C and ctrl+V to clicking Allow
We are at a point were model capability is not a limiting factor human cognition is
<3 Aviation Recap! #ffRecap https://plan.foreflight.com/recap/c6c14ff8-ce09-47a0-be09-3ee96bfedf79 via @ForeFlight
Not everyday you get to perform run-up in your tiny cessna next to a T-6 Texan
My latest post is out! 🔄 Introduce "Semi-automatic Performance Engineering" 🏆 A Grand Challenge for CS 🧩 Why it's hard, heterogeneous ✨ Show GitHub Agentic Workflows "Daily Perf Improver" 🔮 What's Possible, Potential, and Delusional 🧹 Let's tackle technical debt! dsyme.net/2025/10/12/t...
Introducing Semi-automatic Performance Engineering
(This blog post is written in personal tone, but relates to our work at GitHub Next and may be moved to in future. A huge thank you to Peli de Halleux, Joe Zhou, Eddie Aftandilian, Russell Horton, …
dsyme.net
Degrees are merely proof of the money you've spent on a course. Education is knowing how to use what you learned. rahulpandita.me/blog/2025-06...
Rahul Pandita
Rahul Pandita - Software Engineer, Developer, and Researcher
rahulpandita.me
There is high likelyhood of the Bidding wars in AI companies to converge on global maxima of Nelson Bighetti (aka Big Head).
Leadville (KLXV) North Americas highest elevation (9934’) airport ✅ Mt Elbert Colorado’s tallest peak in the background!!
I'm now a FAA-certified private pilot(ASEL) and can joke that I'm a pilot working on @githubnext.com #Copilot.
Growing up, I always found it both frustrating and amusing to see my dad glitch in real-time when we switched from a 9-channel TV operated by a knob to a 999-channel TV operated by a remote.
One interesting thing I've learned when talking with GitHub CEO @ashtom.bsky.social last week is how GitHub is hiring *more* early-careers devs (interns / new grads) as before. I asked: why? Given lots of companies are hiring less assuming AI might be at the level of a junior dev (cont'd)
AI won't resolve process issues; it only rapidly reduces the time to failure due to them.
The Prometheus principles to human-centered AI design t.co/xMJzA75GUj
https://arxiv.org/abs/2002.04087
t.co
"Automatically setting up, building, and testing a developer environment to aid Agentic Software Development." with tamás and davidslater githubnext.com/projects/dis...
GitHub Next | Discovery Agent
GitHub Next Project: Agentic Setup, Build, and Testing of Repositories.
githubnext.com
Devcontainers are the way forward to share the runtime between vibe coding devs and fledgling autonomous agents…