Punch card headed Software Developers on AI
Document envelope stamped TO BE REVIEWED
AR glasses with a temple-mounted camera
Magic 8-ball showing 43
Six-channel mixing desk
Folding multitool
Headphones with green earpads
Six keycaps spelling AGENT
Human × AI
Code-fluent LLMs are reshaping how software gets made. Not every new interaction scenario helps — finding the ones that do, and designing for them.
Evidence bag labelled BENCHMARK RESULTS
Torn research label
Kotlin compiler printout, annotated by hand
Green tray of electronic components
Two lottery balls, 0 and 1
Hammer stamped FORGE BETTER PROOFS
Pocket compass
Spark plug
Tangle of cables and earphones
Verified AI
As AI writes more of the code, the hard part is knowing whether it’s right. The tools and methods that hold it to the same bar as everything else.
Stack of labelled file folders
Sheet of fingerprint samples
Orange drawing stencil
Label embosser printing WORDS WORDS
A pen, a marker and a mechanical pencil
Embossed biscuit stamped with Kodee
Folding shears with translucent amber handles
Twenty-sided die showing 9
Tomorrow Lab
Today’s LLMs weren’t built for your codebase. Tomorrow’s learn its languages, its compiler, its runtime — and we’re building them.
Scroll

The Home of Research

JetBrains Research studies software engineering itself — the tools, workflows, and habits of mind behind how modern software gets built. We publish in the open, partner with universities worldwide, and feed what we learn back into the products millions of developers rely on.

KairosJetBrains Kairos