Jones Codes

All posts →

Agentic Research →

Projects →

Home

❯

tags

❯

Applied AI

Applied AI

Nov 08, 20261 min read

Building and shipping real agentic products, not demos. A coaching agent that plans, logs, remembers, and schedules. Live, with users, on a tool-calling loop. The work here is making an agentic feature ship and hold up in production. The failure-mode writeups are the depth underneath it: I go deeper than most on how agents break, because that’s what decides whether one is shippable.

9 items with this tag.

  • Nov 17, 2026

    A Workout Title Wrote to My Coach's Memory

    • applied-ai
    • agent-reliability
    • security
    • prompt-injection
    • memory-poisoning
    • testing
  • Nov 08, 2026

    What It Takes to Serve an LLM to a Million People

    • ai-infrastructure
    • distributed-systems
    • applied-ai
  • Oct 23, 2026

    Agents Playing Telephone and Hanging Out in Your Bank Account

    • applied-ai
    • agent-reliability
    • security
    • agent-isolation
    • tool-misuse
  • Oct 08, 2026

    Give an Agent a Goal and a Gap

    • applied-ai
    • agent-reliability
    • security
    • agent-isolation
    • law
  • Jul 01, 2026

    Every Model Could Already Do It

    • applied-ai
    • agent-reliability
    • production
    • shipping
    • security
    • prompt-injection
  • Jun 21, 2026

    You Can't Secure an Agent From the Prompt

    • applied-ai
    • agent-reliability
    • security
    • method
  • Jun 21, 2026

    Researching agentic success & failure patterns

    • applied-ai
    • agent-reliability
    • prompt-injection
    • method
  • Jun 19, 2026

    The Trust Boundary Inside Your Prompt

    • applied-ai
    • agent-reliability
    • shipping
    • security
    • prompt-injection
  • Jun 19, 2026

    Prompt Injection, and Asking an Agent for Its Tools

    • applied-ai
    • agent-reliability
    • security
    • prompt-injection

Subscribe

Get an occasional email when I publish something new. No spam, unsubscribe anytime.