AI做为生产力工具，还是蛮不错的

前一阵子试过用AI做画，效果尙可。要想用AI创做出理想的画作，关键词很重要，所以，你可能需要测试很多的关键词。这阵子ChatGPT蛮流行的，也来整整，体验下。

https://openai.com/

到openai上去注册，有个手机验证码一直无法完成，这有点和Gmail注册有点像，都是卡在这一关上。先试了下免费的接码平台，发现都用过了，没办法注册。折腾了半天，最后想到了万能的淘宝！

果然，拍了5块大洋就搞到了一个帐号！openai测试都还好，但是ChatGPT却是登不上，老是显示一个1020的错误。谷歌上折腾半天，终于登上了！

把准备好的几个问题拿来测试，感觉很有潜力，比如替代搜索引擎、百科、知乎.......

ChatGPT还是蛮全能蛮给力的，回答得非常不错！它感觉是对数据做了第二次的处理，综合，再显现给用户。和搜索引擎直接显示网页相比，ChatGPT无疑更深入、更智能，感觉更纯粹、简洁，效率更高。

用open apikey也测试了下，蛮顺利的，就是速度有点慢。代码如下：

let { Configuration, OpenAIApi } = require('openai')

const configuration = new Configuration({
  apiKey: "sk-Y27xxxxxxxxx",
});

const openai = new OpenAIApi(configuration);

async function main(){
  let prompt = "中医如何治疗失眠?"
  console.log(prompt)
  let response = await openai.createCompletion({
    model: "text-davinci-003",
    prompt: `${prompt}`,
    temperature: 0, // Higher values means the model will take more risks.
    max_tokens: 3000, // The maximum number of tokens to generate in the completion. Most models have a context length of 2048 tokens (except for the newest models, which support 4096).
    top_p: 1, // alternative to sampling with temperature, called nucleus sampling
    frequency_penalty: 0.5, // Number between -2.0 and 2.0. Positive values penalize new tokens based on their existing frequency in the text so far, decreasing the model's likelihood to repeat the same line verbatim.
    presence_penalty: 0, // Number between -2.0 and 2.0. Positive values penalize new tokens based on whether they appear in the text so far, increasing the model's likelihood to talk about new topics.
  })
  let res = response.data.choices[0].text
  console.log(168, "req:", res)
}
main()