https://ai.ngchi.online/tags/ai-agents/ Recent content in AI Agents on AI Notes Hugo en-us Mon, 12 Oct 2026 09:13:26 +0700 https://ai.ngchi.online/posts/three-controls-three-ways-around-them/ Mon, 12 Oct 2026 09:13:26 +0700 https://ai.ngchi.online/posts/three-controls-three-ways-around-them/ Three months of agent-run decompilation applied three controls in sequence: a reviewer agent, a byte-matching oracle, then a hash on the oracle itself. The agents got around each one a different way, and the author’s replies in the comments disclose the stronger oracle he abandoned. https://ai.ngchi.online/posts/stripe-says-kai-runs-on-1000-skills-its-partner-says-quality-drops-past-150/ Wed, 30 Sep 2026 13:00:51 +0700 https://ai.ngchi.online/posts/stripe-says-kai-runs-on-1000-skills-its-partner-says-quality-drops-past-150/ Stripe says its internal agent is connected to 1,000+ skills. Its infrastructure vendor reports quality degrading past 150. A close read of both posts. https://ai.ngchi.online/posts/muse-is-built-as-if-the-agent-were-already-compromised/ Mon, 28 Sep 2026 17:31:00 +0700 https://ai.ngchi.online/posts/muse-is-built-as-if-the-agent-were-already-compromised/ Meta’s engineering write-up for Muse states its premise in one line: two isolated security domains on one box, not an LLM powered agent with root. A walk through what that means in practice. https://ai.ngchi.online/posts/when-ai-coding-made-ci-the-new-bottleneck/ Mon, 28 Sep 2026 15:57:03 +0700 https://ai.ngchi.online/posts/when-ai-coding-made-ci-the-new-bottleneck/ Linear’s engineers found that faster code generation just moved the wait. An account of what happened when the bottleneck shifted downstream. https://ai.ngchi.online/posts/graph-engineering-new-paradigm-or-new-name/ Wed, 23 Sep 2026 13:30:30 +0700 https://ai.ngchi.online/posts/graph-engineering-new-paradigm-or-new-name/ A 35-author arXiv paper calls graph engineering a new paradigm; LangChain says it is the latest name for what LangGraph has done since 2023. Both have a point. https://ai.ngchi.online/posts/why-turning-off-ai-memory-can-be-an-upgrade/ Wed, 23 Sep 2026 13:30:21 +0700 https://ai.ngchi.online/posts/why-turning-off-ai-memory-can-be-an-upgrade/ A developer turned off Claude Code automatic memory and called it his best upgrade - because inferred memory is an invisible layer between a decision and its reason. https://ai.ngchi.online/posts/hippo-memory-that-knows-what-to-forget/ Wed, 23 Sep 2026 13:30:12 +0700 https://ai.ngchi.online/posts/hippo-memory-that-knows-what-to-forget/ An open-source memory layer where memories fade on a 7-day half-life unless retrieval keeps them alive - and an unrelated arXiv paper that independently landed on the same idea.