- tags: AI
AI Burnout
Links to this note
在 AI 时代继续热爱编程
tags: AI, AI Burnout 随着工作年限积累,在当前这个环境下我陷入一种职业倦怠里:一种苦练多年的技艺无处施展的无所适从。 当计算机算法逐步碾过象棋和围棋时,那些热爱象棋和围棋的人应该会有同感吧。 当然我也在一直积极的尝试挣脱这种状态:消极的使用 AI。毕竟像我这样脑袋空空没有多少创意的人是无法感受 AI 所带来的美好愿景:Making things。 好的产品创意从来都是一个稀缺品,我除了制造一堆写完再也没有碰过的 AI Slop 之外,AI 之于我成了一个献祭大脑助长懒惰的工具。 最近读到的一篇文章: How to keep enjoying programming in a world of LLMs 给了我很大的启发, 让我来重新审视我和 AI 的关系。作为一个热爱编程的开发者,我享受编程过程中的学习和发现带来的乐趣和成就感。 毋庸置疑,用 AI 节省了我大量的时间,但是催生出了一种失控感,且剥夺了乐趣和成就感。 在懒惰和不甘于人后的焦虑下我心怀抵触的使用 AI,工作完成的很快而且好像还不错,我则看似掌控了什么了不得的工具, 但是我知道这工具不属于我,不像我电脑里装的 Emacs 那样属于我:运行在我自己的电脑上、我自己的配置、 可以不依赖网络运行。实践是检验真理的唯一标准,我已经开始参考这篇文章进行实践: 用 AI 进行计划,自己进行大部分的编程和实现,可以留下一些无聊、低风险、重复的工作给 AI,让 AI 来完成; 让 AI 执行 review-cycle:编写代码 -> AI review -> 调整/重复; 让 AI 研究不是仅仅生成总结:让 AI 列出来源,自己验证并搞懂; 不要只和 AI 聊天,和真实的人建立真实的连接。 当计算机算法战胜了围棋,依然有人热爱围棋。The Problem is not the AI Code, but Nobody Knows Anything Anymore 中提到当前 AI 还只是生成平均水品的代码,Coding is NOT solved 中也提到:那些认为大模型生成的代码足够好的人大概率写代码不如 AI, 所以当 AI 才可以编写平均水平的代码,我为什么不热爱编程呢?未来不远处即使 AI 能编写超高水平的代码,我会选择向 AI 学习编程,而不再献祭大脑。
Coding is NOT solved
tags: AI Burnout, Resist “AI” source: https://blog.alexewerlof.com/p/coding-is-not-solved Coding is not only about creation, but maintenance, reliability, security, scalability, etc. for scaled production software. Here’s the exceptions: Personal software POC(proof of concept) Weaponized AI Those who claim LLM-generated software is good enough: ❌ Haven’t written code in ages ❌ Cannot spot if their code figuratively had 6 fingers! ❌ Have a low bar for what good looks like ❌ Don’t care about quality or NFR ❌ Have difficulty understanding an S-curve ✅ Are honest: AI genuinely writes better code than them ...
The Problem is not the AI Code, but Nobody Knows Anything Anymore
tags: AI Burnout, Resist “AI” source: https://www.ssp.sh/brain/the-problem-is-not-the-ai-code-but-nobody-knows-anything-anymore/ For now, AI can’t drive itself. And it’s hard to maintain large codebase. AI produces average code or levels up small codebase to average. Here’s a smimilar post:Coding is NOT solved
How to keep enjoying programming in a world of LLMs
tags: AI Burnout, Resist “AI” source: https://discourse.haskell.org/t/how-to-keep-enjoying-programming-in-a-world-of-llms/14705 Read if you like programming Researching to understand Instead of letting AI Agents summarization for you, Let AI agents write the results with source links. Understand it and make yourself be better at the domain, even better than the agent. You make the decision. Change the coding work flow AI Plans and then I code instead of “AI Do them All” Use agents to do the low-risk or boring or repeatable jobs that you’ve already known good. ...