Claim◆Audio · 2:22 — 3:52
Coding agents and custom enterprise agents run into the exact same infrastructure problems — model/harness portability, session sharing, and security — so they should be built on one common layer instead of being treated as separate categories.
Matei explains that Omnigent emerged from noticing internal coding-agent tooling and custom enterprise agents kept hitting identical problems — switching models and harnesses, sharing sessions, security — so Databricks built one common layer to serve both. ✦ AI generated
Matei Zaharia · Latent Space · 2026-06-24 · original ↗
plays this moment only · 2:22 — 3:52
Elicited by
“What led you to it?”
at first people thought it was weird. They're like, "Why are you doing coding agents and custom agents in the same thing?" But I said it's, it's the same problems and, you just wanna build the stuff that lets you deliver the agent, maybe control it if you care about security, and, make it portable across things.
verbatim transcript · starts at 2:22
Transcript · around this moment
2:22Omnigent and the Agent Infrastructure Layer
- ·Coding agents & custom enterprise agents hit identical problems
- ·Shared issues: model/harness portability, session sharing, security
- ·Databricks built one common layer to serve both
- ·Initially seen as odd to combine the two
- ·Same problems justify a shared foundation
- ·Goal: deliver, secure, and port agents easily
Around this claim
Evidence · 2
Agentic coding works well because code outcomes are verifiable (does it compile? do the unit tests pass?), but building reliable AI outcomes in domains like finance requires capturing the right input-output data so the agent can validate against a known correct result — this demands a mindset shift toward writing evals and boundary conditions first.Philipp Herzig · No Priors · conf 80%Decagon's short-term moat is the infrastructure and software required to make AI models deployable and safe within enterprise environments.Jesse · a16z Podcast · conf 75%