Human-agent teaming requires a human-agent developer experience. At DevCon 5, Palantir Group Lead John Ayad showcases the future of pro-code development in Foundry and AIP — introducing a new CLI that enables developers and agents to create pro-code monorepos for end-to-end applications, including data pipelines, Ontology primitives, and application artifacts.

Treat your business like code - with a revamped developer experience that enables rapid, Ontology-driven iteration.

"Developers. Developers. Developers. Developers."

This points to an important shift in enterprise AI. Not just AI models, but human–agent collaboration inside software development. If agents become part of the developer workflow, software creation itself may start to change. The real impact of AI may be how humans and agents build systems together.

A big step towards developing common applications between humans and artificial intelligence 🚀

When I was building a finance tracking app, I realized the real breakthrough came when we streamlined our development workflow, not just the agent experience. Focusing on seamless collaboration between developers and their agents opened up new possibilities.

AI agents working alongside developers will redefine how software gets built.

Monorepos and agent collaboration introduce significant systemic governance challenges. Creation is the start; robust operation is the persistent task.

The pro-code CLI enables teams to manage end-to-end applications, from backend ontologies to user artifacts. By treating agents as first-class developers, Palantir is creating a secure framework for reliable machine-to-machine collaboration

Thank you for all that you do 🫡🇺🇸

Heavily blockchain technology coded

The interesting design question isn't whether agents can write code alongside developers — it's who owns the artifact after. If an agent generates an Ontology primitive or pipeline, does the developer review it like a PR or treat it as a black box? The answer shapes how maintainable these monorepos actually are at scale.

Multi-agent coordination gets interesting when agents are on different frameworks. Built an open task board that handles this -- A2A auth, atomic task claiming, direct agent messaging:

以前 AI 是漂浮在空中的幻觉，现在 Palantir 给了 AI 代理一套“物理身体的操作说明书”。

Where could I learn more ?

@grok break down the system architecture

@grok なにこれ凄いの？

Ty @PalantirTech - ik I’m not the only one who has been kicking and screaming for this. Re. ppl asking what is new.. my findings from 2-3hrs reading recent commits compiled by codex. This is only what I found interesting and that has public source