In the early morning of March 24, OpenAI released ChatGPT plugins (ChatGPT plug-in set), which can incorporate third-party applications into GPT to provide services to users.
Plugins is expected to help OpenAI build an ecosystem based on large models, if Gpt3.5 is to let data learn to speak, then the introduction of plugins is to plug in the wings of the data, or can be expanded at will.
Looking at some brief analysis, most look at this "App Store" moment from the perspective of OpenAI, and the importance of this moment will not be repeated. This article will analyze the impact of this event on the data industry and the Internet industry from the perspective of "data".
This article will focus on the following insights
1, Microsoft's Newbing's ambition
2. The transformation of Internet applications
3. Data as a Service brings about the transformation of the data service industry in the era of digital economy
First, the subversion of information aggregation applications by plugins
In the not-too-distant future (perhaps tomorrow or next month), Chatgpt is likely to replace applications as the traffic entry point for all information retrieval.
Plugins is equivalent to removing the middle layer of the application, allowing users to interact directly with the data, using the GPT model to directly read the data and associate the data in it, and directly feedback to the information contained in the user data.
For example, with "Dianping", you need to perform multiple search browsing comparisons of location, taste, price, and evaluation to make a decision, but if the large model accesses the "Dianping" plugin, the large model will directly call all the data on Dianping to answer your question "Where to eat tonight?" "This problem is a direct interaction with the data and the user, and direct feedback of data insights/decision-making to the user.
All information aggregation applications must adapt to new models and face the dilemma of being eliminated if they are not integrated, after all, people need the insights behind the data, not waste time browsing countless data and facing countless difficult choices. (Of course, there is still a "shopping" demand, that is, I want to browse and browse in an app to find inspiration)
If the Chatgpt big model is compared to a brain, then plugins are allowed to directly instill skills into the brain in the form of indoctrination, remember the plot of Neo learning martial arts in The Matrix, martial arts skills are implanted into Neo's brain like a plugins, and since then Neo can martial arts no one can stop!
Pugins is a big step forward in the commercialization of large models
In addition to open API direct sales of underlying capabilities, advertising is also a must, advertising and interaction integration can well cover the cost of computing power generated by interaction, and the advertising generated in the interaction will be more accurate and efficient.
Plugins as an "App Store" ecology, I judge that there will still be two profit models in the future, the first is to pay for loaded plugins, and the second is to call relevant plugins when the user interacts, and plug-in providers can come with ads.
The first is easy to understand and equates to paying for your own data knowledge base. To understand the sea of stars, you may need to spend 5 yuan to buy an "astronomical geography" plugin, to understand business data to help strategic decision-making, you need to buy a consulting company's "business consulting" plugin.
The second is precision advertising, when your question needs to call the plug-in database to answer comprehensively, the plug-in provider can add targeted advertising recommendations and links at the end of the text or in the text. I take Newbing's interface as an example to look forward to the advertising form in the future era of large models——
On March 13, BlueFocus (SZ300058)$ BlueLabel Media officially announced a strategic cooperation with $Microsoft (MSFT)$ Advertising and became its official agent; At the same time, the company will cooperate with Microsoft on OpenAI-based technology products, so that more overseas customers can enjoy the dividends of AI development through keen insight and excellent strength. Customers going overseas are expected to have better interaction with overseas consumers through the ChatGPT function integrated in New Bing.
On March 15, Microsoft is holding a roadshow and holding a private meeting with advertisers and agencies to showcase the ad format of Bing Search in its beta version of ChatGPT. Microsoft hopes to capitalize on the focus on New Bing and OpenAI to drive business monetization.
News linked together - chatgpt-based new advertising marketing must also be an important step in OpenAI business, chatgpt and other large models of advertising monetization ideas will be similar to the legend I made above.
OpenAI plugins give wings to data
The core link in the digital economy, "data transactions", has always been in an awkward position, and there are many controversies about how to mine the value of a large data package, how to price it, and how to generate explicit benefits. But the emergence of OpenAI chatgpt can complete the direct transformation from data to insights, plugins give professional data a direct monetization channel.
1, public data (data that search engines can search) can directly enter the large model by being used as a search engine plug-in, from the perspective of Microsoft's control of OpenAI, the browser plug-in in OpenAI in the future is likely to be monopolized by Newbing.
Continue to connect the news - on March 24, people familiar with the matter revealed that Microsoft threatened to prevent search engine competitors from obtaining relevant data if they continued to use Microsoft's Internet search data as the basis for their own AI chat products.
Combined with the above Microsoft Bing's advertising strategy progress everything becomes clear, Newbing will first master the traffic entrance of Chatgtt through the exclusive or priority strategy, and all the real-time data that needs to be called for interaction in chatgtt will be provided by Newbing, and the magic weapon of search engine traffic monetization will be exclusively mastered by Newbing.
(Corresponding to China, this ambition is Baidu's.) If Baidu hasn't thought of it, Li Yanhong believes Baidu also understood after reading this article)
2. Professional data, such as consulting company data, is often sold to institutions in the form of commercial research reports, but plugins allow consulting companies to access their own data to GPT as plug-in providers, increasing C-end subscriptions from a single institutional sales model.
For example, the basic data that a consulting company needs to do to make a business analysis report of hundreds of pages may be tens of thousands of pages, and these basic data are good call resources, and the number of calls can measure the value of a database, and can also realize secondary sales monetization by calling advertisements and subscription fees.
For example, data consultants such as $0 (SZ301169)$ can independently provide data calls as plug-ins
For example, the research report of a brokerage can be made into a plug-in through a data distributor such as a flush and integrated brokerage report
For consulting companies that collect, sort, process, and sell
1. Improvement of sales efficiency - customer group expansion, toB + toC, B-side provides molding reports, C-end data insight call
2. Reshaping the business model - the efficiency of data processing on the cost side is improved, and all data has become a sellable product, which can be retailed and wholesale, and processed and blank.
Following this idea, through the integration of many vertical data through the large model, you can find the data monetization ideas and paths, so as to realize the closed loop of data from production, processing to sales flow, truly realize that the market value determines the price of data, and the closed-loop business path realizes the distribution of benefits of data producers, processors and distributors, which will greatly stimulate the development of the digital economy.
Combining the above three points, OpenAI plugins will give wings to data, and data-as-a-service will enter everyone's life.
In the era of digital economy, data is oil, and OpenAI + plugins (large models + vertical calls) drilled the first oil well!
Source: Snowball Author: Florist Show