Gary Marcus 反驳 Jack Clark：Claude Code 并非仅靠规模
Gary Marcus 拿 Anthropic 自家的 Claude Code 打脸 Jack Clark：50 万行符号代码说明专门化没死，别只信 scaling。
Anthropic 联合创始人 Jack Clark 表示 AI 进步仅靠规模扩展就能实现，但 Gary Marcus 引用该公司自己的 Claude Code 予以反驳，指出该工具使用了 50 万行符号代码、harnesses、符号工具和正则表达式。
Marcus 认为专门化系统并非无用，规模扩展也非唯一路径。该争论涉及“bitter lesson”假说与实用工程系统的平衡。
this is so not true. anthropic’s own Claude Code uses harnesses, symbolic tools, regular expressions and 500,000 lines of symbolic code. it’s not scale alone and it *is* specialized. Haider. @haider1 Anthropic's Jack Clark says he underestimated how far AI would improve through scale alone The "bitter lesson" kept proving true: add more compute to generic neural networks, and they keep getting smarter with new abilities "specialized systems and cautious predictions keep losing to scaling"
This report is categorized under AITOP's "Industry" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Gary Marcus, Jack Clark, Claude Code, Anthropic.
Source attribution: curated from Gary Marcus (@GaryMarcus). Cross-source corroboration: 11 independent sources have covered this development (including Lenny Rachitsky, IT之家, Claude), confirming industry-wide signal strength.
