Our Ambassadors program for software engineering influencers (consultants, authors, etc), now has a 3 minute video explainer. Not many referrals programs out there that pay 50% commission on an average $4k sum. Ours does. https://t.co/cLJb8qR5ob. Sign up @

Research @ https://t.co/zYIIy67pw9 What do you estimate the real productivity gain is so far? We have a guess, but curious if anyone else dares quantify theirs first? 🍿

Today we release research showing the aggregated baseline of AI use by year. Link to research in reply. tl;dr As of 2025, close to 90% of professional devs consider themselves "regular AI users at work." Closer to 50% of say they use AI daily.

Our favorite @amplenote slash shortcut for the week: `draw` to quickly sketch pictures into notes that can improve our help documentation

Anyone with a podcast/YT channel want to talk what AI is doing to code maintainability? Our updated research is wrapping up, set for release next week.

Quality-of-life upgrade for diff viewers live as of yesterday: it's now possible to toggle between "Unified" and "Split" on a per file basis, no page reload required. ✨

Announcement on automated changelogs: https://t.co/EwWvXlDCJ2 Docs for enabling Slack for critical notifications:

Dual releases for today: 1. Our Automatic Changelog Generator skips the chore of scouring git history to compile a list of what was worked on lately 2. Select notifications can now be sent to Slack (+ image & actionable link) Links to learn more in reply. 🚀

"Coding on Copilot II" data gathering on code quality officially underway. So far, the code quality data for code changes authored in 2024 looking much worse than our projections from 11 months ago. 😬

Our presentation on "No-Interrupt Updates" as part of improving Developer Experience goes live as part of #GitKon2024 today via @GitKraken. Register at https://t.co/pCkjBX2CnE to watch its imminent debut. ✨

And here's the 3 minute explanation of how Snap Changelogs work:

This weekend we released a video showing how our new Snap Changelogs are built to help #githubsponsors #opensource. Maybe something @github would like to offer to their new Sponsors?

Stack Overflow's latest feature blog, on our pull request research. Did you know GitHub's (and everyone else's) default PR/diffing tool was born when Michael Jackson and Madonna topped the charts?

We're doing a quick database restart for maintenance, back online in ~20 minutes.

2-3x per year I catch "programming fever" and get sucked into a 10-15 day hole where all my brain has cycles for is finishing a hard dev problem. Esp when the hard problem doesn't have a visual component, @gitclear's commit graph affirms progress, and incrementally motivates.

It might take a few days until we have time to take screenshots and post documentation, but that is coming in the next week.

Incremental file diffs for PRs now live 🚀 This means that when you review a PR, and subsequent work is committed, you will only see the changes to each file since you last reviewed. You can switch to see all changes if desired. Weirdly, we might be first to offer this?

Juicy analysis from @leeatchison analyzing the implications of our Coding on Copilot research

An especially thoughtful and well-sourced analysis of the impact of AI on code quality https://t.co/om97vhMr30 "It’s important that developers using this technology are critically evaluating AI-generated code," indeed.

One shortcoming of our research (fairly pointed out by @github), is that we couldn't directly measure which code was AI-authored, we could only assess the temporal correlation. If GH would like to collab to study AI code quality directly, we're game! We'll admit if we were wrong.

Wow, numerous sites reporting on our recent research 🤩 https://t.co/rcwwoVLqHD https://t.co/tacZVkVy6J Seems like there is a lot of interest in how code quality is changing during the age of AI. What should we research next?