English · 中文

Human Approval, Idempotency, and Traceable Tool Calls for AI Agents

Jul 2, 2026

Constrain AI agent side effects with provider-neutral action-time approval, business idempotency keys, and an append-only execution trail.

Multimodal AI in SwiftUI: Image Input, Dynamic Profiles, and Model Switching

Jun 21, 2026

Separate SwiftUI image selection, Foundation Models attachments, dynamic profiles, and model policy into testable iOS 27 beta boundaries.

Foundation Models at WWDC26: From One On-Device Model to a Unified Abstraction

Jun 10, 2026

A bounded look at the WWDC26 LanguageModel abstraction, image input, Private Cloud Compute, tools, and evaluation in the iOS 27 beta cycle.

Local Tool Calling with Llama 4: Model, Runtime, and Permission Boundaries

May 30, 2026

Separate Llama 4 output, local inference, tool parsing, and authorization to build a testable local agent that remains safe on bad calls.

Gemini 3.5 Flash and Managed Agents: What Hosted Agents Change

May 21, 2026

Understand Gemini 3.5 Flash and hosted agent runtimes through execution boundaries, durable state, approvals, observability, and portability.

  • « Prev
  • Page 4 of 14
  • Next »
Portrait of Lukes Lu

Lukes Lu

Software Developer

  • lukes.lu@yahoo.com
  • github.com/ilukes
  • @lupingui
  • Home
  • About
  • Keywords →
    • SwiftUI
    • OpenAI
    • Responses API
    • cancellation
    • iOS 26 beta
    • reasoning effort
    • AI Agent
    • Codex

©2026 All rights reserved. Made with Jekyll and ♥