This is the third time I talk about ChatGPT. The first two articles mainly talked about the concept of ChatGPT, the reason why it became popular and its impact on people, as well as the possible security risks and regulatory issues that ChatGPT will face. This article mainly talks about the scenarios of ChatGPT and the logic of dealing with problems.
This time I deliberately used the ChatGPT Chinese web version to experience it. And asked ChatGPT to help me with a few questions. The following is a java encryption and decryption tool class written by ChatGPT for me.
The following is a java encryption and decryption tool class written by ChatGPT for me.
import javax.crypto.Cipher;
import javax.crypto.KeyGenerator;
import java.security.Key;
import java.security.KeyPair;
import java.security.KeyPairGenerator;
import java.security.SecureRandom;
public class EncryptionUtil {
/**
* 对称加密算法AES加密
* @param content 待加密的字符串
* @param password 秘钥
* @return 加密后的结果
* @throws Exception 异常
*/
public static byte[] aesEncrypt(String content, String password) throws Exception {
KeyGenerator kgen = KeyGenerator.getInstance("AES");
SecureRandom secureRandom = SecureRandom.getInstance("SHA1PRNG");
secureRandom.setSeed(password.getBytes());
kgen.init(128, secureRandom);
Key key = kgen.generateKey();
Cipher cipher = Cipher.getInstance("AES");
cipher.init(Cipher.ENCRYPT_MODE, key);
return cipher.doFinal(content.getBytes("utf-8"));
}
/**
* 对称加密算法AES解密
* @param content 待解密的字符串
* @param password 秘钥
* @return 解密后的结果
* @throws Exception 异常
*/
public static String aesDecrypt(byte[] content, String password) throws Exception {
KeyGenerator kgen = KeyGenerator.getInstance("AES");
SecureRandom secureRandom = SecureRandom.getInstance("SHA1PRNG");
secureRandom.setSeed(password.getBytes());
kgen.init(128, secureRandom);
Key key = kgen.generateKey();
Cipher cipher = Cipher.getInstance("AES");
cipher.init(Cipher.DECRYPT_MODE, key);
byte[] result = cipher.doFinal(content);
return new String(result, "utf-8");
}
/**
* 非对称加密算法RSA加密
* @param content 待加密的字符串
* @param publicKey 公钥
* @return 加密后的结果
* @throws Exception 异常
*/
public static byte[] rsaEncrypt(String content, String publicKey) throws Exception {
KeyPairGenerator kpg = KeyPairGenerator.getInstance("RSA");
kpg.initialize(1024);
KeyPair kp = kpg.generateKeyPair();
PublicKey pubKey = kp.getPublic();
PrivateKey priKey = kp.getPrivate();
Cipher cipher = Cipher.getInstance("RSA");
cipher.init(Cipher.ENCRYPT_MODE, pubKey);
return cipher.doFinal(content.getBytes());
}
/**
* 非对称加密算法RSA解密
* @param content 待解密的字符串
* @param privateKey 私钥
* @return 解密后的结果
* @throws Exception 异常
*/
public static String rsaDecrypt(byte[] content, String privateKey) throws Exception {
KeyFactory kf = KeyFactory.getInstance("RSA");
PKCS8EncodedKeySpec spec = new PKCS8EncodedKeySpec(Base64.getDecoder().decode(privateKey));
PrivateKey priKey = kf.generatePrivate(spec);
Cipher cipher = Cipher.getInstance("RSA");
cipher.init(Cipher.DECRYPT_MODE, priKey);
byte[] result = cipher.doFinal(content);
return new String(result);
}
}
I asked ChatGPT to help me write a resignation letter.
尊敬的公司领导：
您好！在这样一个特殊的时刻，我写下这篇辞职申请书，表达我的离职意愿。
回顾在公司工作期间，我感谢公司给予了我很多机会和成长空间，也让我经历了很多挑战与收获。无论是从专业技能还是个人素养上，我都有了极大的提升和进步，同时也结识了很多优秀的同事和朋友。
然而，在个人生活和职业发展方面，我需要做出一些调整和抉择。经过反复思考和评估，我认为现在是一个适合我离开公司的时机。我深感抱歉，但希望您能够理解我的决定。
为了保证公司业务的连续性和稳定性，在此承诺在离职前完成手头的工作，并将所有的工作交接到公司指定的同事，确保工作交接的安全、完整和规范。
最后，再次感谢公司的支持和信任，祝愿公司越来越好，也祝福所有的同事和朋友们取得更加美好的职业与人生。
此致
敬礼！
申请人签名
日期I will ask ChatGPT to help me write a strategy on how to find a girlfriend.
寻找女朋友是一件需要耐心和方法的事情。以下是一些可能有用的策略：
了解自己：首先，你需要了解自己想要什么样的女孩，以及你自己能够提供什么。这将有助于你更有针对性地寻找合适的对象。
广交朋友：在社交场合多认识人，在你感兴趣的领域里结交志同道合的朋友，他们或许会介绍给你一个合适的女孩。
参加活动：参加各种活动和社交聚会，尤其是与你擅长或者热爱的活动相关的聚会，会有更多的机会和女生认识。
在线社交：现在的社交方式已经多种多样，可以通过线上社交平台来认识新的女生，例如微信、陌陌、探探等，注意安全问题。
做个好男人：善良、有礼貌、有责任心、有幽默感、身体健康等优秀品质会吸引更多的女孩子注意到你。
加强自我培养：除了外表条件，内在素质也是很重要的，读书、学习技能、开阔见识、丰富爱好，这些都能让你更加有吸引力。
别心急：恋爱是一件需要缘分的事情，不要着急，多享受单身时光，保持好心态，在合适的时间会遇到合适的人。
总之，寻找女朋友需要耐心和方法，同时也需要保持良好的信心和心态。希望以上策略对您有所帮助。For the above three questions, we can see that ChatGPT’s answers are still very official. In fact, we can find similar answers to these things through simple Baidu or Google. It's just that ChatGPT aggregated for us.
I asked the last question: Where is the best place to live in Shenzhen, and ChatGPT answered as follows:
深圳是一个发达的现代化城市，拥有很多宜居的区域。以下是一些比较适宜居住的地区：
福田中心区：福田中心区拥有完备的基础设施和便捷的交通网络，生活便利程度高，商业、文化、娱乐等资源丰富。
南山区：南山区是深圳的科技与文化区，有很多海滨别墅和高端公寓，同时也有很多优秀学校和医院，适合居住的环境优美。
罗湖区：罗湖区是深圳的老区之一，但近年来得到了很好的发展，是深圳的商业、交通和金融中心之一，而且还有很多古建筑和文化景点。
龙岗区：龙岗区是深圳最大的行政区之一，有很多新型社区和高档住宅小区，适合家庭或个人居住，同时也有很多优秀的学校和商业配套。
宝安区：宝安区是深圳的工业中心和物流枢纽，但随着城市规划的不断完善和发展，已经形成了一些高端住宅区和优质社区。
需要注意的是，每个人对于居住环境的需求、生活习惯和经济实力都不同，选择居住地时应根据自身情况进行选择。以上推荐仅供参考。This answer is less than satisfactory. I feel that ChatGPT will always be bureaucratic and moderate. This reminds me of what I said before that the data center can solve 90% of the problems. But this is a wrong statement, it should be 90% of the problem can be solved.
We can see that the things that ChatGPT can answer are some common things, just like we write code to encapsulate and abstract. If it is a personalized thing, ChatGPT can't solve it well. So I think ChatGPT can only solve 90% of the problem, because ChatGPT can't directly read people's hearts, has no pattern, no emotion, and can't innovate. It can only train step by step based on data, and provide existing experience in the world for easy retrieval. But it is undeniable that ChatGPT can indeed be used as a very useful auxiliary work, able to write code, poetry, papers, etc.
ChatGPT's future development forecast (from China Daily):
ChatGPT's technology development trend
The tools ChatGPT is based on, called Large Language Models (LLMs), work by learning statistical patterns in language from huge online text databases. Based on the existing basic principles and technologies, ChatGPT will inevitably further improve its technical level in the future development process, such as improving the quality of search results.
The way to improve technology, first, is to add massive data, especially real-time new data, to the original text database, correct the false, biased or outdated information, and based on the updated text database, carry out Further model training to keep pace with the times with technical models and application tools.
The second is that it is necessary to conduct specialized model training for certain specific fields to achieve technical improvement. In contrast, the intelligent question answering and retrieval that were born earlier are easier to be identified because the model based on them is relatively simple and the degree of intelligence is not high; while the technical progress of LLM-based ChatGPT enables it to perform logical screening, targeting at The entered question provides a generally reasonable answer. However, its answers are still unreliable in nature, and it will also produce wrong answers when answering some professional field questions. Therefore, in some specific and highly specialized fields, it is necessary to provide specialized text databases and adjust models and data based on targeted training methods. Taking the study of problems in the fields of literature, history, philosophy and science and engineering as an example, the methodologies and logical thinking methods required are different; in different scenarios of the same discipline, the required analytical ideas are also different. Therefore, the future development of ChatGPT requires specialized model training for specific fields.
ChatGPT's commercial development trend and application fields
The further development of ChatGPT is also inseparable from the application and commercialization of technology. Its commercialization method, one is to cooperate with existing large-scale commercial platforms (such as Microsoft, Google, etc.). Although some people say that the cost of LLM-driven search still needs to be greatly reduced, at the same time, Microsoft plans to integrate ChatGPT into their search engine Bing and is expected to launch it in March. In addition, existing large-scale commercial platforms may also conduct technology research and development on their own. Whether the two achieve product substitution depends on the development speed of the two methods.
No matter which method is used to achieve commercial development, it must first focus on specific areas with high concentration of resources and funds and strong technical applicability. For example, technology applications in the financial field are relatively extensive and funds are abundant, and financial technology and financial data governance are the first to be developed in the financial field; commercialization in the field of culture and entertainment is relatively rapid, and it may also become the primary field of technology application; related to daily work In the field of office software and other fields, if partial automation of office work (such as automatic document writing and PPT production) can be realized based on this technology, due to its wide application and large user demand, it may also become the first field of technology application development.
The phased development trend and commercial competition trend of ChatGPT
As a new technology and tool, ChatGPT is likely to be similar to other new technologies and tools. It will receive more research and attention when it first enters the public eye, and a large number of corresponding start-up companies or listed companies will be born at home and abroad. layout. But after a period of fierce competition, not many companies actually succeed and survive. This is also the situation reflected in the technology maturity curve proposed by Gartner. ChatGPT is likely to be at the peak of excessive expectations in this curve, and it is about to reach the trough of the bubble. After its application and implementation, the application effect in some specific fields and scenarios may not be as expected.
In this process, many start-up companies or listed companies will gradually withdraw from the market. From a domestic point of view, it is speculated that the relevant platforms that survive and successfully implement technology applications are likely to be large platforms with sufficient funds, traffic foundation and existing databases, such as Baidu and Toutiao. Some other innovative companies or small specialized companies may encounter problems such as insufficient training sets and inability to achieve effective training of models, and eventually withdraw from the market. During this period, the investment and stock prices of related listed companies may rise rapidly, but the existing large platforms will still be able to realize the localization of this technology.