少点错误 2024年11月02日
Science advances one funeral at a time
index_new5.html
../../../zaker_core/zaker_tpl_static/wap/tpl_guoji1.html

 

文章探讨了科学领域中虽强调创新,但诸多机制却倾向于渐进式而非革命性进步的现象。以多位科学家的经历为例,说明新的科学观点常遭排斥,直至多年后才被认可。同时指出在对齐研究中,人们多倾向安全的研究领域,而真正关键的创新往往来自敢于突破常规的人。

🧐许多科学机构虽强调创新,但实际机制如同行评审、资金决策、学术等级等,明确激励的是渐进式而非革命性的进步,如托马斯·库恩指出的范式转换现象。

😔多位科学家的新发现或理论起初遭排斥,如Ignaz Semmelweis提出医生应在病人间洗手,Barbara McClintock发现基因可在染色体间跳跃等,多年后才被认可。

💡在对齐研究中,人们因职业资本等因素,多倾向“安全”的研究领域,但真正关键的创新将来自敢于突破常规、质疑既定前提并追求初听荒谬想法的人。

Published on November 1, 2024 11:06 PM GMT

Major scientific institutions talk a big game about innovation, but the reality is that many of the mechanisms designed to ensure quality—peer review, funding decisions, the academic hierarchy—explicitly incentivize incremental rather than revolutionary progress.[1] 

Thomas Kuhn's now-famous notion of paradigm shifts was pointing at precisely this phenomenon. When scientists work within what Kuhn called "normal science," they're essentially solving low- to medium-stakes puzzles within their field's accepted framework. While it's fairly easy to evaluate the relative quality of work that occurs within any given paradigm, Kuhn argued it's nearly impossible for scientists to reason about the relative power of different paradigms for a given field—especially when they have already drank the paradigmatic kool-aid.

Max Planck captured this idea succinctly in his biting statement that "science advances one funeral at a time."[2]

There are no shortage of examples of this occurring throughout the history of science:

The takeaway here is quite relevant (and not all that unfamiliar) for alignment research. The still-young field attracts brilliant people who want to help solve the problem—and, by virtue of their technical chops, also care about their career capital. In attempting to check both of these boxes, many naturally gravitate toward "safer," already-somewhat-established research areas. However, when we polled these very researchers, most acknowledged they don't think these sorts of approaches will actually solve the core underlying problems in time. This seems quite familiar to the old story of incentives driving forward incremental work when what is desperately needed are breakthroughs.

The alignment innovations that will be most-critical-in-hindsight will have come from people who were willing to step outside the bounds, question the premises everyone took for granted, and pursue ideas that initially sounded ridiculous. 

Got a crazy hunch that doesn't fit nicely into the current alignment landscape? Come talk to us at EAG Boston—or apply to work on your idea with us here.

  1. ^

    This is not to say that incremental progress is unimportant or that revolutionary progress is all that matters—only that mainstream science is mostly in the business of operating under established paradigms rather than creating new ones.

  2. ^

    While this is the better-remembered variant, Planck's actual statement was "a new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents die, and a new generation grows up that is familiar with it." But this would have made for too long a title.



Discuss

Fish AI Reader

Fish AI Reader

AI辅助创作,多种专业模板,深度分析,高质量内容生成。从观点提取到深度思考,FishAI为您提供全方位的创作支持。新版本引入自定义参数,让您的创作更加个性化和精准。

FishAI

FishAI

鱼阅,AI 时代的下一个智能信息助手,助你摆脱信息焦虑

联系邮箱 441953276@qq.com

相关标签

科学创新 范式转换 对齐研究
相关文章