The technological change detonated by ChatGPT is coming faster than expected.
On March 24th, ChatGPT ushered in a major update. The latest update shows that ChatGPT supports third-party plug-ins, can acquire new knowledge online, and can interact with 5000+ applications, and 11 plug-ins have been put on the shelf in one go.
The launch of these plug-ins means that ChatGPT will really lift the seal, because, in the previous version, ChatGPT could only query news before September 2021, but now it can not only directly retrieve the latest news but also help you query flights and hotel information, and even help you plan business trips, access major e-commerce data, help you compare prices, and even place orders directly. This means that ChatGPT will play a more important role in more fields. Some people said that this is like Apple’s “App Store.”.
The most important thing is that ChatGPT not only provides some plug-ins, but you can also generate some plug-ins through it. If you want a certain plug-in function, you can describe it in your natural language, and then GPT will help you generate this plug-in. This means that you don’t need to know how to program, you can also develop the plug-ins you want to use.
This update is more like a depth bomb. I don’t know what kind of waves it will cause in the field of large models, but the development of the application layer will definitely become more extensive.
In fact, before these plug-ins were really widely used, we found in foreign application communities that besides ChatGPT, there are actually many GPT applications, such as Jasper, YouChat, and so on. With the continuous advancement of GPT technology, more and more domestic companies have begun to pay attention to and apply GPT technology to improve their business efficiency and user experience.
This time ChatGPT “unsealed,” ChatGPT not only turned itself into a certain sense of the operating system but is also likely to lead to the birth of more ChatGPT-based applications.
At this point in time, we also collected and sorted out some popular GPT applications that have appeared around the world, and analyzed the main directions. Although these directions may still be very similar to the first batch of tool applications such as Horns that appeared after the emergence of Android, they may not be able to reach the end of the AI era, but they may also provide a good reference for latecomers.
On March 21, Bill Gates stated in the blog post “The Era of Artificial Intelligence Has Been Opened” that OpenAI’s GPT artificial intelligence model is a technological change comparable to the introduction of the GUI (graphical user interface) in 1980. On the same day, NVIDIA CEO Huang Renxun said at the press conference that “we are in the iPhone era of artificial intelligence.”
Under the new technological changes represented by GPT, an era of AI native application explosion may be coming.
More than 600 GPT applications have been uploaded
When the eyes of the whole world are on ChatGPT and are shocked by the capabilities it has demonstrated, we also know what other vertical applications of GPT are there and what kind of effects they can achieve.
Finding these GPT applications today is a complex project. Fortunately, we found some community websites that summarize GPT applications, where we can find almost all related GPT application information.
GPT-3 DEMO is one of them. This website collects various applications and application cases based on ChatGPT and GPT-3. There are currently more than 600 applications, including AI writing, AI search engines, AI editing, AI programming and testing, AI design, and other aspects. In addition to this kind of community dedicated to GPT applications, we also found some other community websites with broader categories.
For example, Futurepedia, which is a brand new online platform for retrieving AI tools and has the largest AI tool database, updates AI tools every day. Futurepedia brings together many categories of AI tools, including 3D, art, voice editing, character creation, customer service, development tools, educational assistants, fashion, interesting tools, etc., as well as the most familiar text AI and image generator applications. There are also GPT applicant applications that are included.
In addition, we also found Cre AI Times, the AI (artificial intelligence) tool collection navigation website that claims to browse the largest AI database, SaaS Al Tools, Futurepedia, and the AI Research Institute. These websites will collect and update new AI tools every day to help people discover new GPT applications faster.
These GPT application communities are of great reference value to practitioners who think about how to combine GPT applications with existing business scenarios.
Through the application community described above, we learned that GPT has a wide range of applications. In addition to chatbots like ChatGPT, they can exert unique advantages in many other aspects, such as generating articles, images, and other fields. We have also sorted out some of the popular applications of these categories.
dialogue system
GPT can be used to build dialogue systems, including chat robots, customer service systems, etc. Representative applications include:
ChatGPT: A Language Model for Optimizing Conversation. The conversational format allows ChatGPT to answer follow-up questions, admit mistakes, challenge incorrect premises, and reject inappropriate requests.
• Andi: Andi is iterating with generative AI. Andi is like a new search engine, it not only provides web links but also provides you with answers through dialogue.
• Quickchat: Quickchat is a fully conversational AI chatbot powered by OpenAI’s GPT-3, a state-of-the-art language model with 175 billion parameters. Users can use it to automate customer support, online applications, search internal knowledge bases, and many other tasks.
language generation
Language generation is one of the most common applications of GPT, which can be used to generate various types of text, such as copywriting, news reports, novels, poetry, etc. Representative applications include:
• Writesonic is an AI writer that creates SEO-friendly content for Blogs, Facebook Ads, Google Ads, and Shopify for free.
• Copy.ai is an artificial intelligence copywriter that can generate high-quality copywriting for businesses.
• Rytr is an AI writing assistant that helps users create high-quality content in seconds and at a fraction of the cost
• Compose.ai is one of several automated writing tools built on top of GPT-3. This is a free Chrome extension that automates writing.
• EduRef dissertation writing is also one of the automated writing tools based on GPT-3, which can help fresh graduates write dissertations.
• Inspo is an AI-based inspiration writing robot. With its natural language ability and internal engine of continuous learning, it helps users write efficiently, stimulate creativity, acquire knowledge, and make decisions.
image generation
GPT can also generate images through text descriptions. Representative applications include:
• The DALL·E application is built on a 12 billion parameter version of GPT-3, trained to generate images from text descriptions using the text-image pair dataset. Now the updated second-generation DALL·E 2 can create realistic images and art based on descriptions in natural language.
These different applications have already played a role in the corresponding fields. In the future, these GPT applications may play a unique role in different industrial scenarios. With the advancement of technology, we will also see GPT applications with more categories and a wider range of applications.
What are the current mainstream applications?
We summarize and classify the popular applications in the community and share them with you.
The first is the hot ChatGPT, a language model for optimizing conversations. The conversational format allows ChatGPT to answer follow-up questions, admit mistakes, challenge incorrect premises, and deny inappropriate requests. We also found tools similar to ChatGPT, including Andi. Andi is more like a new search engine, but it doesn’t just provide links, it also provides answers—like having a conversation with a friend.
The previously mentioned popular tool Fliki can create a video from a script or a blog post in 2 minutes with realistic sound, turning this text content into a video.
In addition, there are some powerful productivity tools, such as Namelix, which can quickly generate a short brand name for a business with a few simple choices from the user. Another app, Tome, uses AI to help you make presentations. The application of Codeium is more like a gospel for programmers. It can help programmers write better code and help them develop products. In addition, there is Browse AI, an application that can monitor and extract data from websites, and Adcreative.ai, an application that helps advertisers provide better ideas. I have to say that after the experience, we feel that these tools will greatly improve our production efficiency.
There are other useful and entertaining tools, such as Writesonic, that can help you write articles, such as creating some SEO-friendly article content for Facebook Ads, Google Ads, and Shopify. Similar to this is Copy.ai, an artificial intelligence copywriter that can generate high-quality copy for businesses. Looka Logo Maker can help create a brand logo that belongs to you. Another interesting app is Talk to Books because it allows people to find books and their contents by asking questions. Another app, Tutor AI, is more like an AI tutor. You can enter the content you want to know on it, and it will display it in categories to help you understand it. There is also an app, Piggy To, that can help workers do PowerPoint presentations. By typing in prompts, the tool will generate 7-10 slides in a format suitable for mobile devices, saving time and ensuring that the content is visually appealing.
Of course, there are some applications in vertical fields. One such popular app is Midjourney, which has been gaining popularity recently and helps people create art. It can produce pictures through your text description, and the style can also be defined by you. There are also some image generation applications, such as Playground AI, DALLE 2, and NightCafe Studio, that can also help people create art. In the field of music, Riffusion can generate music based on text prompts. Boomy also helps people who’ve never made music create original songs in seconds
It is not easy to win the favor of investors.
Any technological progress must be budgeted, and the GPT application that investors are optimistic about is likely to be an application with more “money”.
We found that many applications in the application community have also received capital support. For example, Andi is a startup from San Francisco, created by a small team. Angela Hoover is the co-founder and CEO, and Jed White is the co-founder and CTO. Andi’s investors include Y Combinator, Gaingels, Goodwater Capital, K20 Fund, Acacia Venture Capital Partners, Fepo Capital, and BBQ Capital. The company currently has $2.5 million in financing.
Another text-generating application, Writesonic, also comes from a San Francisco start-up company. Their founder and CEO is Samanyou. They have received a $2.6 million seed round of financing led by HOF Capital. We also found CopyAI, a startup building an AI-powered copywriting tool for business clients. Their investors include Wing Venture Capital, Tiger Global Management, Sequoia Capital, etc., and they have received $13.9 million in financing.
In addition, Compose AI, which offers a free Chrome extension that claims to reduce writing time by 40% through auto-completion, has raised $2.1 million in funding. Looka, another artificial intelligence company that helps companies design logos, has received more than $6 million in funding, and their founder, Dawson Whitfield, is a logo designer himself.
Another app, Playground AI, a free-to-use online AI image creator, has raised an undisclosed amount in an oversubscribed pre-seed round led by East Ventures and Mirana Ventures, with participation from Arc Capital and angel investors.
Most of these applications that have been favored by investors are text generation applications, precisely because large language models are ubiquitous, but they are most obvious in the field of writing assistance, where the technologies they apply are relatively mature and the products are easier known to the general public.
Fliki and Namelix are the ones that have not received financing but are relatively popular. Fliki is a text-to-speech and text-to-video converter that helps users create audio and video content using AI voice in less than a minute. In the Futurepedia community, 2140 people like this tool, ranking it second. Another app, Navelix, offers advertising solutions for brands, publishers, and programmatic buyers. In the Futurepedia community, 1894 people like this tool, ranking it third. They are all applications for targeted layouts on some vertical tracks.
From here, we can find that most of these very popular applications are related to video, audio, and images or provide solutions for enterprises, but this type of technology has not been verified more, and these contents It seems that it can be realized after the plug-in of the first-party application comes out, so investors seem to be more rational and have not invested in these future models.