Shulou(Shulou.com)11/24 Report--

The technologies that change people's lives do not need the media to write long speeches to tell ordinary people how good they are, and the power of definition is in the hands of the public.

I don't know if you are like me, some time ago, I was suddenly brushed by ChatGPT.

When the number of ChatGPT users broke through 1 million at the end of last year, it became popular for a while, but it was obviously more out of the circle this time. Almost all the accounts I follow are talking about ChatGPT, everything from how powerful it is to what it can do to how it will upend the world.

In fact, the type of chatbot has a long history.

The history of chat robots can be traced back to the 1960s, when the first chat robot ELIZA was founded by Joseph Weizenbaum, a computer scientist at the Massachusetts Institute of Technology.

ELIZA is designed as a proof of concept for natural language processing and uses simple pattern matching and response to simulate conversations with humans.

The design of ELIZA is based on the concept of a psychotherapist and uses a simple rule-based system to determine the appropriate response. When a user enters a declaration, ELIZA recognizes the keywords and answers the relevant follow-up questions, creating the illusion of the actual conversation.

Although ELIZA is simple, it quickly caused a sensation, with users having meaningful conversations with the chatbot and even revealing personal information to the chatbot.

In the years since the birth of ELIZA, other chat robots have emerged, such as PARRY, developed in 1972 by Kenneth Colby, a psychologist at Stanford University. PARRY aims to simulate conversations with paranoid patients, using cognitive psychology and linguistics to create a more real conversation.

ALICE is an abbreviation for artificial language Internet computer entity, founded by Richard Wallace in 1995. ALICE is designed as an open source platform for building conversational AI systems and uses artificial Intelligence markup language (AIML).

Unlike ELIZA and PARRY, ALICE is designed to be more accessible and intuitive, allowing users to create their own chat bots and customize responses.

In the early 2000s, instant messaging platforms such as AOL and MSN Messenger became popular, and chat robots such as SmarterChild began to provide quick and simple information, such as weather and stock updates. SmarterChild is one of the first chat robots integrated into the instant messaging platform, and quickly caused a sensation, millions of users rely on it for information and entertainment.

In 2007, the launch of iPhone and Siri marked a new era of chat robot technology.

Siri is designed as a personal assistant chat robot, taking advantage of advances in artificial intelligence and machine learning to provide a more complex chat robot experience. With Siri, users can perform a variety of tasks through the dialog interface, such as setting reminders, sending messages, and doing searches.

In recent years, chat robots have been increasingly integrated into various industries, such as customer service, health care and e-commerce.

The company uses chat robots to provide round-the-clock customer support and improve the overall customer experience. With the progress of artificial intelligence technology, chat robots have become more advanced and can handle complex tasks, such as medical diagnosis and financial planning.

So how is ChatGPT different from the chatbots of the past? In fact, the chatbot history of about 700 words above, which I asked ChatGPT to help write, has hardly changed. Compared with the chatbots in the past, it is indeed very strong.

02 of course, some of the media have made it clear that various industries will lose their jobs as a result, and I think it is still too early. Because recently, my friends often ask me, what do you think of ChatGPT? it can generate copywriting. Are you under a lot of pressure? Or whether it can improve your productivity, subvert the way you work, and so on.

In my opinion, at least in my own writing process, its help to me is limited. For me, it is never a problem to write a manuscript. The important thing is to consult the data, understand and build your own point of view.

ChatGPT is a very powerful chatbot, but it's not the AI we thought it was, and it can't really think like humans.

For example, when I told it that I needed a history of the development of chatbots, the text it gave me was very rigid, lacked a lot of details I needed, and the information content did not exceed my own reserves.

So I still need to consult the data and read some papers to understand the key nodes and technical differences of the chat robot, such as the operation mode of the first chat robot ELIZA, and so on.

Then I went back to ChatGPT and told him that what I needed was an article detailing the history of chatbots, including ELIZA, PARRY, ALICE, SmarterChild, and Siri.

Many people believe that search engines are the biggest threat to ChatGPT, which is why both Google and Baidu immediately announced that they would launch their own chat robots.

But at least for now, ChatGPT can't replace search engines. On the one hand, the way you use ChatGPT is quite different from the way you use search engines. On the other hand, the language library used by ChatGPT is still 2021, so it doesn't know a lot of things that are happening now, such as if you ask it if Messi won the World Cup, it will answer no.

However, when it is combined with the search engine, the effect is not necessarily. Microsoft recently announced that the latest Bing has integrated ChatGPT's AI capabilities and is using an updated version of GPT-3.5, the language model used by ChatGPT. Unfortunately, I didn't get an appointment, so I don't know what it will be like to use AI in combination with a search engine.

After all, search engines have not changed over the years, and Google has held more than 90% of the global market share for a long time. It is always exciting to have new possibilities.

Microsoft, on the other hand, may be the big winner after the ChatGPT explosion. The last time AI went so far, it was because of AlphaGo, and the company that made it, DeepMind, belonged to Google. Although AlphaGo and DeepMind's subsequent achievements in AI are surprising, compared with the popularity of ChatGPT, it is clear that there is a lack of imagination in commercial applications.

I'm not a technology blogger, and I don't know much about technology, so I can only look at it from a business point of view.

Judging from the current hot environment, I don't think it's ordinary consumers who really need AI chat robots-- what's most exciting are the capital markets and the tech giants.

03 Why do you say so?

In fact, the popularity of the concept of ChatGPT and even AI chatbot is not uncommon in recent years. The closest to us are meta-universe, Web3.0 and NFT, and before that there are 5G, VR and so on. At that time, it was very popular, and it was the most cutting-edge direction of exploration at that time, but today the volume is much lower.

This is not to say that these concepts and technologies are not developed and useless. I believe these technologies are improving and changing some things, but on the consumer side, the feeling is not so obvious, and it seems to be a long way from the level claimed by the major media at that time, such as the Internet changing the world and iPhone changing the mobile Internet.

I have experienced the VR industry before, 2015, 2016 was once called the first year of VR, and then it died down in less than a year.

The year before last was the first year of meta-universe, and after Facebook was renamed Meta, the concept of virtual world began to heat up again. As a result, when we look back, VR hardware has been upgraded in terms of parameters in the past few years, and some fundamental problems have not been solved, which is not essentially different from that of a few years ago. Some time ago, PicoVR also did a wave of marketing, but there was no storm.

And 5G. Because of looking up the data, I have read too many articles and analyses about how many changes 5G can bring. Although they are all ambiguous, I always think that when 5G is used, I can always see any changes, right? Now that I use 5G myself, I still don't know what 5G brings to the civilian end. I didn't seem to have any of the subversive scenes I imagined at first.

Ask the friends around you, it seems that there is no difference after changing 5G, it seems that the Internet speed has become faster, but there is no feeling that you have to use it. When you go to the restaurant, you still have to ask what your Wi-Fi password is. Perhaps the biggest feeling is that the set meal has become more expensive.

I would like to stress again that I am not denying these technologies and concepts. I believe they are good, the investment behind them is worth it, and they will be rewarded in the future. I'm talking about a sense of disparity, a sense of disparity that has been promised to change the world, but I don't feel it when I look at it a few years later. I don't know if I'm the only one who feels that way.

Last time I talked about VR with my friends, I said that I think the VR industry may be in the steam era. It is true that it leads to a definite goal, and it is indeed more productive than in the past, but it still has many insurmountable shortcomings, such as no one can bear to wear such a heavy head for a few hours, so the goal may be achieved in the end. Maybe it's something else.

This steam age may still exist in many industries.

So how to define or judge whether this technology or concept can be subversive?

In fact, there is no need to judge at all, because the technologies that change people's lives do not need the major media to write long speeches to tell ordinary people how good they are, and the power of definition is in the hands of the public. Most people will be naturally affected by these technologies, so that it won't be long before they feel that life becomes more difficult without it.

This is almost always the case with a detailed count of all kinds of technological progress in human history.

Why is the concept like meta-universe so popular? Why does it seem that there are "sensational" technology products every year, especially at the beginning of the year?

I think it is the capital markets that really need them.

As we all know, the dividend of the global mobile Internet has been eaten up, and the growth has already peaked. All kinds of industries that can be transformed by the Internet have basically been transformed, so that today you want to find another field to start a business, and find that they have already been done.

So we urgently need new technology to create new demand and create new opportunities.

Don't underestimate this matter, in fact, it is good for all of us. With new things, investment institutions have new investment goals, technology companies have new projects to do, and at the same time provide new jobs, we have our own media, right? there are new topics to write about.

So when ChatGPT quickly broke through so many users, people will spare no effort to pursue, to stir up the concept of chat robots. Seeing that Google is about to launch its own chat robot, many reports say that Google is in a hurry, which I think is completely exaggerated.

What ChatGPT can do in the end is a question mark, and it has not grabbed any AI market. The effect of the new version of Bing has yet to be determined. What is Google's hurry? I think many technology companies should be happy. If ChatGPT can really open up the market and get out of the business model, it will finally be able to recover some of the huge expenses they previously invested in AI.

As for whether ChatGPT is just a man-made pseudo-tuyere, I chatted with friends in the AI industry and felt very contradictory.

Compared with the empty-headed concept of meta-universe, ChatGPT is obviously more pragmatic and more down-to-earth.

The first is that its number of users is there, which is not to be underestimated in the short term.

Second, its commercialization is also very clear, subscription, charging API fees and deep binding cooperation with Microsoft all seem to have the possibility to make money.

Third, because the technology is really powerful, it still makes people feel quite smart, otherwise there would not be so many users so soon.

However, it also has something in common with previous industries.

First of all, its technology is not a revolutionary change, the underlying logic has not changed, it can be regarded as a "miracle" of spending money, and there must be a big gap between it and the AI in the human dream.

Second, my feeling of using it for a few days is that I don't have to use it. In fact, after the freshness has passed, I have been too lazy to use it. Even as a personal assistant, the function seems too crude to wait for other applications that use its language model to emerge.

But in the process of using it, I really found a good use scene, that is, to practice language. For example, if you want to learn English, it is obviously much cheaper to chat to such a powerful AI than to find an one-on-one foreign teacher. And facing the machine, the sense of shame is much less than facing the human.

In any case, I wonder whether ChatGPT will really lead an AI subversion of the industry, or will no one be interested at the end of the year.

Now that there are signs of a new trend, it's always okay to keep paying attention to new things. It may be more interesting to look back at what happened today when we wait until the end of the year.

This article comes from the official account of Wechat: IC Lab (ID:InsightPlusClub)

Welcome to subscribe "Shulou Technology Information " to get latest news, interesting things and hot topics in the IT industry, and controls the hottest and latest Internet news, technology news and IT industry trends.