cloudnativeboy.bsky.social

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

16/17 Try Kev for self-hosted experiments: it provides a Jev-compatible decision API, but it’s an independent model—not Jev’s weights/training. Compare accuracy, calibration, latency and cost on your own eval set: github.com/jaredpalmer/...

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

14/17 Project #1: build a support-ticket router. Use Choice for team, Score for urgency/frustration and Noul for “Is this an outage?” Send low-confidence cases to a human. Runnable Python/JS quickstart: docs.typesafe.ai/introduction...

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

12/17 Current status: Jev 1.14.0 was announced on Oct 14. TypeSafe says the direction includes small quality/stability improvements and better multilingual support. A detailed public delivery roadmap hasn’t been announced. typesafe.ai/blog/jev-1-14

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

10/17 Limitations: Jev doesn’t write prose, code or explanations. It works best with narrow, well-scoped questions and known choices. For ambiguous problems or multi-factor reasoning, decompose the task or keep the reasoning step with an LLM.

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

8/17 Example: an SRE copilot receives an alert. The LLM summarizes the incident and drafts a runbook step. Jev scores severity, selects the owning team and checks whether escalation criteria are met. Code gates paging; humans approve risky changes.

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

6/17 Jev isn’t simply “an LLM with JSON mode.” TypeSafe describes a different model stack: a parallel sampler and RLCD (Reinforcement Learning for Calibrated Decisions). Its full architecture/training recipe isn’t public, so test the claims yourself.

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

4/17 The key difference: an LLM generates a flexible sequence of tokens—great for language, but apps often need to validate/parse the output. Jev constrains answers to a predeclared type/options, reducing glue code for bounded decisions.

LLMs can write essays, generate code, and reason through tasks. But what if an AI system doesn’t need to write anything, just make a fast, structured decision? Meet Jev, TypeSafe AI’s “System One” model. 🧵👇

2/17 Jev’s core idea: send a piece of state (ticket, incident, document) plus typed questions. It returns machine-consumable answers and probability/confidence signals not a paragraph your app must parse. Think decision API, not chatbot.

AI agents are moving beyond simple chatbots. They can now use tools, access APIs, write code, interact with cloud infrastructure, make decisions, & execute multi-step workflows w/increasing autonomy. How do we know what an AI agent is doing before something goes wrong? Link 👇

🚨 What if your K8s dashboard could investigate an incident for you? Radar starts with something familiar: a live K8s resource experience. It connects to your kubeconfig, detects the resources and CRDs in your clusters, and gives you a detailed view without requiring extensive configuration. 📽️ 👇

In this clip from @cloudnativefm 147, Nadav Erell explains how Radar Cloud extends the Radar experience from individual K8s clusters to fleet-level operations. 🔹 Fleet-wide visibility across clusters 🔹 Search and health checks across your K8s fleet 🔹 A unified MCP endpoint for AI agents Link 👇

I had the privilege of contributing as an Exam Developer for the Model Context Protocol (MCP) Associate exam in partnership with @linuxfoundation.org 🔗 Credly: www.credly.com/badges/ffddd... (1/1)

And for those already working with MCP: What should an MCP certification test that a tutorial or simple demo cannot? I'd love to hear perspectives from AI engineers, platform engineers, DevOps practitioners, open-source maintainers, certification experts, and MCP builders.

I had the privilege of contributing as an Exam Developer for the Model Context Protocol (MCP) Associate exam in partnership with @linuxfoundation.org 🔗 Credly: www.credly.com/badges/ffddd... (1/1)

This is also where my interests across open source, Kubernetes, platform engineering, DevOps and AI increasingly converge. The next generation of platform engineers may not just manage infrastructure for applications.

I had the privilege of contributing as an Exam Developer for the Model Context Protocol (MCP) Associate exam in partnership with @linuxfoundation.org 🔗 Credly: www.credly.com/badges/ffddd... (1/1)

As one of the SME selected for this initiative, I helped: 🔹 Define the job task domains and competencies that the exam is built around 🔹 Write and review certification content 🔹 Determine the passing score benchmark for the exam

I had the privilege of contributing as an Exam Developer for the Model Context Protocol (MCP) Associate exam in partnership with @linuxfoundation.org 🔗 Credly: www.credly.com/badges/ffddd... (1/1)

MCP is quickly becoming an important piece of the AI ecosystem, connecting models/agents with tools, data sources, APIs, & external capabilities.

I had the privilege of contributing as an Exam Developer for the Model Context Protocol (MCP) Associate exam in partnership with @linuxfoundation.org 🔗 Credly: www.credly.com/badges/ffddd... (1/1)

In this episode of #CloudNativeWisdom, a dedicated CloudNativeFM conversation series, we go beyond the usual K8s vs. AI debate and ask a much bigger question: - Is K8s still the right abstraction for the AI era, or are we heading back toward an HPC-centric model of infra? Link 👇