home.social

#aiandml — Public Fediverse posts

Live and recent posts from across the Fediverse tagged #aiandml, aggregated by home.social.

fetched live
  1. Learn how to use AI agents to detect malicious repositories and prompt injection safely with read-only audits, evals, and microVM sandboxing. hackernoon.com/fake-recruiter- #aiandml

  2. What happened when I stopped treating LLM tool calls like ordinary function calls and started treating them like distributed transactions. hackernoon.com/i-taught-an-ai- #aiandml

  3. "Reflections On Trusting Trust" meets "Attention Is All You Need" — why the "local AI" hype has a fatal security blind spot. Open weight ≠ Safe. hackernoon.com/revisiting-trus #aiandml

  4. Building plugins to extend OpenAI's Codex AI coding agent for improved AGENTS.md handling, with local overlays, import expansion, source provenance. hackernoon.com/fixing-codexs-a #aiandml

  5. AI and Unintended Consequences for Human Decision Making b.mamund.com/3FY2zhI

    "there’s often no transparency regarding how the recommendations are derived nor what data are used to determine them."

    #api360 #AIandML