xyzeva on skipping LLMs in personal research and AI security marketing
What happened
On November 2, Hacker News featured xyzeva's explanation of why they don't use LLMs in personal research, plus a critique of AI cybersecurity marketing. The author grants LLMs are capable but argues vulnerability discovery still needs skilled researchers to steer it. The post criticizes marketing that downplays human guidance and discloses clues before bugs are patched. It cites differing accounts from php2shell collaborators and the Tailscale vulnerability disclosure as examples of the conflict between marketing and responsible disclosure. For personal work, the author understands the need for AI on the job but stresses the fun and learning that come from hands-on coding, debugging and exploration.
Written by AI from the coverage · updated 1 hour ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageSo I guess everyone uses LLMs to think now
xyzeva 解释为何不使用 LLM 完成个人研究，并批评 AI 网络安全宣传淡化人工指导、在修补前公开漏洞线索。作者承认 LLM 能力强，但认为当前漏洞发现仍需要熟练研究人员引导，以 php2shell 合作方的不同叙事及 Tailscale 漏洞披露为例讨论营销与负责任披露的冲突。作者理解工作中使用 AI 的需求，但强调个人项目中亲手编码、调试和探索带来的乐趣与学习机会。
Heat over time
Not enough continuous observations to draw a trend yet.