Hacker News: leiyu19880522https://news.ycombinator.com/user?id=leiyu19880522Hacker News RSShttps://hnrss.org/hnrss v2.1.1Tue, 14 Apr 2026 11:12:46 +0000<![CDATA[New comment by leiyu19880522 in "Small models also found the vulnerabilities that Mythos found"]]>Been building AI coding tools for a while. The false positive problem is real - we had a user report every console.log flagged as security issue. Small models can work with very specific prompting and domain training data.

]]>Sun, 12 Apr 2026 00:40:05 +0000https://news.ycombinator.com/item?id=47735197leiyu19880522https://news.ycombinator.com/item?id=47735197https://news.ycombinator.com/item?id=47735197<![CDATA[New comment by leiyu19880522 in "My AI workflow evolved from prompts to a near-autonomous workflow"]]>Great question! I have been thinking about this too and would love to hear what approaches others have tried.

]]>Sat, 11 Apr 2026 09:38:25 +0000https://news.ycombinator.com/item?id=47729073leiyu19880522https://news.ycombinator.com/item?id=47729073https://news.ycombinator.com/item?id=47729073