Why Python was first

Part of Java is Coming for Python’s Lunch - The Rise of JVM AI Agents. Slides 5–17.

Why Python was first slide 5 ↗

and why Java will last

AI started in the lab, and the lab spoke Python slide 6 ↗

research & quick prototyping

numpy
pandas
scikit-learn

Researchers joined companies and brought Python with them slide 7 ↗

research lab

companies

LangChain LangGraph

Python was the glue, the heavy lifting always ran in C and C++ slide 8 ↗

the API you call

the code that actually does the math

LLMs moved the heavy lifting into the datacenter slide 9 ↗

your application HTTPS

thousands of GPUs

What travels over the wire is JSON and JSON Schema slide 10 ↗

type safety

{
  "name": "create_invoice",
  "parameters": {
    "type": "object",
    "properties": {
      "customer": { "type": "string" },
      "total":    { "type": "number" },
      "dueDate":  { "type": "string", "format": "date" }
    },
    "required": ["customer", "total", "dueDate"]
  }
}

An agent is mostly concurrent network IO, and the JVM is built for it slide 11 ↗

performance

your agent JSON LLM API JSON tools & services

Jackson turns a record into the schema and the answer back into a record slide 12 ↗

type safety

Java

Java
record Invoice(String customer, BigDecimal total, LocalDate dueDate) {}

JsonSchema schema = schemaGenerator.generateSchema(Invoice.class);

Invoice invoice = mapper.readValue(llmResponse, Invoice.class);

The context an agent needs already lives in JVM systems slide 13 ↗

context

the data, the domain model, the business rules

banking
insurance
logistics
manufacturing
healthcare

Put the agent inside the system, not next to it slide 14 ↗

integration

an MCP server and a Python agent per legacy system

the agent lives in the Java project you already have

A production agent needs the same safety net as any other service slide 15 ↗

reliabilityobservability

your agent

retries
security
OpenTelemetry
Prometheus
Grafana

The big JVM vendors are building best-in-class agent frameworks slide 16 ↗

Spring AI Koog

Most AI projects fail because they automate the wrong thing slide 17 ↗

replacing human interaction

boring, repetitive, time-consuming processes