https://siqi-liu.com/en/categories/ai/Recent content in Ai on Yosgi BlogHugoenFri, 12 Jun 2026 04:28:41 +0000https://siqi-liu.com/en/post/from-extracting-drawing-text-to-placing-2d-annotations-in-a-3d-scene/Thu, 11 Jun 2026 00:00:00 +0000https://siqi-liu.com/en/post/from-extracting-drawing-text-to-placing-2d-annotations-in-a-3d-scene/A practical engineering recap of how a simple drawing text extraction task turned into a 2D-to-3D mapping workflow for digital twins, combining SVG parsing, vision models, geometry rules, registration math, and agent tool design.https://siqi-liu.com/en/post/ai-won-t-replace-software-engineers-it-will-replace-the-parts-of-coding-we-never-loved/Tue, 19 May 2026 00:00:00 +0000https://siqi-liu.com/en/post/ai-won-t-replace-software-engineers-it-will-replace-the-parts-of-coding-we-never-loved/AI will not replace software engineers as a whole. It will remove much of the repetitive, mechanical work and give engineers more leverage to focus on judgment, systems, users, and building better products.https://siqi-liu.com/en/post/your-agent-is-not-inconsistent-because-it-is-dumb-it-just-has-too-many-tool-paths/Thu, 07 May 2026 00:00:00 +0000https://siqi-liu.com/en/post/your-agent-is-not-inconsistent-because-it-is-dumb-it-just-has-too-many-tool-paths/Multi-tool agents become unreliable when the same task takes different tool paths each time. This article explains why Skills should preserve validated execution paths, not one-off results, and why those Skills need to evolve over time.https://siqi-liu.com/en/post/tool-for-control-code-for-analysis3/Wed, 28 Jan 2026 00:00:00 +0000https://siqi-liu.com/en/post/tool-for-control-code-for-analysis3/In large-scale Agent systems, pure Tool-based approaches tend to collapse under context pressure, while pure Code-based approaches introduce excessive latency. Based on real-world experiments, this article introduces a dual-path execution model that uses a “Context Off-Ramp” to switch between Tool and Code execution.https://siqi-liu.com/en/post/why-tool-and-code-fail-differently-in-agent-systems2/Sat, 27 Dec 2025 00:00:00 +0000https://siqi-liu.com/en/post/why-tool-and-code-fail-differently-in-agent-systems2/Tool and Code do not just differ in expressiveness; they also fail at different times. This article explains why multi-round Agent systems amplify that difference.https://siqi-liu.com/en/post/an-mcp-as-code-refactor-and-why-it-did-not-work-the-way-i-expected1/Thu, 18 Dec 2025 00:00:00 +0000https://siqi-liu.com/en/post/an-mcp-as-code-refactor-and-why-it-did-not-work-the-way-i-expected1/A real-world MCP-as-Code refactor on where Code helps, where Tool still fits better, and why Agent runtime needs both.https://siqi-liu.com/en/post/early-experiences-with-building-mcp-tools/Tue, 01 Jul 2025 00:00:00 +0000https://siqi-liu.com/en/post/early-experiences-with-building-mcp-tools/How Cut 6,000 Tokens Down to 500