AnqiRocks Profile picture
Feb 10 5 tweets 1 min read
#闲聊
有个朋友(在大厂做算法),说在高强度的工作状态下,长时间和异常异性同事一起待着,很容易搞出事。他以前不理解,现在忽然理解了..

我说当然了,国内大厂和异性同事一起待着的时间至少是伴侣的2倍。而且工作强度大,带来的高压、高兴奋状态下的激素效果,也有一定致幻作用。
我们之前爬雪山,也是类似的场景(高压高兴奋、长时间相处)。而且相对来说更隔绝。非常容易出情侣。
但这算是特殊场景下的适配结果。如果后续场景发生很大的迁移,就很不稳定... 爬山的情侣,除了后来一起做户外创业的情侣,基本都分手了。

大厂情侣,要是分手了,基本为了大家体面,其中一个人要活水换部门换团队。退出成本还是挺高的。
后来就很理解赚大钱的行业很容易乱搞哈哈哈(排除那些权色交易)。
因为 high pressure + anxiety +long working hours, 就很难不对身边的异性产生荷尔蒙。
而且这个时候甚至发生关系,对两个人都是获得愉悦激励的事情哈哈(因为性行为确实可以帮助减轻压力,缓解焦虑和抑郁)
这个时候,忽然想到了《色戒》🤣

• • •

Missing some Tweet in this thread? You can try to force a refresh
 

Keep Current with AnqiRocks

AnqiRocks Profile picture

Stay in touch and get notified when new unrolls are available from this author!

Read all threads

This Thread may be Removed Anytime!

PDF

Twitter may remove this content at anytime! Save it as PDF for later use!

Try unrolling a thread yourself!

how to unroll video
  1. Follow @ThreadReaderApp to mention us!

  2. From a Twitter thread mention us with a keyword "unroll"
@threadreaderapp unroll

Practice here first or read more on our help page!

More from @AnqiRocks

Feb 10
#ChatGPT

AI语言生成领域,有不同量级的模型,谷歌的LaMDA参数为1350万,ChatGPT模型的参数为1750万,英伟达-微软的MT-NLG模型参数高达5300万,然而参数越高,成本也越高。大模型优点是功能齐全,各类问题都能回答,不过精准度和成本也高。
ChatGPT的费用至少包括训练费用、对话使用费、硬件软件费用,Semianalysis估算,一次性训练用约为8.4亿美元,每天的硬件维护费为69.4万美元,每次对话费用为0.36美分。
按此估算,ChatGPT一亿的月活用户每个月只做一次对话,每月也至少需要5682万美元(约合3.86亿美元)的维护费,这还不包括除开文字标记等人工费用,以及之前一次性的训练费用。
Read 4 tweets
Feb 9
#5am
在做搜索的时候,学到一个很重要的进步方式分类: exploration & exploitation

Exploration, 一般是引入新的方法或者参数优化目标。搜索优化中,主要是算法上的优化。

Exploitation, 指的是用已有的知识和方法,最大限度优化目标。搜索优化中,主要是通过 case study, 发现现有算法排序中
还有哪些问题。通过解决这些问题来迭代。
Exploration
一般很难预测,对于搜索,尤其是模型上的迭代。胜率低。bert成了,xlnet废了。很难解释。深度学习模型,俗称“炼丹”。
但万一成了,就是大招。

Exploitation
就是通过对用户query进行大量的case study,会更理解用户的需求分布。成功率高。
但会过拟合当前样本。
Read 6 tweets
Feb 8
When chatting with friends about a new technology, reliable people generally discuss three things:

1. Technology: implementation constraints, model papers, technological competition, etc.
2. Use case: users, scenario, pain points
3. Investment: valuation, ongoing rounds, existing investors
For example, ChatGPT:
1. Technology: LLMs, training data, model, computational power
2. Use case: ChatGPT, perplexity.ai etc.
3. Investment: valued at 29B USD, Microsoft is the main investor, previously invested a total of 3B USD.
Read 4 tweets

Did Thread Reader help you today?

Support us! We are indie developers!


This site is made by just two indie developers on a laptop doing marketing, support and development! Read more about the story.

Become a Premium Member ($3/month or $30/year) and get exclusive features!

Become Premium

Don't want to be a Premium member but still want to support us?

Make a small donation by buying us coffee ($5) or help with server cost ($10)

Donate via Paypal

Or Donate anonymously using crypto!

Ethereum

0xfe58350B80634f60Fa6Dc149a72b4DFbc17D341E copy

Bitcoin

3ATGMxNzCUFzxpMCHL5sWSt4DVtS8UqXpi copy

Thank you for your support!

Follow Us on Twitter!

:(