思考与写作daemonology.net
谈谈人生的用法
On the Use of a Life
2020 — 2022，它一次次回到首页
黑点是上首页（≥ 50 分）的那次投稿，空心点是没冲上去的投稿，点越大分数越高。
导读
Colin Percival 在文中回顾自己的职业选择，比较学术界与工业界/商业环境在做研究上的差异，并解释为何选择以独立开发者/创业者身份经营 Tarsnap 等事业。他认为学术界受经费、终身教职和委员会牵制，未必适合做真正新颖的工作；而商业上的独立让他能自由追随兴趣、正面解决困难问题。文章更像是对“怎样度过一生”和如何安排研究工作的个人反思。
为什么经久不衰：它把科研理想与体制现实摆在一起，对每个在学术、工业或独立工作之间权衡的人都持续有启发。
HN 评论精选
赞同作者：工业界更能产生实际影响，真正投入一个难题会打开大量研究空间。
我强烈赞同 Colin：我在工业界做高影响力、真正新颖工作的机会远多于学术界；学术系统不是产生巨大影响的地方，每年 50 万或 100 万美元的经费规模也不是创新扩展的方式。Colin 经验的核心是，真正承担并试图解决一个现实难题，很快就会看到大量研究空间。
killjoywashere查看英文原文
I'm in violent agreement with Colin here. I have had far more opportunity to do seriously novel work on hard hitting problems in industry than I could ever accomplish in academia. And I have an academic appointment. Which, in pandemic land, pretty much means I have a .edu email address that I also have to attend to. On the industry side, I generate and move data at scale and my teams produce industrial strength models where the SoTA won't be attempted again for years, if ever. Household brands are changing their businesses based on we've done so far, and we're just getting started. The academic system is not where you go to have staggering impact in your lifetime. $500k or $1M a year isn't how innovation scales. Eisenhower didn't pay 100 labs $5M each to come with possible ballistic missile designs. He told Bennie Schriever: I need to deliver a payload to Moscow in 30 minutes, gave him $500M, and kept everyone out of Schriever's business. And, have a read of Rickover: http://ecolo.org/documents/documents_in_english/Rickover.pdf Finally, I would point to the real meat of Colin's experience: if you take on a problem, one problem, like really take it on and try to develop it, you will find 100 PhDs worth of work in short order. Mind-breakingly hard problems abound when you actually try to solve hard problems in the world.
个人经历：学术界的经费压力和不快乐促使他转向软件业，但仍在体制外继续数学研究。
我也曾是好学生，原以为会走学术道路，并在澳大利亚 CSIRO 做过全职研究、拿过奖。最终让我离开的是学术界普遍不快乐：不断为经费争吵、 frustration，还要迎合大矿业/石油公司提出有明确经济收益的课题。后来我投身软件行业，但仍自学数学并有所收获，觉得自己选对了。
nsainsbury查看英文原文
I'm nowhere near as smart as Colin but I'll also share here a little personal story which relates to the topic. I was also a very high achieving student in high school and university and was similarly all set for a career in academia (also studying mathematics). In my final year however, I had a full-time position doing research with CSIRO, which is a leading research organisation in Australia. I did some interesting work there - applying neural nets for classifying micro-seismic events around mine sites, and won some awards for my research. If I had wanted to, I could have stayed on and continued down that path. But I didn't. What ultimately pushed me away was everyone I bumped in to in academia was so unhappy. There was constant bickering and frustration around getting funding (a common sentiment in the division I worked in was that you had pander to big mining/oil companies and propose research topics with clear financial gain for them). It was not a happy place to be, and at the end of my time there I jumped head first in to a software job instead. I later found time to still do mathematics on my own, and have written about that journey and shared it previously on HN: https://www.neilwithdata.com/mathematics-self-learner and have had some other little successes in my life that make me feel like I made the right choice: https://www.neilwithdata.com/how-i-built-bbsmart Tangentially, this I think is also why I'm more open to hearing ideas from organisations like Numenta, and seeing research done outside of academia by folks like Stephen Wolfram. I think increasingly much of the most novel interesting research will be done outside of traditional academia.
反驳：学术界孕育了许多核心技术，并能保护研究者免受市场压力；作者可能从个人坏体验过度概括。
“学术界是做新颖研究的糟糕地方”这一说法，与大量源自学术界的核心技术相矛盾：ENIAC、互联网、万维网、编程语言设计的许多核心思想、AI/神经网络的历史、算法与数据结构研究等。这不一定推翻作者观点，但我觉得他忽略了学术界相对创业的主要优势：让研究者免受市场效应干扰；对开发周期很长或难以商业化的技术，创业未必是替代方案。
ookdatnog查看英文原文
> academia is a lousy place to do novel research This observation seems a bit at odds with the abundance of core technologies that originated in academia. The ENIAC (and a good chunk of the history of early computers), the internet, the world wide web (counting CERN as an academic institution here), most of the core ideas in programming language design (most programming languages we use aren't academic, but most of their ideas can be traced back to academia), most of the history of AI (including neural nets), probably most published work on algorithms and data structures, ... And that's just talking about CS. This doesn't necessarily disprove the author's point: most of the things I've listed are quite old, academia may have worsened in the meantime. But I think they glossed over the main advantage of academia over entrepreneurship: shielding the researcher from market effects. Although investors can do the same thing, as a system academia is more reliable, and for some technologies which take very long to develop or are too hard to monetize (such as algorithms) I don't think entrepreneurship offers an alternative. EDIT (small addition): I think the author is also a bit quick to generalize from "I had a lousy experience in academia" to "academia is a lousy place to do nover research".
批评学术 IP 制度与企业化：大学专利垄断和应用研究被既有企业利益绑架。
学术界不是做有用新颖贡献的地方，干扰因素太多。大学拥有你产生的知识产权，并乐意把独家许可交给没兴趣开发它的实体；垄断企业也不以培育新竞争著称。如今美国学术界的应用研究，最好理解为为制药、科技、工业化学等既有利益集团做企业研发，蓝海可再生能源研究经费极少，“学术自由”已成神话。明显的解决办法是取消学术 IP 的独家许可（如废除 Bayh-Dole 法案）。
photochemsyn查看英文原文
Academics are not the place to go to make useful novel contributions, there are too many confounding factors. First, the university owns whatever IP you come up with, and will happily exclusively license that IP to entities with no interest in developing it. Monopolistic corporate power centers aren't known for fostering novel competition to their established businesses. This is why it took so long - and serious capital - to get something like Tesla going, as the likes of Toyota, Ford, GM etc. were closely tied to Exxon, Chevron etc. via their investors, and had no interest in developing electric cars. Any kind of applied research in the US academic sector these days is best understood as corporate R & D for established interests in pharmaceuticals, technology, industrial chemistry, etc. You're certainly not free to do blue-skies renewable energy research (budgets for that are still miniscule at best, and have been since the 1980s), for example. The job description has become indistinguishable from that of in-house corporate researcher - narrowly defined assigned problems are what you get to work on, and 'academic freedom to pursue the research wherever it leads' is a quaint myth. The obvious fix is to eliminate exclusive licensing of academic IP (i.e. repeal the 1980s Bayh-Dole Act), which would force corporations to spend capital on their own private R & D divisions if the wanted exclusivity, and simply make any university-held patented discovery available to any entrepreneur who wanted to develop it for a small flat fee. Until then the author is 100% correct about entrepreneurship being the only way out of the trap.
历次讨论
同类还值得读
保持你的身份认同足够小
Keep Your Identity Small (2009)
Paul Graham 2009 年的一篇短文：人一旦把某个话题当成自己身份的一部分（宗教、政治、编程语言皆然），讨论就会退化成站队和争吵，因为此时争论的不再是事实，而是"我是谁…
思考与写作上首页 5 次 · 2009–2021paulgraham.com编辑精选苏格拉底式教学法：通过提问而非讲授来教学
The Socratic Method: Teaching by Asking Instead of by Telling
记录了作者用纯提问的方式，引导一群小学生从零推导出二进制计数法的课堂实验，展示了苏格拉底式教学法的惊人效果。
思考与写作上首页 5 次 · 2011–2022garlikov.com你不能说的话
What You Can't Say (2004)
Paul Graham 2004 年的经典文章，探讨那些因道德时尚而不能公开说出的观点。
思考与写作上首页 5 次 · 2014–2022paulgraham.com编辑精选龙暴君的寓言（2005）
The Fable of the Dragon-Tyrant (2005)
哲学家 Nick Bostrom 于 2005 年写的一则寓言：一条每天吞噬数千人的巨龙（隐喻死亡与衰老）长期被人类当作不可抗拒的自然规律接受，直到有人提出并最终用导弹将其杀死，…
思考与写作上首页 5 次 · 2009–2021nickbostrom.com编辑精选现实有着惊人的细节量
Reality has a surprising amount of detail (2017)
2017 年的博文指出，现实远比我们以为的充满细节：我们习惯用简化的心智模型看世界，直到真正动手做事（比如作者举的造楼梯例子）才发现处处是未曾预料的细节。
思考与写作上首页 4 次 · 2018–2025johnsalvatier.org编辑精选