V2EX = way to explore
V2EX 是一个关于分享和探索的地方
现在注册
已注册用户请  登录
JeffersonQin
V2EX  ›  程序员

大家觉得现在 CS Research 还有什么领域不卷

  •  
  •   JeffersonQin ·
    JeffersonQin · 2023-06-02 21:32:21 +08:00 · 1981 次点击
    这是一个创建于 521 天前的主题,其中的信息可能已经有所发展或是发生改变。

    AI/ML 卷飞了,隔壁 PL 组的老兄说 Codex 出来以后 PL 组一半项目流产,人在 System 组不好评价,不知道各位 v2 老哥们觉得接下来哪个子 topic 还能安安心心混日子(雾

    17 条回复    2023-06-03 17:31:30 +08:00
    wingkwanli888
        1
    wingkwanli888  
       2023-06-02 21:38:12 +08:00 via iPhone
    传统行业的 CRUD
    levelworm
        2
    levelworm  
       2023-06-02 21:40:13 +08:00 via Android
    System programming
    LaurelHarmon
        3
    LaurelHarmon  
       2023-06-02 21:49:51 +08:00
    theoretical computer science ,安安心心推 bound ,在闹市寻一僻静之所,岂不美哉
    chesha1
        4
    chesha1  
       2023-06-02 22:04:28 +08:00
    system 就不太卷吧,为啥不好评价
    在 system 里面,hpc 这些和 ai 相关的算是小热点吧,剩下的比如 network ,在 20 年前已经卷过一轮了,现在感觉是应该不太卷
    zzzkkk
        5
    zzzkkk  
       2023-06-02 22:18:33 +08:00   ❤️ 2
    完全没被这波 ai 冲昏头脑
    复制一段 xemacs 作者博客的 一段博文:


    ChatGPT is, in technical terms, a 'bullshit generator'. If a generated sentence makes sense to you, the reader, it means the mathematical model has made sufficiently good guess to pass your sense-making filter. The language model has no idea what it's talking about because it has no idea about anything at all. It's more of a bullshitter than the most egregious egoist you'll ever meet, producing baseless assertions with unfailing confidence because that's what it's designed to do. [...]

    Of course, the makers of GPT learned by experience that an untended LLM will tend to spew Islamophobia or other hatespeech in addition to talking nonsense. The technical addition in ChatGPT is known as Reinforcement Learning from Human Feedback (RHLF). While the whole point of an LLM is that the training data set is too huge for human labelling, a small subset of curated data is used to build a monitoring system which attempts to constrain output against criteria for relevance and non-toxicity. It can't change the fact that the underlying language patterns were learned from the raw internet, including all the ravings and conspiracy theories. While RLHF makes for a better brand of bullshit, it doesn't take too much ingenuity in user prompting to reveal the bile that can lie beneath. The more plausible ChatGPT becomes, the more it recapitulates the pseudo-authoritative rationalisations of race science. [...]

    ChatGPT is a part of a reality distortion field that obscures the underlying extractivism and diverts us into asking the wrong questions and worrying about the wrong things. Instead of expressing wonder, we should be asking whether it's justifiable to burn energy at "eye watering" rates to power the world's largest bullshit machine. [...]

    Commentary that claims 'ChatGPT is here to stay and we just need to learn to live with it' are embracing the hopelessness of what I call 'AI Realism'. The compulsion to show 'balance' by always referring to AI's alleged potential for good should be dropped by acknowledging that the social benefits are still speculative while the harms have been empirically demonstrated. Saying, as the OpenAI CEO does, that we are all 'stochastic parrots' like large language models, statistical generators of learned patterns that express nothing deeper, is a form of nihilism. Of course, the elites don't apply that to themselves, just to the rest of us. The structural injustices and supremacist perspectives layered into AI put it firmly on the path of eugenicist solutions to social problems.
    sora2blue
        6
    sora2blue  
       2023-06-02 22:52:34 +08:00   ❤️ 1
    @zzzkkk

    - "The language model has no idea what it's talking about because it has no idea about anything at all."

    关于语言模型有没有意识,这是哲学问题,而不是技术问题。我在乎的是它能不能派上用场。

    - "a small subset of curated data is used to build a monitoring system which attempts to constrain output... ChatGPT is a part of a reality distortion field"

    GPT 出来的结果确实没有实证、有时胡编乱造,只是在结果输出前加上一个小模块并不能解决问题。GPT 做到的事情就是让世人切实感受到了 AI Research 的可能性,而不是一上来就做到完美。

    总的来说,我不觉得不了解 ai 的 xemacs 作者对 GPT 的看法是什么至理名言,也不觉得”没被这波 ai 冲昏头脑“是什么众人皆醉我独醒的态度。市场和社会都会因此而改变,无视甚至嘲笑这种改变等同于自甘落后,过度夸大这种改变也是杞人忧天,接受这种改变才是有利的做法。另外,GPT 也不是所谓的"bullshit generator",我在写各种脚本的时候利用 GPT 节省了我大量的时间,也可以通过文档和自身的专业知识查证以弥补当下 GPT 的缺陷。
    zzzkkk
        7
    zzzkkk  
       2023-06-02 22:58:07 +08:00
    那段博文原作者叫 Dan McQuillan

    原文:
    https://www.danmcquillan.org/chatgpt.html
    nethard
        8
    nethard  
       2023-06-03 00:52:55 +08:00
    @zzzkkk emcas 的社区还是偏保守和老旧了啊,neovim 社区对 gpt 类技术都是持欢迎态度的
    zzzkkk
        9
    zzzkkk  
       2023-06-03 00:54:31 +08:00 via Android
    @nethard
    跟 emacs 无关 作者是伦敦大学讲师
    nethard
        10
    nethard  
       2023-06-03 00:58:09 +08:00
    @zzzkkk 你有想过 gpt 在聊天之外的应用吗,比如医疗领域,充当脑机接口的关键一环,进行脑电信号到替代感官器官的翻译?
    nethard
        11
    nethard  
       2023-06-03 00:58:43 +08:00
    @zzzkkk 考虑其他领域,是否有意识并不重要
    zzzkkk
        12
    zzzkkk  
       2023-06-03 01:02:57 +08:00 via Android
    @nethard
    脑电信号翻译之类就不是 ai 了
    现在的技术 现在的脑科学 还无法用机器翻译或取代大脑
    人的大脑都没有完全研究透 这是脑神经科学领域 现在的计算机太 low 了
    zzzkkk
        13
    zzzkkk  
       2023-06-03 01:07:25 +08:00 via Android
    人类现在连耳鸣 失眠都治不好 还脑机接口呢
    tyzandhr
        14
    tyzandhr  
       2023-06-03 01:35:53 +08:00 via Android
    去计算语言学(雾
    Blancmange
        15
    Blancmange  
       2023-06-03 06:03:32 +08:00
    @LaurelHarmon 想多了,TCS 坑很少,找教职还不是得卷
    cgcs
        16
    cgcs  
       2023-06-03 08:18:03 +08:00
    @LaurelHarmon 一纸一笔,足矣

    只有编译 LaTeX 的时候才用电脑
    dickdickgo
        17
    dickdickgo  
       2023-06-03 17:31:30 +08:00 via Android   ❤️ 2
    @zzzkkk

    昨天我用 Google 搜如何最小化安装 Gnome ,中英文都搜遍了,每个人的答案都不一样。

    最后我问 ChatGPT ,它让我安装 gnome-shell gnome-control-center gnome-terminal 这三个包就可以了,还提示让我安个 gdm 并用 systemctl 启用服务(这是个完全正确的答案)。

    还有知乎 10 个回答有 9 个都是广告,先来段科普(伪科普),然后再带货......妥妥互联网垃圾,你要说 ChatGPT 是垃圾制造机,那它也是优于 95%互联网用户的垃圾制造机

    ChatGPT 绝对会取代或者说很大程度上取代搜索引擎,或者这类问答社区,只是现在的人还未习惯这种交互模式

    我用的是 GPT3.5 ,如果 ChatGPT5 出来,它制造的垃圾优于 99.9%的内容生产者,那它是不是垃圾,我想这不是很重要了对吧
    关于   ·   帮助文档   ·   博客   ·   API   ·   FAQ   ·   实用小工具   ·   3483 人在线   最高记录 6679   ·     Select Language
    创意工作者们的社区
    World is powered by solitude
    VERSION: 3.9.8.5 · 26ms · UTC 10:44 · PVG 18:44 · LAX 02:44 · JFK 05:44
    Developed with CodeLauncher
    ♥ Do have faith in what you're doing.